JPWO2005111880A1 - Behavior analysis device - Google Patents

Abstract

Although the sales itself can be grasped from the purchase history of the product, it is difficult to identify the cause of the sales.
In the customer behavior analysis device 20, the position acquisition unit 30 acquires the position information and time information of the customer in the store. The space acquisition unit 36 acquires space information that represents the space in the store as a position in the virtual space. The route acquisition unit 34 generates route information that indicates the customer's travel route in the space of the store as a position in the virtual space based on the position information, the time information, and the space information. The speed acquisition unit 32 generates speed information of the customer based on the position information and the time information. The pattern determination unit 42 determines which of the action pattern types the customer's purchasing action matches, based on the action pattern data, which is the type of purchasing action, route information, and speed information. The output processing unit 48 outputs the determination result.

Description

  The present invention relates to an action analysis device, and more particularly to a technique for analyzing an action of a moving body in space by position detection using a technique such as wireless communication.

In recent years, fierce service competition has been occurring in the retail industry centered on large stores. Each retailer is not only trying to cut costs from all angles in order to reduce the selling price of products, but also tackling the issue of how to provide customers with added value that other stores do not have. Customer needs are diversifying, and it is important to respond to such needs, provide accurate product information at the right time, and arrange sales floors and product lines that meet the customer's needs. Directly connected to efficiency.
JP, 2003-223548, A

  One of the tools that captures the diversifying needs of customers is the POS (Point Of Sales) system. The POS system can collect statistics on the sales of products and the demographic of consumers based on the input data to the cash registers of retail stores. However, although the POS data reveals “sellable products” and “unsellable products”, it is difficult to identify the causes related to the customer's behavior such as “why sell or why not sell”. Therefore, it is difficult to determine whether the unsold product has a problem in the product itself or simply in the display method, and it is difficult to take appropriate measures.

  On the other hand, in order to promote sales promotion, it is necessary not only to analyze the behavior of customers in stores but also to improve the efficiency of employees, and the behavior analysis of employees using IT technology can be realized with high accuracy. It can be an effective tool across all industries. In other business and applications, various data that can be used for marketing can be obtained by analyzing human behavior.

  The present invention has been made under the above background, and an object of the present invention is to analyze the behavior of a moving body in a space and provide useful information.

  In order to solve the above problems, the behavior analysis device according to an aspect of the present invention is a position that acquires position information indicating a position of a moving body that changes with time in a predetermined real space and time information indicating the positioning timing. An acquisition unit, a space acquisition unit that acquires spatial information that represents the real space as a position in the virtual space, and a movement path of the moving body in the real space in the virtual space based on the position information, the time information, and the spatial information. Relationship between the route acquisition unit that generates the route information indicated by the position, the speed acquisition unit that generates the velocity information indicating the moving velocity of the moving body based on the position information and the time information, and the movement route and the moving velocity of the moving body Based on the route information, the speed information, and the plurality of action pattern data, the action of the moving body is based on the pattern storage unit that stores a plurality of action pattern data in which different action patterns are defined as the action patterns based on A pattern determination unit that determines which of the types matches, and an output processing unit that outputs the determination result are provided.

  Here, the "moving object" is, for example, a customer who has visited the store, a shopping cart or shopping basket of the customer, an employee of the store, a visitor or employee in a recreation facility such as a theme park, an athlete in a competition space, or the like. Indicates a person or its adjuncts moving in space. The "behavior pattern data" defines a type of action pattern of the moving body in the entire predetermined space and a type of action pattern of the moving body in each of a plurality of regions included in the predetermined space. It may be either. According to this aspect, not only the moving route of the moving body in the space but also its moving speed is acquired, so that even if the moving route is the same, for example, if it is a store space, it takes time to simply pass through the sales floor and consider purchasing products. You can tell the difference. Therefore, it is possible to accurately infer a person's behavior such as a customer's purchase behavior or an employee's behavior, and the factors of such behavior.

  The space acquisition unit acquires, as space information, respective coordinates for specifying the arrangement of a plurality of regions included in the real space in the virtual space, and the route acquisition unit calculates the movement route of the moving body in the real space in the virtual space. The route information shown at the position above is generated together with information regarding which one of the plurality of regions the mobile body has traveled, and the speed acquisition unit, as speed information, the average speed of the mobile body in all of the plurality of regions, The information indicating at least one of the average speed of the moving body and the instantaneous speed of the moving body in each of the plurality of areas is generated, and the pattern determining unit determines which of the plurality of areas the moving body has traveled. The matching of the action patterns may be determined based on the route information including the information, the speed information, and the plurality of action pattern data.

  The position acquisition unit acquires position information indicating the position of the moving body in a space in a predetermined store as a real space and its time information, and the space acquisition unit arranges a plurality of sales floors included in the store in a virtual space. Each of the coordinates for identifying is acquired as spatial information, and the route acquisition unit displays the route information indicating the moving route of the moving body in the space in the store in a position in the virtual space, among the sales floors where the moving body is located. The speed acquisition unit generates the average speed of the moving body in all of the plurality of sales floor areas, the average speed of the moving body in each of the plurality of sales floor areas, and the speed of the moving body. Generates information indicating at least one of the instantaneous speeds, the pattern determination unit, the route information including information regarding which one of the plurality of sales floors the mobile has passed, speed information and a plurality of behavior pattern data. It is also possible to determine the matching of the action patterns based on. In this case, since the behavior of the customer can be analyzed for each sales floor, accurate useful information that leads to sales promotion can be obtained for each sales floor or product.

  This behavior analysis device, intensity information indicating the intensity of the radio waves transmitted or received by the communication device carried by the moving body, at least one of a plurality of wireless communication devices installed in different positions in the real space It may further include a strength acquisition unit that acquires the position information and a position recognition unit that generates position information indicating the position of the moving body and time information based on the strength information. Here, the “wireless communication device” may be a wireless base station also called a so-called access point. In this case, the position information and the time information can be acquired by using the widely used wireless LAN technology as the hardware, and the behavior of the mobile body such as the customer who visits the store can be analyzed at a relatively low cost. The “communication device carried by the mobile unit” is mainly assumed to be a wireless LAN terminal, but this may be realized by an IC tag (RFID) having a built-in wireless function.

  This behavior analysis device is configured such that at least one of a plurality of wireless communication devices installed at different positions in a real space and a wireless communication medium carried by a mobile body, one of which transmits a radio wave and then the wireless communication device. A position recognition unit may be further provided that generates position information indicating the position of the mobile body and time information based on the length of time until one receives a radio wave as a reply by the other that receives the radio wave. The “wireless communication device” and the “wireless communication medium carried by a moving body” here are the same as above, and may be a wireless LAN or an IC tag.

  In this behavior analysis device, an IC tag or a wireless communication device carried by a moving body communicates with any one of a plurality of wireless communication devices or a plurality of IC tags installed at different positions in a real space. A position recognition unit that generates position information and time information indicating the position of the moving body based on whether or not the position is being detected may be further included.

  This behavior analysis device identifies an object that is moving between a plurality of images acquired by the image acquisition unit that acquires an image captured from a predetermined angle in a real space as a moving body, and And a position recognition unit that generates position information and time information of the moving body by recognizing the position of the object in the space. The “imaging” referred to here may be not only imaging using a CCD sensor or a CMOS sensor but also special imaging using reflected waves when ultrasonic waves, microwaves, infrared rays, electromagnetic waves, etc. are applied to an object. .

  The pattern determination unit may acquire the data indicating the purchase content of the customer who is a mobile body and estimate the factor of good or bad of the purchase based on the data indicating the purchase content and the type of the matching action pattern. For example, the POS data may be used to acquire data indicating the purchase content of the customer. Since it is considered that there is a correlation between the degree of sales of the product and the behavior pattern of the customer, it is possible to relatively easily grasp the cause of the good sales and the cause of the poor sales by combining the two.

  The pattern determination unit determines the position of the settlement area where the customer stayed to purchase the product and the date and time of the product purchase based on the route information, and outputs the data indicating the purchase content corresponding to the position of the settlement region and the date and time of the product purchase. The acquired data may be associated with the route information.

  The pattern determination unit is based on route information, speed information, and a type of matching action pattern, and attribute pattern data in which conditions of the route information, speed information, and type of action pattern are defined for each customer attribute. The attributes of the mobile customer may be estimated. There may be a high degree of correlation between customer attributes and behavior patterns, and by preliminarily converting such relationships into attribute pattern data, the customer attributes can be estimated relatively easily according to the behavior patterns. You can

  The pattern determination unit acquires the attribute of the customer who is a mobile body, stores the determined behavior pattern in the pattern storage unit in advance as attribute pattern data associated with the attribute, and stores the attribute pattern stored in the pattern storage unit. By referring to the data, the attribute of the customer whose attribute is unknown may be estimated.

  A product information storage unit that stores information that can be provided to customers regarding products to be sold in association with the sales floor of the product and the type of action pattern, and can be provided to customers who are mobile according to the type of matching action pattern An information distribution unit that transmits information to a communication device held by the customer may be further included. Since the current state can be estimated from the current position of the customer and the behavior pattern of the customer, it is possible to promote the sales of products and improve the sales efficiency by actively distributing the information that matches the position and the state. .

  In the plurality of behavior pattern data stored in the pattern storage unit, the first pattern showing a case of passing through the sales floor area as a type of the purchase behavior of the customer who is the moving body in the sales floor area to be determined, and comparison in the sales floor area For each of the second pattern showing a case of purchasing or considering purchasing after stopping for a short period of time, and the third pattern showing a case of staying for a long time in one sales floor area, stay time, presence/absence of stoppage, and A criterion relating to at least one of acceleration and deceleration is set in advance, and the pattern determination unit determines which of the first pattern, the second pattern, and the third pattern the behavior of the customer who is the moving body matches. You may judge. The term “stop” as used herein may include a case where the vehicle stays to the extent that it is considered to be substantially stopped, and a case where the traveling speed is low such that the traveling speed is significantly lower than the average speed. For example, the presence/absence of sudden deceleration may be set as the reference regarding the “acceleration/deceleration”. Thereby, the purchase behavior of the customer can be accurately estimated.

  The pattern determination unit calculates the staying time from the entry to the exit of each moving body for each of the plurality of sales floor areas, and based on the route information and the data indicating the purchase details, the moving body that purchased the product and The sales area of the purchased product is specified, and the staying time of the moving body in the sales area is specified as the staying time when the product is purchased. The actual shortest time and the longest time for each zone are specified, and the average shortest time is calculated by averaging the actual shortest time over multiple days or multiple time zones for each sales floor area, and this is used as the first pattern. The first threshold value for distinguishing the second pattern is stored in the pattern storage unit in a form corresponding to the sales floor area, and the substantially longest time over a plurality of days or a plurality of time zones is averaged for each sales floor area. The calculated average longest time is stored in the pattern storage unit as a second threshold value for distinguishing the second pattern from the third pattern, and is stored in a certain sales floor area of the moving body that is the determination target of the action pattern. If the staying time is less than the first threshold value associated with the sales floor area, it is determined that the behavior pattern of the mobile object in the sales floor area matches the first pattern, and the mobile object that is the determination target of the behavior pattern. If the staying time in a sales floor area is equal to or more than the first threshold value associated with the sales floor area and falls below the second threshold value, the action pattern of the mobile body is the second pattern with respect to the sales floor area. When the staying time in a certain sales floor area of a moving object whose action pattern is to be determined is equal to or more than a second threshold value associated with the sales floor area, It may be determined that the action pattern matches the third pattern.

  Here, "when the value is below the first threshold value" as the determination criterion of the first pattern does not mean that values in the vicinity of the first threshold value are strictly distinguished. It may be determined that the first pattern and the first threshold value are also in the predetermined range. Similarly, the criterion of the second pattern "when it is equal to or more than the first threshold value and falls below the second threshold value" or the criterion of the third pattern "the second threshold value Also in the case of "when it is lower than", it does not mean that values near each threshold value are strictly distinguished. In addition, “substantially the shortest time” and “substantially the longest time” are intended to exclude an abnormal value that is extremely short or long, and is a plurality of values regarding the stay time specified for one day. Of these, for example, the minimum time and the maximum time of excluding the values within the range of 10% above and below may be set as the actual minimum time and the maximum time, respectively.

  Further, the "plurality of days" may be designated by a day of the week, a combination of a month and a day of the week, or may be designated by a method of dividing weekdays and holidays. The “plurality of time zones” may be specified, for example, by only the time zone, by the combination of the day of the week and the time zone, or by the combination of the month, the day of the week, and the time zone, It may be specified by a combination of how to divide weekdays and holidays and time zones.

  The pattern determination unit displays, for each sales floor area, the relationship between the staying time in that area and the number of customers who stayed in the area, and among the customers whose staying time was shorter than the first time value, The first time value is specified as the first time value such that the ratio of the total sum and the total sum of customers who have not purchased products in the sales floor area becomes a predetermined ratio, and multiple time values are specified for each sales floor area. Calculate the average value of the first time value over a day or a plurality of time zones and store this as the first threshold value for distinguishing the first pattern and the second pattern in the pattern storage unit, and for each sales floor area When expressing the relationship between the staying time and the number of staying customers in the area, the total sum of the customers who purchased products in the sales area among the customers who stayed longer than the second time value and the purchases in the sales area The second time value is specified as the second time value such that the ratio with the total number of customers who did not have a predetermined ratio is specified for each day or each time zone, and each sales area is divided into multiple days or multiple time zones. The average value of the time values of 2 is calculated and stored in the pattern storage unit as the second threshold value for distinguishing the second pattern from the third pattern, and it is for the moving object that is the determination target of the action pattern. If the staying time in the sales floor area is less than the first threshold value associated with the sales floor area, it is determined that the behavior pattern of the mobile body in the sales floor area matches the first pattern, and the behavior pattern is to be determined. If the staying time in a certain sales floor area of a mobile object is equal to or more than the first threshold value and is less than the second threshold value associated with the sales floor area, the action pattern of the mobile object in the sales floor area is If it is determined that the moving object is determined to match the second pattern and the staying time in a certain sales floor area of the moving object that is the determination target of the behavior pattern is equal to or more than the second threshold value associated with the sales floor area, It may be determined that the action pattern of the moving body matches the third pattern.

  The pattern determination unit determines a customer, which is a moving body, according to at least one of a flow line distance from entry to exit in the entire space of the store, entry time, stay time, average speed, and total number of stops. Are divided into a plurality of segments and the first threshold value and the second threshold value are calculated for each segment and stored in the pattern storage unit. After determining which of the following applies based on at least one of the flow line distance from entry to exit in the entire space in the store, entry time, stay time, average speed, total number of stops , The first pattern, the second pattern, or the third pattern may be determined based on the first threshold value and the second threshold value corresponding to the corresponding segment. ..

  The pattern determination unit classifies the behavior of the customer, who is a mobile object, into the first pattern and the second pattern and the third pattern, and calculates the ratio of product purchase in each pattern. Alternatively, a reference value for estimating a factor of whether or not a product is purchased may be determined.

  In this case, the presence or absence of the customer's interest in the product can be estimated for each sales floor area based on which pattern matches. For example, if it matches the first pattern, it can be estimated that there is no interest in the product in the sales floor area, and if it matches the second pattern or the third pattern, it can be estimated that the purchaser has an interest in the product in the sales floor area. Depending on the presence or absence of such interest, it is possible to estimate whether the product was not purchased due to lack of interest as a factor that did not purchase it, or if it was interesting but did not buy it, It is possible to estimate factors such as whether it was based on the purchase or impulse purchase, and it is possible to obtain useful marketing information.

  The pattern determination unit determines the number of purchases or the purchase amount of the product for any number of the first pattern, the second pattern, and the third pattern for each sales floor area as a factor of the quality of purchase by the customer who is a mobile body. Alternatively, the number of times of the first pattern, the second pattern, and the third pattern may be calculated. As a result, it is possible to accurately estimate the factor of whether the customer is good or bad.

  It should be noted that any combination of the above constituent elements, and those in which the constituent elements and expressions of the present invention are interchanged among methods, devices, systems, etc. are also effective as an aspect of the present invention.

  According to the present invention, useful information can be provided by analyzing the behavior of a moving body in a space.

It is a functional block diagram which shows the structure of a customer behavior analysis system. It is a figure which shows typically the classification|category based on the state of a customer, and the relationship between the number of purchasers. It is a figure which shows the external appearance of a shopping cart. It is a figure which shows typically the sales floor arrangement in a store in the virtual space defined by the space information. It is a figure which shows the sales floor area hierarchized typically. It is a figure which shows typically the example of the moving path of the shopping cart in a sales floor area. 6 is a functional block diagram showing the configuration of a customer behavior analysis system in Example 2. FIG. It is a functional block diagram which shows the structure of the customer behavior analysis system in Example 3. It is a functional block diagram which shows the structure of the customer behavior analysis system in Example 4. An example of an image taken from a bird's eye view of a sales floor area is illustrated. It is a figure which shows the imaging condition of the sales floor image in FIG. 10 typically from a side. It is a figure which applied the cylindrical body model to the sales floor image of FIG. It is a figure which shows typically the method of detecting the same moving body in the overlapping imaging range. It is a figure which shows typically the example of the imaging range of each camera in a store. It is a functional block diagram which shows the structure of the customer behavior analysis system in Example 5. It is a functional block diagram which shows the structure of the customer behavior analysis system in Example 6. It is a figure which shows the relationship between the stay time and the number of customers in a sales floor area. It is a figure which shows the example of a screen which displays the list of the length of time for every action state of an employee.

Explanation of symbols

  10 customer behavior analysis system, 20 customer behavior analysis device, 22 communication unit, 24 strength acquisition unit, 26 position recognition unit, 28 position information storage unit, 30 position acquisition unit, 32 speed acquisition unit, 34 route acquisition unit, 36 space acquisition Section, 38 spatial information storage section, 40 data setting section, 42 pattern determination section, 42 action pattern data storage section, 46 determination result storage section, 48 output processing section, 50 information distribution section, 56 display section, 58 product information storage section , 60 shopping cart, 62 liquid crystal display for cart.

(Example 1)
Even if you can understand "sellable items" and "unsellable items" by referring to the data indicating purchase contents such as POS data, it is possible to purchase "why sell" and "why not sell" It is difficult to identify. According to the system of the present embodiment, the intention of the customer to go to the sales floor and the process leading to the product purchase can be grasped based on the statistical data, and each customer's interest in the products in the sales floor can be objectively estimated. You can get useful marketing information.

  FIG. 1 is a functional block diagram showing the configuration of the customer behavior analysis system. The customer behavior analysis system 10 includes a plurality of wireless LAN base stations, a first access point 12, a second access point 14, a third access point 16, and a customer behavior analysis device 20. The “plurality of wireless LAN base stations” is provided with at least three access points for positioning between the mobile body and each base station using the strength of the radio wave. In this embodiment, three access points including the first access point 12, the second access point 14, and the third access point 16 are illustrated, but actually four access points are provided depending on the size of the store and the size of the sales floor area. The above access points may be used.

  Here, as the wireless LAN system, for example, IEEE 802.11a/b/g is used. The first access point 12, the second access point 14, and the third access point 16 communicate with each other via a wireless LAN with a communication device provided in a customer's shopping cart, and a customer via a LAN 18. Data is transmitted to and received from the behavior analysis device 20. The customer behavior analysis device 20 acquires POS data as data indicating purchase details from the POS server 19 via the LAN 18. The POS data includes information regarding customer purchases. The POS server 19 is connected to a plurality of cash registers (not shown), and obtains information related to customer purchases such as the product name or product ID sold, the number of products sold, the sales date and time, and the customer ID from each cash register. The customer ID may be, for example, the ID of a membership card issued to the customer, in which case the POS server 19 does not acquire the customer ID from a customer who does not own the membership card.

  The customer behavior analysis device 20 includes a communication unit 22, a strength acquisition unit 24, a position recognition unit 26, a position information storage unit 28, a position acquisition unit 30, a speed acquisition unit 32, a route acquisition unit 34, a space acquisition unit 36, and a space information storage. Unit 38, data setting unit 40, pattern determination unit 42, action pattern data storage unit 44, determination result storage unit 46, output processing unit 48, information distribution unit 50, operation input unit 52, control unit 54, display unit 56, product The information storage unit 58 is provided. The communication unit 22 transmits/receives data to/from the first access point 12, the second access point 14, the third access point 16, and the POS server 19 via the LAN 18.

  The strength acquisition unit 24 receives the data indicating the radio field strength from the first access point 12, the second access point 14, and the third access point 16 via the communication unit 22. This radio wave intensity is the intensity of a radio wave transmitted/received to/from the communication device provided in the shopping cart used by the customer. The communication device of each shopping cart transmits data indicating the intensity of a radio wave transmitted/received to/from the first access point 12, the second access point 14, and the third access point 16, respectively. It transmits to at least one of the access point 14 and the third access point 16, and the intensity acquisition unit 24 acquires this via any of the first access point 12, the second access point 14, and the third access point 16. .. The position recognizing unit 26 recognizes the position of the shopping cart from the difference in the radio field intensity from the same shopping cart received by the strength acquiring unit 24 from the first access point 12, the second access point 14, and the third access point 16. .. That is, the position recognition unit 26 obtains the distance from each access point to the shopping cart based on the radio field intensity received from each access point, and recognizes the point where the distance intersects as the current position of the shopping cart. The position recognizing unit 26 generates the position information, which is the current position of the customer that changes with time in the store, and the time information indicating the timing of positioning the position information, and stores the time information in the position information storage unit 28. This position information indicates the relative position of the shopping cart with respect to the three access points, and is represented by, for example, three-dimensional coordinates (x, y, z). Further, it may be represented as (x, y, z, t) together with the time information, or may be represented by two-dimensional coordinates (x, y) or (x, y, t). The position information of the three access points is preset based on the information input by the administrator. That is, when installing each access point, the administrator inputs each approximate position through the operation input unit 52, and the position recognition unit 26 corrects the input value based on the radio field intensity between the access points to correct each position. The position information is stored in the position information storage unit 28.

  The position acquisition unit 30 acquires, from the position information storage unit 28, position information and time information of a shopping cart used by a customer who has shopped or is shopping in the store. That is, the position acquisition unit 30 acquires the position information and the time information continuously measured from the time the customer enters the store and starts using the shopping cart until the use ends.

  The speed acquisition unit 32 generates speed information indicating the moving speed of the customer based on the position information and the time information. For example, the speed acquisition unit 32 determines the average speed of all shopping carts, the maximum speed, and the average stop count, the average speed from the start of use to the end of use of each shopping cart, the maximum speed, and the stop count, and the average speed of each sales floor area, The maximum speed, the average number of stoppages, the average speed for each sales area for each shopping cart, the maximum speed, the number of stoppages, and the speed of the shopping cart currently in use are calculated. Shop floors that are determined to match the flow line distance from the start to the end of use of shopping carts throughout the store, entry time, stay time, average speed or average speed excluding the third pattern, total stop count, and the first pattern A combination or permutation of areas, a combination or permutation of sales floor areas determined to match the second pattern, a combination or permutation of sales floor areas determined to match the third pattern, according to at least one of the aspects, Divide customers into multiple customer segments. For example, it may be classified into high speed, medium speed, and low speed based on an arbitrary speed. In this case, for example, customers who are slow on average, such as the elderly and customers with disabilities, and customers who are fast on average, such as young people, are separated in terms of data. Can be increased. Parameters such as a minimum value, a maximum value, an average value, and a threshold value, which will be described later, are calculated for each customer segment, and the action pattern determination and the customer attribute estimation are processed for each customer segment. However, as a modification, the parameters such as the minimum value, the maximum value, the average value, and the threshold value are calculated for the entire customer, not for each customer segment, and the action pattern is determined and the customer attribute is estimated based on the parameter. May be processed. Further, as the “average speed” here, the average speed in the period excluding the “freeze” state described later is also calculated.

  The space acquisition unit 36 acquires space information that represents the space in the store as a position in the virtual space. In the space information, coordinates indicating the entire area of the store and coordinates for specifying the arrangement of a plurality of sales floors contained therein are defined in the virtual space. The positions of the first access point 12, the second access point 14, and the third access point 16 are also defined in the spatial information. The route acquisition unit 34 generates route information that indicates the customer's movement route in the space in the store by the position in the virtual space based on the position information, the time information, and the space information. The route acquisition unit 34 generates route information together with information regarding which of the plurality of sales floor areas the customer has passed through. In this way, the route information is generated in units of sales floor areas, and if the route information for one shopping cart is connected in time series, all routes from the start of use to the end of use can be recognized.

  The behavior pattern data storage unit 44 stores a plurality of behavior pattern data in which different behavior patterns are defined as types of purchasing behavior based on the relationship between the travel route and the travel speed of the customer. The type of purchasing behavior defined in a plurality of behavior pattern data is, for example, “passing”, which is the first pattern of passing through the sales floor area without showing the behavior of the purchase consideration such as stop or sudden deceleration. There are a second pattern, "active", in which a person stops for a short time and puts a product in a shopping cart, or at least considers purchasing a product, and a third pattern, "freeze", in which one product area stays for a long time. In each of the plurality of behavior pattern data, the average staying time, the average traveling speed, the presence/absence of a cart stop, and the presence/absence of a sudden deceleration are previously determined for each sales floor area for various purchasing behaviors. A plurality of action pattern data is determined for each sales floor area, and the pattern determination unit 42 calculates the average staying time and the average traffic speed in each sales floor area regularly or irregularly, and sets the average value as the average value. Based on this, the plurality of behavior pattern data stored in the behavior pattern data storage unit 44 is updated. Since the standard for distinguishing the behavior pattern differs depending on the product area and the layout for each sales floor area, the standard is frequently updated for each sales floor area to reflect the change in the product composition and the layout for the distinction of the behavior pattern. The pattern determination unit 42 stores in the determination result storage unit 46 the determination result as to whether the customer's behavior matches passing, active, or freeze, in association with the sales floor area. The pattern determination unit 42 stores, in the determination result storage unit 46, a combination or permutation of sales floor areas in which a customer's action matches passing, a combination or permutation of sales floor areas in which active customer coincides, and a combination or permutation of sales floor areas in which freeze is matched. Store.

  The pattern determination unit 42 determines, based on the route information, the speed information, and the plurality of behavior pattern data, which of the behavior pattern types the customer's purchasing behavior matches. The pattern determination unit 42 acquires POS data as data indicating purchase details from the POS server 19 via the communication unit 22 and uses the POS data as a determination material for the purchase behavior of the customer. The POS data includes information about the customer's purchase such as the product name or product ID, the number of purchases, the purchase price, the purchase date and time, the customer ID, the settlement area ID, and the information about the customer himself. The customer ID is identification information given to identify the customer, and is registered in, for example, a membership card, and the ID is input via a cash register at the time of payment. The customer ID is not included in the POS data regarding the customer who does not own the membership card. The settlement area ID is identification information given in advance for each cash register in order to distinguish a plurality of cash registers. The position coordinates of the plurality of settlement areas in which each cash register is installed are defined in the spatial information, and each is associated with the settlement area ID in advance. The pattern determination unit 42 associates the POS data with the route information based on the purchase date/time, the settlement area ID, and the route information. In other words, the route information records which settlement area was passed and when, and refers to the purchase date and time and the settlement area ID included in the POS data to associate the POS data with the route information.

  The pattern determination unit 42 extracts the customer ID from the POS data and acquires the customer attribute from the POS server 19 based on the customer ID. As a modified example, the POS data may include the customer attribute data itself. The customer attribute is an element that can distinguish the type of customer, such as gender and age group, and can affect the behavior of the customer. The pattern determination unit 42 stores the action pattern determined to match in the action pattern data storage unit 44 in a form associated with the type of customer attribute. Then, the pattern determination unit 42 estimates the attribute of the customer whose attribute is not known by referring to the correspondence between the action pattern and the attribute stored in the action pattern data storage unit 44. The pattern determination unit 42 estimates the purchase behavior of the customer based on the information regarding which of the plurality of sales floors the customer has passed through and the type of the matching behavior pattern. Details of the determination method and the estimation method by the pattern determination unit 42 will be described later. The determination result and the estimation result of the pattern determination unit 42 are stored in the determination result storage unit 46.

  The output processing unit 48 outputs the determination result stored in the determination result storage unit 46 through the display unit 56 or the communication unit 22. That is, the output processing unit 48 displays the determination result on the screen of the display unit 56 and transmits the determination result to the outside via the communication unit 22. The product information storage unit 58 stores the information that can be provided to the customer regarding the product to be sold in association with the sales floor of the product and the type of action pattern. The information distribution unit 50 sends information that can be provided to the customer to the output processing unit 48 according to the type of the customer's action pattern determined by the pattern determination unit 42, and the output processing unit 48 sends the information to the customer's shopping cart. It is transmitted to the liquid crystal display device and the POS server 19 provided. The operation input unit 52 receives an operation by the administrator of the customer behavior analysis system 10, and the control unit 54 controls each unit of the customer behavior analysis device 20 based on the operation.

Hereinafter, a method of determining an action pattern and a method of estimating a purchase action by the pattern determining unit 42 will be described. Two examples will be given as the determination method of the action pattern. One is (A) a method to judge according to the time from entry to exit to the exit area and whether it has stopped or suddenly slowed down, and another is (B) entry to the exit area It is a method to judge only by the staying time from the entry to the exit.
In the case of (A), the following five patterns are defined as the plurality of action pattern data.
(1) Passing ・・・ Stay time from entry to exit area and exit is less than X seconds or there is no sudden deceleration (2) Active ・・・ Stay time from entry to exit area is X Seconds or more and less than Y seconds, or stay time from entry to exit to exit area is less than X seconds, or there is sudden deceleration (3) Freeze: Stay time from entry to exit to exit area Y seconds or more (4) Re-freeze: Repeat freeze in the same sales floor area (5) Re-passing: Repeat passing of the same sales floor area Here, the pattern determination unit 42 exits from entering the sales floor area. Calculate the staying time required for each sales floor area, and set the actual minimum value on that day or that time zone to XX [seconds] and the actual maximum value on that day or that time zone to YY [seconds] .. Here, the actual minimum value for the cart that purchased the product in the sales floor area, the actual minimum value for the cart that stopped or suddenly slowed down in the sales floor area, the stop or sudden slowdown in the sales floor area Then, the administrator selects one of the practical minimum value for the cart that purchased the item and the value set by the system administrator, and the selected value is XX [seconds] for the day or the time zone. ] In addition, the actual maximum value for a cart that purchased goods in the sales floor area, the actual maximum value for a cart that stopped or suddenly slowed down in the sales floor area, and stopped or suddenly slowed down in the sales floor area The administrator selects either the actual maximum value for the cart that purchased the item above, or the value set by the system administrator, and the selected value is YY [seconds] for the day or the time period. And The pattern determination unit 42 averages the substantial minimum value XX and the substantial maximum value YY over a plurality of days or a plurality of time zones, respectively, and obtains an average value X [second] and an average value Y [second]. These average values X and Y serve as a threshold value as a reference for distinguishing between passing, active, and freeze. Note that XX, YY, X, Y may be calculated for each day of the week instead of being calculated for each day or each time zone.

In the case of (B), the following five patterns are defined as the plurality of action pattern data.
(1) Passing ・・・ Stay time from entry to exit area is less than X seconds (2) Active ・・・ Stay time from entry to exit area is less than X seconds and less than Y seconds (3) Freeze: Staying time from entry to exit to exit is Y seconds or more (4) Re-freeze: Repeat freeze in the same store area (5) Re-passing: Repeat pass of the same store area Here, the pattern determination unit 42 calculates the staying time from the entry to the exit of the sales floor area to the exit of the sales floor area for each sales floor area, and sets the actual minimum value of that day or that time zone as XX [seconds], and the day or Let YY [seconds] be the actual maximum value in that time period. Here, the actual minimum value for the cart that purchased the product in the sales floor area, the actual minimum value for the cart that stopped or suddenly slowed down in the sales floor area, the stop or sudden slowdown in the sales floor area Then, the administrator selects one of the practical minimum value for the cart that purchased the item and the value set by the system administrator, and the selected value is XX [seconds] for the day or the time zone. ] In addition, the actual maximum value for a cart that purchased goods in the sales floor area, the actual maximum value for a cart that stopped or suddenly slowed down in the sales floor area, and stopped or suddenly slowed down in the sales floor area The administrator selects either the actual maximum value for the cart that purchased the item above, or the value set by the system administrator, and the selected value is YY [seconds] for the day or the time period. And The pattern determination unit 42 averages the substantial minimum value XX and the substantial maximum value YY over a plurality of days or a plurality of time zones, respectively, and obtains an average value X [second] and an average value Y [second]. These average values X and Y serve as a threshold value as a reference for distinguishing between passing, active, and freeze.

  The pattern determination unit 42 determines which of (1) to (5) the movement of the shopping cart matches for each sales floor area. In the case of (1), it is estimated that the product has simply passed the sales floor area, and in the case of (2), it is estimated that the product in the sales floor area has been purchased or at least the purchase has been considered. In the case of (3), it is presumed that the product in the sales floor area has a strong interest, but the product is not found or the purchase is unclear. In the case of (4), it is presumed that freezes are repeated and the purchase is lost. In the case of (5), it is presumed that the more the passing is repeated, the more lost and the product cannot be found. As described above, the pattern determination unit 42 can estimate the purchasing behavior of the customer if the type of the behavior pattern is known.

  The pattern determination unit 42 determines the traveling direction, the moving speed, and the presence/absence of a stop for each shopping cart based on the position information and the time information. The pattern determination unit 42 can determine whether the market area has entered in the forward direction or the reverse direction based on the traveling direction. The pattern determination unit 42 can also determine whether the customer is in a hurry based on the moving speed of the shopping cart. In addition, the pattern determination unit 42 can determine whether the customer is deciding the product or is shopping based on the speed of the shopping cart, the presence/absence of sudden deceleration, and the presence/absence of a stop.

  The pattern determination unit 42 estimates the attribute of the customer based on the behavior pattern or the purchasing behavior of the customer. For example, on Saturdays, Sundays, and holidays, a customer whose moving speed is slow and the average residence time in each sales floor area is relatively long is extracted, and the customer is estimated to be “with family”. Further, in a store that handles non-food items such as clothing and miscellaneous goods, customers who pass through most or all of the sales areas for men, women, and children are presumed to be "with family." In this way, the pattern determination unit 42 may make different determinations depending on the day of the week, the time of day, and the sales floor. The pattern determination unit 42 determines the number of passing or freezing areas, that is, the number of passing or freezing areas, for example, the number of selling or freezing areas for a moving object whose time exceeds a predetermined abnormal value criterion. A moving body having a ratio of higher than a predetermined standard may be determined to be a suspicious person. In addition, a moving body whose freeze time in a specific sales floor area is longer than a predetermined abnormal value or a mobile body whose freeze count is larger than a predetermined abnormal value may be determined to be a suspicious person. In addition, the flow line distance from entry to exit, entry time, stay time, average speed or average speed excluding the third pattern, total stop count, combination or permutation of sales floor areas determined to match passing, When at least one of the combination or permutation of sales floor areas determined to match active and the combination or permutation of sales floor areas determined to match freeze corresponds to a predetermined abnormal manner, the pattern determination unit 42 The moving body is determined to be a suspicious person. For example, a suspicious person is a moving body that has a very long stay in the store for several hours, or a moving body that has a long flow line distance in the store that far exceeds the normal time. May be determined. Alternatively, a moving body whose average speed is lower than a predetermined abnormal value or a moving body whose average speed is higher than a predetermined abnormal value may be determined to be a suspicious person. The output processing unit 48 may transmit the determination result of the suspicious individual to, for example, a terminal carried by a security guard.

  Next, the pattern determination unit 42 estimates factors of good or bad sales of the product by using the cumulative total of the behavior patterns of the customers and the determination result regarding the purchase behavior. For example, the pattern determination unit 42 can compare the product sales capabilities of each sales floor area from a plurality of cut points. The behavior of the customer in each sales floor area and the process in which the product was purchased were aggregated as a "performance index" for each sales floor area, and compared with past aggregations, or between different sales floors. Compare tally. The comparison result is stored in the output processing unit 48. For example, the sales and the number of purchases in each sales area, the area of the sales area, the number of passing carts for passing, the number of active carts for freezing, the number of freezing carts, the total staying time of carts, the total number of cart stopping, Numbers divided by density etc. are totaled and compared. The output processing unit 48 reads the comparison result from the determination result storage unit 46 and displays it on the display unit 56.

  The details of the analysis will be described below. First, the pattern determination unit 42 performs (1-1) purchase count, (1-2) purchase amount, (1-3) passing count, active count, and freeze count from entry to exit for each sales floor area. Probability of purchase as a ratio of the number of purchases, (1-4) Probability of purchase as a ratio of the number of purchases to the number of stoppages, (1-5) Average speed, (1-6) Stay time, (1-7) From entering a store Information on each of the flow line distance until the store is closed, (1-8) customer attribute, (1-9) number of stops, (1-10) passing ratio, active ratio, and freeze ratio is calculated. Here, the “passing ratio” is, for example, the ratio of the number of times of passing to the sum of the number of times of passing, the number of times of active, and the number of times of freeze. Similarly, the "active ratio" is, for example, the ratio of the active count to the sum of the passing count, the active count, and the freeze count, and the "freeze ratio" is the ratio to the sum of the passing count, the active count, and the freeze count. This is the ratio of the number of freezes.

  The pattern determination unit 42 collects each of the information items (1-1) to (1-10) for each shopping cart by all customers, and classifies or compares the following elements as cut points according to the user's request. The elements of the classification are, for example, (1-A) sales floor area, (1-B) time and day of the week, (1-C) average speed, (1-D) stay time, (1-E) flow line distance, (1 -F) Customer attributes, (1-G) stop count, (1-H) purchase amount, purchase count. The output processing unit 48 is a graph in which any one of (1-A) to (1-H) is on the x-axis and any one of (1-1) to (1-10) is on the y-axis. Is output to the display unit 56 or the like. The user selects the elements for the x and y axes.

  The pattern determination unit 42 also includes (2-1) the number of traffic, that is, the number of customers, (2-2) the number of purchases, (2-3) purchase amount, (2-4) passing ratio, and active ratio for each sales floor area. , Freeze ratio, (2-5) purchase probability as a ratio of the number of purchases to each of the passing number, active number, and freeze number, (2-6) purchase probability as a ratio of the number of purchases to the number of suspensions, (2-7 ) Average speed, (2-8) average stay time, (2-9) average flow line distance from entry to the sales floor area, (2-10) customer attribute, (2-11) average number of stoppages, ( 2-12) Area information of the sales floor area, (2-13) Shopping cart density in the sales floor area, and (2-14) Purchased information are calculated.

  The pattern determination unit 42 classifies or compares each of the information items (2-1) to (2-14) for each sales floor area with the following elements as cut points according to the user's request. The elements of classification are, for example, (2-A) passing, active, and frozen customer status, (2-B) sales floor area, (2-C) hours and days of the week, (2-D) exiting from a store. Average speed up to, (2-E) staying time from entry to exit, (2-F) flow line distance from entry to store area, (2-G) customer attribute, (2-H) The stop count, (2-I) purchase amount, purchase count, and (2-J) purchase product. The output processing unit 48 is a graph in which any one of (2-A) to (2-J) is on the x-axis and any one of (2-1) to (2-14) is on the y-axis. Is output to the display unit 56 or the like. The user selects the elements for the x and y axes.

  FIG. 2 schematically shows the relationship between the classification based on the state of customers and the number of purchasers. The pattern determination unit 42 determines the tendency of the product purchase, such as whether the customer often makes a prompt decision to purchase the product for each sales floor area or often purchases the product after careful consideration. Judge based on the number of freezes. As an example, the number of purchases of products, the number of passing times, the number of active times, and the number of freezes for each sales area are aggregated for all customers. In the figure, the number of prospective purchasers 130 is the sum of the number of active times and the number of freezes in a certain sales floor area, and is positioned as at least the number of customers who are interested and have entered products in the sales floor area. The actual purchaser number 132 is the number of customers who actually purchased the product in the sales floor area calculated based on the POS data. The number of customers 136, which is a portion where the circle of the expected purchaser 130 and the circle of the actual purchaser 132 overlap, is the number of customers who actually purchased among the customers who were interested in the sales floor area and entered. The number of customers 134 is the number of customers who are interested in the sales floor area and have entered the sales floor area but have not actually purchased it. When the number of customers is large, the number of customers is out of stock, poor quality, low product strength, and It can be inferred that a negative factor such as high price exists in the sales floor. On the other hand, the number of customers 138 is the number of customers who were in a state of passing that was not planned to be purchased at this place, and it can be estimated that there are positive factors such as impulsive buying and noticing the necessity of the product on the spot. In this way, the existence of negative factors and positive factors can be revealed for each sales floor area, and measures for reducing negative factors and increasing positive factors can be taken according to the results.

  The pattern determination unit 42 may calculate the following performance index. For example, 134 is the number of customers who have an interest in the sales floor but have not purchased, and 136 is the number of customers who have an interest in the sales floor and have made a purchase. When the number of customers 138 who have reached to c is set as c, and the number of customers 139 who are not interested in the sales floor and who did not purchase is set as d, b/(a+b) is connected to the purchase among the interested customers (a+b). Rate, and if this rate is high, the quality and price of the product can be highly evaluated. c/(b+c) is the rate of customers (b+c) who did not originally have an interest in the sales floor. If this rate is high, in-store advertisements, product arrangements, campaigns, etc. can be highly evaluated. c/(c+d) is the ratio of customers (c+d) who were not interested in the sales floor to purchase, and if this ratio is high, it is possible to highly evaluate the in-store advertisement, product arrangement, and the like. a/(a+d) is the rate of customers who were originally interested in the sales floor among the customers who did not purchase at that sales floor. If this rate is high, the potential customer needs of that sales floor are highly evaluated. be able to.

  The pattern determination unit 42 estimates the cause of a product that does not sell well. For example, the pattern determination unit 42 estimates that there is a problem in the shelving allocation or arrangement method when the number of times of passing is relatively large in the sales floor area where the unsold products are displayed, and the information indicating that is used as the determination result. It is stored in the determination result storage unit 46. For example, when the number of purchases is relatively small compared to the number of active times and the number of freezes, the pattern determination unit 42 estimates that there is a problem in the product lineup, product strength, and product price, and the determination result is information indicating that fact. It is stored in the storage unit 46. Instead of the purchase history such as sales or the number of purchases, the factor of good or bad of purchase may be determined based on the number of times the shopping cart is stopped in each sales floor area. Stopping the shopping cart includes not only when the product is purchased but also when the product is not actually purchased but at least when the purchase is considered, so useful data can be obtained as marketing information. Be done.

  The pattern determination unit 42 can also analyze the effect of a special sale campaign or leaflet performed on a specific product or a specific sales floor area. The quality of sales itself is known from the POS data, but it is difficult to determine the cause of the sales quality from the POS data. According to this embodiment, it is possible to estimate the factor of sales quality based on the behavior pattern of the customer. For example, the pattern determination unit 42 does not increase the number of visitors to the sales floor advertised by the campaign or the leaflet or the number of interested customers (a+b) before or after the campaign or the leaflet, or the value of c does not increase so much. If you don't, you can judge that the effect of the campaign or the leaflet was not so great. Also, if the number of visitors to the sales floor and the number of interested customers (a+b) have increased, but the purchase probability for the number of interested customers (a+b) has decreased and sales have not increased significantly, campaigns and leaflets Although there was an effect, it can be judged that sales did not increase due to problems with product strength and product price. As described above, the pattern determination unit 42 determines the effect of the campaign or the flyer based on the number of times of passing, the number of times of activation, the number of freezes in the sales floor area of the product for sale on the special sale item or the leaflet and the purchase number and the purchase amount for each number. And the factor can be determined.

  Next, the pattern determination unit 42 monitors the behavior of the shopping cart of the customer who is currently shopping in real time. Each performance index is displayed on the display unit 56 by the output processing unit 48. The administrator can grasp the performance index for each sales floor area in real time while looking at the screen of the display unit 56, find a sales floor with poor performance, and take immediate action. For example, when the display unit 56 indicates that the average value of the number of “active” carts in a sales floor area does not change from normal but the sales growth is worse than usual, the product in the sales floor area is not fresh. It can be inferred that there is a problem such as, or the product price is high.

  The information distribution unit 50 sends, to the output processing unit 48, information for introducing the product displayed in the sales floor area according to the sales floor area in which the shopping cart is currently located. The output processing unit 48 transmits the information received from the information distribution unit 50 to the shopping cart via the wireless LAN and the POS server 19 via the LAN 18. The information distribution unit 50 acquires the product information according to the type of the customer's action pattern from the product information storage unit 58 and sends it to the output processing unit 48. For example, you may send information that introduces the product as information to customers who are in a frozen state or information that is being passed, or you may send store guide map data as information to customers who are lost. Good.

  As another method of providing information introducing products to the customer, the pattern determination unit 42 refers to the POS data and the customer's purchase history to identify the customer's route information, speed information, and action pattern types and purchase contents. The information distribution unit 50 may determine the information content and the providing method based on the determination. The pattern determination unit 42 acquires the customer ID and the purchase details of the customer staying in the settlement area based on the POS data received from the POS server 19, and associates the customer ID with the route information of the customer as a purchase history in the determination result storage unit 46. Store. After the storage in the determination result storage unit 46, when the same customer visits the store again, the pattern determination unit 42 reads the purchase history of the customer from the determination result storage unit 46 based on the customer ID, and analyzes the behavior of the customer. .. The customer ID may be acquired, for example, in the form of being read from the membership card to the shopping cart at the start of shopping, or in the form of being read from the membership card in the settlement area at the time of settlement. In these cases, the information delivery unit 50 may send the information to the cash register in the checkout area and have it printed on the customer's receipt, or deliver it by e-mail if the customer's e-mail address is registered in advance. You may.

  The information to be printed on the receipt and the information to be delivered by e-mail are exemplified below. For example, when the pattern determination unit 42 determines that the frequency of visits to a specific sales floor area for a customer who relatively frequently visits the shop has suddenly decreased, it is estimated that a product in the sales floor area is being purchased at another store. be able to. In that case, the information distribution unit 50 distributes the data of the discount ticket for the product in the sales floor area or the data of the sales promotion advertisement for the products in the sales floor area to the customer. When the pattern determination unit 42 determines that a customer who stayed in the same sales floor area for a relatively long time, that is, a customer who is active or frozen does not purchase the products in the sales floor area, the products in the sales floor area or We can assume that we were not satisfied with the price. In that case, the information distribution unit 50 provides the customer with the data of the discount ticket for the product in the sales floor area or the data of the sales promotion advertisement for the products in the sales floor area. When the pattern determination unit 42 determines that the number of purchased items is small despite the fact that the flow line distance is relatively long or the stay time in the store is relatively long, the information distribution unit 50 determines that the discount coupons for all the products in the store are available. Provide the customer with data or promotional data for the entire store product. When the pattern determination unit 42 determines that the number of products purchased in a specific sales floor area is relatively small even though the stay time is relatively long, the information distribution unit 50 causes the information distribution unit 50 to discount the products in the sales floor area. Providing the customer with the data of the ticket or the data of the promotion advertisement regarding the product in the sales area. When the pattern determination unit 42 determines that the customer has a short flow line distance since the membership card is issued shallowly, the information distribution unit 50 provides the customer with discount voucher data regarding products of the entire store or Provides sales promotion data for all store products. Customers who are active or frozen in a special sales floor where a sales promotion campaign is held are presumed to be customers who are interested in the theme of the campaign. For such a customer, the information distribution unit 50 provides the data of the discount ticket regarding the target product of the similar campaign or the data of the sales promotion advertisement regarding the product.

  The information distribution unit 50 provides, as the information to be distributed by e-mail, for example to a customer who has purchased a product targeted for the campaign during a specific sales promotion campaign, information regarding the sales promotion campaign of other products or related products. You may. As a method of displaying information on the liquid crystal display device of the shopping cart, the information distribution unit 50 immediately provides sales promotion information regarding products in the sales floor area when a predetermined time has passed in the same sales floor area. Alternatively, information about sales promotion regarding products in the sales floor area may be provided to a customer who passes through the same sales floor area a predetermined number of times or more. The information for sales promotion to be sent to the customer may be information on the products in the sales floor where the customer did not pass or the products in the sales floor that the customer passed by.

  FIG. 3 shows the appearance of the shopping cart. The shopping cart 60 is provided with a cart liquid crystal display device 62. The cart liquid crystal display device 62 has a built-in wireless LAN communication function, and wirelessly communicates with the first access point 12, the second access point 14, and the third access point 16 in the store through the cart antenna 64. The cart liquid crystal display device 62 acquires the strength of the radio wave transmitted/received to/from the first access point 12, the second access point 14, and the third access point 16. The cart liquid crystal display device 62 has a liquid crystal display unit, and causes the liquid crystal display unit to display the information about the product received from the customer behavior analysis device 20.

  FIG. 4 schematically shows the sales floor arrangement in a store in a virtual space defined by the space information. Inside this store, there is a sales floor for products classified as "vegetables," "fruits," "meat," "fresh fish," "sweets and dried foods," "retorts and seasonings," "dairy products," and "beverages." The area through which the customer can pass is composed of an entrance area 70, a first sales floor area 72 to an eleventh sales floor area 92, a first settlement area 94 to a sixth settlement area 104, and an exit area 106, and the arrangement and coordinates of each of these areas. Is defined in the spatial information.

  The entrance area 70 is an area that a customer passes through when starting shopping at this store, and a shopping cart is prepared in this area in advance and serves as a movement start point. The first sales floor area 72 and the second sales floor area 74 are areas adjacent to the entrance area 70, the first sales floor area 72 is associated with the “vegetables” sales floor, and the second sales floor area 74 and the fourth sales floor area 78 are It is associated with the "fruit" sales floor. The third sales floor area 76 is associated with the “meat” sales floor, and the fifth sales floor area 80 and the sixth sales floor area 82 are associated with the “confectionery/dry food” sales floor. The ninth sales floor area 88 is associated with the “fresh fish” sales floor, and the seventh sales floor area 84 and the eighth sales floor area 86 are associated with the “retort/seasoning” sales floor. The tenth sales floor area 90 is associated with the “dairy product” sales floor, and the eleventh sales floor area 92 is associated with the “beverage” sales floor.

  The first checkout area 94, the second checkout area 96, the third checkout area 98, the fourth checkout area 100, the fifth checkout area 102, and the sixth checkout area 104 are the first cash register, the second cash register, the third cash register, and the third cash register, respectively. It is associated with the fourth cash register, the fifth cash register, and the sixth cash register. For example, the route information about the shopping cart passing through the first checkout area 94 is associated with the POS data recorded by the first cash register at the time of the passing. The exit area 106 is an area through which the customer passes when he/she finishes shopping at this store, and is an end point of movement of the shopping cart. Note that in both the entrance area 70 and the exit area 106, it is considered as entering the store when passing in the direction from outside the store to the inside of the store, and as exiting when passing in the direction from the inside of the store to the outside of the store.

  FIG. 5 schematically shows a hierarchical sales floor area. Although FIG. 4 shows an example in which the sales floor areas are not hierarchized, in this figure, one sales floor area is subdivided into a plurality of sub sales floor areas, and the relationships between the sales floor areas are hierarchized. For example, when the sales floor area 120 is a clothing sales floor and is divided into a plurality of brands, the sales floor area 120 that is the “clothing floor” is defined as “A brand”, “B brand”, “C brand”, “D brand”. It is divided into a plurality of sub-sales areas 122, 124, 126, 128 called “brand”. The pattern determination unit 42 not only counts the behavior patterns for the sales floor area 120 but also counts the behavior patterns for the plurality of sub sales floor areas 122, 124, 126, and 128, so that the customer's preference and degree of interest can be further calculated. It can be estimated in detail.

  On the other hand, the relationship between the sales floor areas may be hierarchized in the same manner as in the example of FIG. 5 by setting a partition that is a higher area so as to include a plurality of sales floor areas. In this case, not only the judgment result of the behavior pattern in each sales floor area but also the judgment result of the behavior pattern in the upper hierarchy is obtained. You can get the result.

  FIG. 6 schematically shows an example of the movement route of the shopping cart in the sales floor area. The sales floor area 110 in this figure is a horizontally long quadrilateral area, and the shopping cart enters the sales floor area 110 from the left side, moves back and forth several times inside the sales floor area 110, and once outside the sales floor area 110 from the right side. Come out. After that, a movement path is shown in which the right side of the sales floor area 110 enters the sales floor area 110 again, and after a while, the right side edge of the sales floor area 110 leaves the sales floor area 110.

  Depending on the sales floor area, the correct traveling direction may be determined in advance depending on the layout, and the route information generated by the route acquisition unit 34 indicates whether the shopping cart has entered in the forward direction or the reverse direction. It is recorded in correspondence with the sales floor area. In addition, after entering an area that does not become active or frozen in the sales area that passes through from the time you enter the store until you reach a certain sales area, and after becoming active or frozen in that area, you also enter the sales area once. When the user goes to the settlement area without becoming active or frozen, the determination result by the pattern determination unit 42 is recorded that the shopping cart "goes straight" to the sales area.

  Next, the contents of each table generated by the pattern determination unit 42 will be described. First, the pattern determination unit 42 records, for each shopping cart, position information and time information indicating the behavior of the customer from entering the store to leaving the store as a position attribute table. In the position attribute table, as parameters relating to the whole, entrance date and time, entrance course [forward course, reverse course, direct], settlement method [yes/no], exit course [forward course, reverse course], checkout cash register number, Settlement date and time, traffic line distance that is the total traveled distance from entering to leaving the store are recorded.Average speed as a parameter related to speed, average speed during the period excluding the frozen state, maximum speed, and cart stop as a parameter related to whether or not to stop The number of times, average cart stop time and maximum cart stop time are recorded. The pattern determination unit 42 determines that the movement of the shopping cart is stopped when the movement of the shopping cart is within a predetermined radius range. For example, when the movement range from a certain positioning point is continuously within the range of the predetermined radius, the pattern determination unit 42 regards this as “stopped”. In addition, when the state of staying within the range of the predetermined radius from the previous positioning point measured, for example, one second ago continues for a predetermined time or more, the pattern determination unit 42 regards this as “stopped”. You may.

  In the position attribute table, the number of times of passing, the average time of passing, the shortest time of passing, the longest time of passing, and the number of passing sales floor areas are further recorded as parameters relating to passing.

  In the position attribute table, the number of active times, the average active time, the shortest active time, the longest active time, and the number of active sales floor areas are further recorded as parameters relating to active. In the position attribute table, the number of freezes, the average freeze time, the shortest freeze time, the longest freeze time, and the number of freeze sales floor areas are further recorded as parameters relating to freeze.

  The pattern determination unit 42 records, for each shopping cart, a history of actions in the passing sales floor area as a cart action history table. In the cart behavior history table, sales floor area name, entry date and time, exit date and time, status name (passing, active, freeze), flow distance from entry to the sales floor, average speed, maximum speed, number of stops, average stop time, The maximum stop time, the number of purchased products, and the total purchase price are recorded. When a certain shopping cart travels through the same sales floor area a plurality of times, a separate cart action history record is generated each time the shopping cart travels.

  The pattern determination unit 42 records the state in each sales floor area as a sales floor area position attribute table for each sales floor area and for each time zone. The sales floor area position attribute table contains the number of passing states, the average time of the passing states, the shortest time of the passing states, the longest time of the passing states, the average number of stopping times of the passing states, and the average of the passing states in one selling area and time zone. The purchase amount and the average number of purchases in the passing state are recorded. The sales floor area position attribute table includes the number of active states, average active time, minimum active time, maximum active time, average active stop count, and average active status in one sales floor area and time zone. Purchase amount, average number of active purchases, number of freezes, average time of freezes, shortest time of freezes, longest time of freezes, average number of stops in freezes, average amount of purchases in freezes, average of freezes The average number of purchases is also recorded. In addition, for each customer who passes through the sales floor area for each time zone, the average distance, minimum movement distance, and maximum movement distance from entry to the sales floor area, and for all customers who purchased products in that sales floor area Average distance from entry to exit, minimum distance to flow, and maximum distance to flow, average distance from entry to exit for all customers who stopped in that sales floor area, minimum flow path The distance and the maximum flow line distance are also recorded.

  In the sales floor area position attribute table, the average travel time, the shortest travel time, the longest travel time of customers who purchased products in the sales floor area, the average travel time, the shortest travel time, and the longest travel time of customers who stopped in the sales floor area. Further recorded are the time, the average travel time, the shortest travel time, and the longest travel time when the product was stopped and purchased in the sales area. The XX, YY for that day or time zone and the X, Y that apply to that day or time zone are also recorded.

The pattern determination unit 42 can estimate each of the following items by matching the purchase history based on the POS data with the behavior pattern of the customer.
(1) The pattern determination unit 42 calculates the number of purchases per stop for each sales floor area from the formula (the number of purchases/the number of stoppages). Since it is considered that there is a strong correlation between the stop of the shopping cart and the purchase/non-purchase of products, if the number of purchases per stop is relatively small, it can be presumed that there is a problem in the product lineup or product strength.
(2) The pattern determination unit 42 calculates the number of purchases per shopping cart passage (passing, active, freeze) for each sales floor area. If the number of purchases per pass is relatively small, factors such as inappropriate product placement, low attractiveness of products, and inappropriate advertising content or placement of advertisements in the sales floor are estimated. )
(3) The pattern determination unit 42 calculates the number of purchases and the purchase amount per shopping cart that is “passing” for each sales floor area. If the number of purchases and the amount of purchase are large, it can be inferred that attractive products are arranged easily.
(4) The pattern determination unit 42 calculates the number of purchases and the purchase amount per “active” shopping cart for each sales floor area. If the number of purchases and the amount of purchase are low, factors such as improper product placement, low product appeal, and high product prices are estimated.
(5) The pattern determination unit 42 calculates the number of purchases and the purchase price per “freezing” shopping cart for each sales floor area. If the number of purchases and the amount of purchase are low, it is estimated that the product may have missed the opportunity to sell due to improper product placement or out of stock despite high demand.
(6) The pattern determination unit 42 calculates the number of purchases and the amount of purchase per shopping cart that is “re-freeze” for each sales floor area. If the number of purchases and the amount of purchase are low, it is estimated that there are problems such as improper product placement and out of stock.
(7) The pattern determination unit 42 calculates the number of shopping carts that have left the store without purchasing anything because they are “active” even though they are “direct”. In this case, it is presumed that the product was out of stock.
(8) The pattern determination unit 42 calculates the number of shopping carts that are “direct” and “freeze” but have left the shop without purchasing anything. In this case, it is estimated that the product lineup was insufficient.
(9) The pattern determination unit 42 calculates the number of shopping carts that have left the store without moving after moving in the “forward direction”. In this case, problems such as product lineup, product quality, and product arrangement in the entire store are estimated.
(10) When the pattern determination unit 42 once “freezes” in a certain sales floor area and then enters the same sales floor area again and becomes “active” or “freeze”, if no product is purchased, It is presumed that the customer either returned the item or did not make the purchase decision.

  Next, the pattern determination unit 42 estimates the attribute and state of the customer based on the behavior pattern and the purchase behavior of the customer. Specifically, (1) day of the week and time zone [weekdays, holidays, 11:00 to 17:00, etc.], (2) entrance course [forward course, reverse course, direct], (3) whether or not payment is [Yes, None], (4) Withdrawal course [Forward course, reverse course], (5) Store stay time [High, Middle, Low], (6) Average speed or average speed during periods excluding the frozen state [High, Middle] , Low], (7) Cart stop frequency [High, Middle, Low], (8) Cart stop average time [High, Middle, Low], (9) Passing frequency, (10) Active frequency [High, Middle, Low] ], (11) freeze frequency [High, Middle, Low], (12) re-freeze probability [High, Middle, Low], (13) re-passing probability [High, Middle, Low], (14) active or freeze. [15) At least 6 of food, confectionery, meat, fresh fish, vegetables, dairy products, dry foods, seasonings are included in the combination or permutation of the sales floor area, (15) Distance from entry to exit Conditions are defined in advance as attribute pattern data for each item such as total, and stored in the action pattern data storage unit 44. The pattern determination unit 42 estimates the customer's attribute as follows with reference to the attribute pattern data. Here, in (2) entrance course and (4) exit course, it is determined whether or not the envisioned entrance course and exit course have been followed in advance. (12) The re-freeze probability is calculated by (number of freezes)/(number of frozen store areas). (13) The re-passing probability is obtained by (number of passing times)/(number of sales floor areas that have been passed). Here, the “passing frequency” is the number of times of passing per hour, and is the following: passing frequency×in-store stay time=passing number of times.

  For example, in the case of the attribute "shopping with family on a holiday", (1) is "holiday, 11:00 to 17:00", (2) is "forward course", (3) is "payment", and (4) is "Forward course", "High" as (5), "Low, Middle" as (6), "High" as (7), "High" as (10), "High" as (11), (12) As “Middle, High”, (13) As “Middle, High”, (14) As “at least one of food, confectionery, meat, fresh fish, vegetables, dairy products, dried products, seasonings, etc. 6 areas are included”, and (15) is defined as “High”. When the behavior pattern or the purchase behavior of the customer meets the above conditions, the pattern determination unit 42 estimates that the attribute of the customer is “shopping with family on holiday”. Note that “High”, “Middle”, and “Low” correspond to, for example, the case where the distribution of customers is divided into three.

(Example 2)
In this embodiment, between a customer who is a mobile body and a plurality of wireless LAN base stations, the time from when one transmits a radio wave to when the return radio wave is received from the other is measured, and each base is measured. Positioning is based on the difference in the length of time measured at the station. This point is different from the first embodiment in which the position of the moving body is detected by using the difference in the radio field intensity at each base station. Hereinafter, the differences from the first embodiment will be mainly described, and the description of the common points will be omitted.

  FIG. 7 is a functional block diagram illustrating the configuration of the customer behavior analysis system according to the second embodiment. The customer behavior analysis system 10 is different from the customer behavior analysis system 10 according to the first embodiment including the strength acquisition unit 24 in that the customer behavior analysis system 10 includes the time value acquisition unit 140. First, a plurality of wireless LAN base stations within the wireless communication range of the communication device provided in the shopping cart, for example, the first access point 12, the second access point 14, and the third access point 16 transmit a wireless signal and then The time until the wireless signal as a reply is received from the communication device is measured, and the time value is transmitted to the customer behavior analysis system 10. The wireless communication system used in this embodiment may be UWB (Ultra Wide Band). The time value acquisition unit 140 acquires the time value of the time required for transmission/reception with the shopping cart from each base station. The position recognition unit 26 calculates the distance between each base station and the shopping cart from the time value, calculates the position of the shopping cart based on the distances from at least three base stations, and generates position information. Time information indicating the calculated time is generated. The position detection method may be a TOA (Time of Arrival) method or a TDOA (Time Difference of Arrival) method. The position information and the time information are stored in the position information storage unit 28. The position information and time information of the shopping cart can be detected by the above method as in the first embodiment, and the same analysis as in the first embodiment can be performed based on the detected position information and time information. It is also possible to detect which shopping cart is used by detecting the ID of the wireless LAN communication device provided in the customer's shopping cart.

(Example 3)
In this embodiment, a plurality of reader/writers are used as a wireless communication device in which an IC tag is embedded in a shopping cart, a shopping cart, a membership card, a mobile phone, a mobile terminal, or the like carried by a customer and which can wirelessly communicate with the IC tag. Be done. The position of the customer is detected by information indicating which reader/writer the IC tag wirelessly communicates with. For example, a large number of reader/writers are installed in a store, each has its own communication area, and by detecting which reader/writer communication area the IC tag held by the customer is passing through, the customer's position can be determined. judge. This point is different from the first embodiment in which the position of the moving body is detected by using the difference in the radio field intensity at each base station. Hereinafter, the differences from the first embodiment will be mainly described, and the description of the common points will be omitted.

  FIG. 8 is a functional block diagram illustrating the configuration of the customer behavior analysis system according to the third embodiment. The first reader/writer 142, the second reader/writer 144, and the third reader/writer 146, when a customer who has an IC tag in their respective communication areas passes, sends reception information indicating that fact to the customer behavior analysis system 10. To do. The reception information acquisition unit 148 acquires reception information from the first reader/writer 142, the second reader/writer 144, and the third reader/writer 146. The position recognizing unit 26 generates position information indicating the position of the customer who owns the IC tag depending on from which reader/writer the received information is acquired, and also generates time information indicating the timing, which is stored in the position information storage unit 28. Store. With the above method, the customer position information and time information can be detected as in the first embodiment, and the same analysis as in the first embodiment can be performed based on the detected position information and time information. It is also possible to detect which shopping cart or the like is used by detecting the ID of the IC tag embedded in the belongings such as the shopping cart of the customer.

(Example 4)
The customer behavior analysis system 10 according to the present exemplary embodiment is connected to the first camera 152, the second camera 154, and the third camera 156, and includes a reception information acquisition unit 148. In this embodiment, a plurality of cameras continuously capture a plurality of areas included in a target space from a predetermined angle such as a bird's-eye view, and an object that moves due to color changes in a plurality of obtained continuous images. To detect the position of the object. This point is different from the first embodiment in which the position of the moving body is detected by using the difference in the radio field intensity at each base station. Hereinafter, the differences from the first embodiment will be mainly described, and the description of the common points will be omitted.

  FIG. 9 is a functional block diagram of the configuration of the customer behavior analysis system according to the fourth embodiment. At least one camera is installed in each store area in the store so that the camera can shoot from a predetermined angle such as a bird's-eye view. The first camera 152, the second camera 154, and the third camera 156 are examples of cameras that can take images from a bird's-eye view. The first camera 152, the second camera 154, and the third camera 156 continuously shoot the sales floor area, which is each shooting area, from a bird's eye view. A plurality of continuous images taken by the first camera 152, the second camera 154, and the third camera 156 are transmitted to the customer behavior analysis system 10 at any time. The image acquisition unit 150 acquires continuous images from the first camera 152, the second camera 154, and the third camera 156. The position recognition unit 26 detects a moving body from continuous images, recognizes the position of the moving body, generates position information, generates time information indicating the time, and stores the time information in the position information storage unit 28. Store. With the above method, the customer position information and time information can be detected as in the first embodiment, and the same analysis as in the first embodiment can be performed based on the detected position information and time information.

  FIG. 10 illustrates an image of a certain sales floor area photographed from a bird's eye view. If the sales floor image 180 of this figure is two-dimensionally captured, the image of the customer, who is a moving body, appears at an angle except when the customer is directly under the camera. As a result, while the customer directly under the camera does not see the parts other than the head and shoulders, the customer who appears obliquely shows the whole body as well as the head and shoulders. Not only is the appearance completely different depending on the customer's position, but the head is closer to the camera than the feet, which causes a large image, making it difficult to accurately grasp the customer's position. Here, according to the conventional position detection technology, first, a square frame 183 with which the moving body 182 is inscribed is set on the sales floor image 180, and the frame center point 184 of the square frame 183 is regarded as the position of the moving body 182. .. However, as shown in the figure, the actual position of the moving body 182 should originally be the foot position 186, and the frame center point 184 is displaced from the foot position 186.

  FIG. 11 schematically shows a shooting situation of the sales floor image 180 in FIG. 10 from the side. A camera 204 is installed on the ceiling 200, and the camera 204 shoots in the direction of the floor surface 202. The moving body 182 walking on the floor surface 202 is moving from the image center point 188 toward the end, and when the camera 204 takes an image of the figure, the moving body 182 is photographed at an angle rather than vertical as shown in the figure. Therefore, assuming that the floor surface 202 is the sales floor image 180, the moving body 182 is obliquely projected onto the area 206. Therefore, the width of the moving body 182 is larger than that when the moving body 182 is vertically photographed right above the moving body 182, and the center point of the width is far from the actual position of the moving body 182. In addition, the closer to the camera 204, the larger the image, so that the head of the customer, which is the moving body 182, appears larger than the feet.

  FIG. 12 is a diagram in which a cylinder model is applied to the sales floor image in FIG. In the sales floor image 180 of this figure, the moving body 182 shown in the sales floor image 180 of FIG. 10 is replaced with a columnar body 192. Hereinafter, the processing of the position recognition unit 26 that detects the position of the moving body by replacing the image of the moving body with the cylindrical object will be described. The position recognition unit 26 sets a cylinder axis 194 that is a straight line connecting the image center point 188 and the frame center point 184 of the rectangular frame 183, and sets a cylinder body 192 having the cylinder axis 194 as an axis. The bottom surface 196 of the cylindrical body 192 is set to have a radius in the real world of, for example, 50 cm. Since the appearance and size of the cylindrical body 192 change depending on the distance, it is calculated that the bottom surface 196 is drawn with an ellipse having a major axis radius corresponding to 50 cm in the sales floor image 180. The radius of the ellipse can be set to any value. The bottom surface 196 is set at such a position that the center point 190 of the bottom surface is placed on the cylindrical shaft 194 and the circumferential line of the bottom surface 196 abuts the square frame 183. The upper surface 198 of the cylindrical body 192 is drawn as an ellipse whose size is such that its center point is placed on the cylindrical axis 194 and its circumferential line abuts the square frame 183. As shown in FIG. 11, since the head of the moving body 182 appears larger than the feet, the columnar body 192 is set such that the top surface 198 is larger than the bottom surface 196 as shown in FIG. The bottom center point 190 is regarded as the foot position of the moving body 182. As a result, the position recognition unit 26 can acquire more accurate position information even when the moving body is imaged obliquely. Although the process of detecting the position of the foot without performing image correction has been described in this drawing, even if the position of the foot is detected after the size difference due to the aberration or perspective of the sales floor image 180 is corrected in advance. Good. In that case, in particular, the sales floor image 180 may be corrected so that the bottom surface 196 changes from an ellipse to a perfect circle. Since the method of correcting the aberration is known, the description thereof will be omitted.

  The position recognition unit 26 assigns an ID to the detected moving body and recognizes the position information and the position detection time in association with the ID. The position recognizing unit 26 continues to give the same ID to the moving body as long as the same moving body is continuously recognized from the images in the same sales floor area. The initially recognized position information is the position coordinates of the frame center point 184 of the rectangular frame 183 shown in FIG. 10, and the coordinates are only relative coordinates in the sales floor image 180. A straight line connecting the image center point 188 and the frame center point 184 is assumed as a cylinder axis 194. A columnar body 192 having a bottom surface 196 corresponding to a radius of 50 cm on the cylinder axis 194 and having a radius of 50 cm is located closer to the image center point 188 than the frame center point 184 and having an upper surface 198 on the opposite side and contacting the square frame 183. To set. The relative coordinates of the bottom surface center point 190 of the bottom surface 196 at this time are detected as position information of the moving body.

  The position recognition unit 26 detects a moving object in the image for each camera. Even if the same moving body is assigned a different ID for each camera, the position recognition unit 26 determines the correspondence relationship of the IDs for the same moving body based on the arrangement relationship of the images from each camera. Accordingly, the position recognition unit 26 recognizes the flow line of each moving body in the store.

  Here, when the photographing ranges of adjacent cameras partially overlap with each other, a moving body that has entered the overlapping region may be displayed in two images at the same time. In that case, the number of IDs generated from the overlapping area is a multiple of 2. All the IDs shown in the overlapping area are divided into pairs, the distance between the moving bodies in the pair is calculated, and the distances of the respective pairs are summed. Then, the combination of pairs is changed and the distances of each pair are summed in the same manner. A combination having the smallest total distance is assumed as an optimal combination among all possible combinations of such pairs, and the moving bodies of each pair are determined to be the same moving body.

  FIG. 13 schematically shows a method of detecting the same moving body in overlapping shooting ranges. The first image 262 and the second image 264 have overlapping regions 260, respectively, because the shooting ranges partially overlap. In the example of this figure, one moving object is shown in the area other than the overlapping area 260 of the first image 262, whereas four moving objects 252 to 258 are shown in the overlapping area 260. Since the overlapping area 260 is a portion where the shooting ranges of the two images overlap, the number of moving objects shown in this area is a multiple of 2. Therefore, the second moving body 252, the third moving body 254, the fourth moving body 256, and the fifth moving body 258 appearing in the overlapping area 260 are actually recognized as four by duplicating two moving bodies. It's just that.

  Here, the position recognition unit 26 divides the moving body shown in the overlapping area 260 into a plurality of pairs, and determines how to divide the pair that has the smallest value when the total value of the distances is calculated for each pair. In the example of this figure, the second moving body 252, the third moving body 254, the fourth moving body 256, and the fifth moving body 258 are arranged in this order at substantially equal intervals. Therefore, when the pair is divided into a pair of the second moving body 252 and the third moving body 254 and a pair of the fourth moving body 256 and the fifth moving body 258, the total value of the distances becomes the minimum. In addition, since it is known in advance that the moving bodies appearing in the same image are different moving bodies, the second moving body 252 and the fourth moving body 256 are not paired in the figure, and the third moving body is not shown. Neither 254 nor the fifth moving body 258 is paired.

  As a supplementary determination method, the position recognition unit 26 determines whether the speed from the position of the first moving body 250 to the positions of the second to fifth moving bodies 252 to 258 or the moving direction and the speed to the position of the first moving body 250. , The difference in any of the moving directions is calculated, and the method of dividing the pair having the smallest total value when the total value of the differences is calculated for a plurality of pairs is further determined. Further, the position recognizing unit 26 calculates a total value of values obtained by arbitrarily combining some of the parameters such as the distance between the moving bodies, the speed of the moving bodies, and the moving direction, and the total value is minimized. You may judge how to divide a pair.

  It should be noted that the comparison of the moving objects by the above calculation has been described on the assumption that the images in which the shooting ranges partially overlap are processed, but the position recognition unit 26 performs the above calculation and comparison without overlapping the shooting ranges. The images of adjacent cameras may be processed together. In that case, the position recognizing unit 26 recognizes the moving direction of the moving body from the obtained flow line of the moving body, and recognizes the correspondence of the ID of the moving body between the images in the shooting ranges adjacent to the moving direction. With this, even if the cameras are slightly separated from each other, the correspondence relationship between the moving bodies can be maintained, and the number of cameras to be installed can be reduced and the cost can be kept low. Further, even when the positional relationship of a plurality of moving bodies is switched between images, it is possible to accurately recognize the corresponding moving body due to the difference in speed and moving direction.

  FIG. 14 schematically shows an example of the shooting range of each camera in the store. Here, the photographing ranges at three locations are illustrated. The first shooting range 320 and the second shooting range 322 are shooting ranges corresponding to the first image 262 and the second image 264 of FIG. 13. The first shooting range 320 and the second shooting range 322 have a positional relationship in which they are aligned with each other, and their coordinate axes are parallel to each other. On the other hand, the coordinate axes of the second shooting range 322 and the third shooting range 324 are not parallel, and the third shooting range 324 is inclined with respect to the second shooting range 322. Thus, even if the coordinate axes of the shooting ranges are not parallel to each other, the position recognition unit 26 converts the coordinates in the image of each shooting range into absolute coordinates that are the spatial coordinates of the entire store, so that each moving body in the store The position can be recognized.

(Example 5)
The customer behavior analysis system 10 according to the present exemplary embodiment is connected to the first ultrasonic acquisition device 270, the second ultrasonic acquisition device 272, and the third ultrasonic acquisition device 274, and includes a reception information acquisition unit 148. The first ultrasonic acquisition device 270, the second ultrasonic acquisition device 272, and the third ultrasonic acquisition device 274 acquire ultrasonic waves transmitted from a moving body in a plurality of areas included in the target space, and the ultrasonic waves The position of the moving body is detected by the reception angle of. It is premised that each moving body has an ultrasonic transmitter. For example, the mobile is a customer's shopping cart in a store, which shopping cart comprises an ultrasonic transmitter. As described above, the customer behavior analysis system 10 according to the present exemplary embodiment is different from the customer behavior analysis system 10 according to the first exemplary embodiment that detects the position of the moving body by using the difference in radio field intensity at each base station. Hereinafter, the differences from the first embodiment will be mainly described, and the description of the common points will be omitted.

  FIG. 15 is a functional block diagram of the configuration of the customer behavior analysis system according to the fifth embodiment. At least one ultrasonic acquisition device is installed at a high place such as a ceiling in each of a plurality of sales floor areas in the store. The first ultrasonic acquisition device 270, the second ultrasonic acquisition device 272, and the third ultrasonic acquisition device 274 in this figure are examples of the ultrasonic acquisition devices installed on the ceiling, respectively, and detect ultrasonic waves from the sales floor area. Receive continuously. The first ultrasonic acquisition device 270, the second ultrasonic acquisition device 272, and the third ultrasonic acquisition device 274 transmit the reception angle of the received ultrasonic wave to the reception information acquisition unit 148. The position recognizing unit 26 detects the moving body from the reception angle of the ultrasonic wave, recognizes the position of the moving body and generates position information, generates time information indicating the time, and stores each position information. The data is stored in the unit 28. A method of detecting the reception angle of the ultrasonic wave and a method of detecting the position of the moving body based on the reception angle are known and will not be described.

(Example 6)
In the present embodiment, a geomagnetic sensor and an acceleration sensor are mounted on a customer's shopping cart, which is a moving body, and a flow line from the vicinity of the entrance of a store can be recognized by tracking the geomagnetic sensor and the acceleration sensor. The direction information which is the detection result of the traveling direction by the geomagnetic sensor and the acceleration information which is the detection result of the acceleration by the acceleration sensor are transmitted to the customer behavior analysis system 10 by the wireless communication device mounted on the shopping cart. The customer behavior analysis system 10 recognizes the flow line of the customer traced from the entrance of the store based on the direction information and the acceleration information received from the shopping cart. The customer behavior analysis system 10 does not need to constantly receive the direction information and the acceleration information from the shopping cart, and all the direction information and the acceleration information indicating the flow line from the entrance to the payment area when the shopping cart reaches the payment area. Are received together. Therefore, at least one base station in the store that receives information from the shopping cart by wireless communication is sufficient. As described above, the customer behavior analysis system 10 according to the present exemplary embodiment detects the information regarding the position by the sensor on the mobile body side, and thus the position of the mobile body is detected by using the difference in the radio field intensity at each base station. The customer behavior analysis system 10 of FIG. Hereinafter, the differences from the first embodiment will be mainly described, and the description of the common points will be omitted.

  Correction points may be provided at a plurality of locations in the store. An IC tag is installed at each correction point. When the mobile body arrives within the communication range of each IC tag, the position recognition unit 26 rewrites the position information of the mobile body at that time point with a predetermined value indicating the position of the correction point. As a result, the base point of position recognition based on the direction information and the acceleration information is irregularly corrected, so that position recognition with less error is realized.

  FIG. 16 is a functional block diagram of the configuration of the customer behavior analysis system according to the sixth embodiment. At least one base station is installed in the store. The access point 280 in the figure is an example of a base station installed in a store, receives direction information and acceleration information from the sales floor area or the settlement area, and transmits them to the reception information acquisition unit 148. Based on the direction information and the acceleration information, the position recognizing unit 26 recognizes the moving line of the moving body based on the position of the store entrance, which is the detection start point, to generate the position information, and the time information indicating the time. Are stored in the position information storage unit 28. The position recognition unit 26 may further store the direction information and the acceleration information received from the reception information acquisition unit 148 in the position information storage unit 28. The speed acquisition unit 32 may integrate the acceleration information stored in the position information storage unit 28 to calculate the speed of the moving body.

(Example 7)
In the present embodiment, a specific method of determining an action pattern by the pattern determination unit 42 will be described, and the description common to the first to sixth embodiments will be omitted.

  FIG. 17 shows the relationship between the staying time and the number of customers in the sales floor area. The relationship between the staying time and the number of customers in the sales floor area shown in the figure is only an example, but it is assumed that the relationship is roughly as shown in the figure. The horizontal axis represents the customer's staying time in a certain sales floor area, and the vertical axis represents the number of customers. The number of staying customers 160 indicates the number of customers for each staying time in one day in a certain sales floor area, and the number of purchasing customers 162 indicates the number of customers who actually purchased products in the sales floor area among the number of customers for each staying time. .. As shown in the figure, the total number of customers is larger for customers who have a relatively short stay, and smaller for customers who have a longer stay. Although the number of customers who purchased the product is smaller as the stay time is shorter, the number of customers increases as the stay time becomes longer, and the customer number starts to decrease as the stay time becomes longer.

  The area of the first range 164 indicates the total number of customers who have not purchased the product in the sales floor area among the customers whose stay time is shorter than the first time value XX [seconds]. The area of the second range 166 indicates the total number of customers who have purchased the product in the sales floor area among the customers whose stay time is shorter than the first time value XX. The area of the third range 168 indicates the total number of customers who did not purchase the product in the sales floor area among the customers who stayed longer than the second time value YY [seconds]. The area of the fourth range 170 indicates the total number of customers who have purchased the product in the sales floor area among the customers who stay longer than the second time value YY.

  The first time value XX is a value such that the area ratio between the first range 164 and the second range 166 becomes a predetermined ratio. That is, the first time value XX is a value such that the ratio of the customers who did not purchase the product to the customers who purchased the product becomes a predetermined ratio among the customers whose stay time is shorter than the first time value XX. The area ratio between the first range 164 and the second range 166 is 10:1, for example, and is set to an optimum value based on experiments and verification.

  The second time value YY is a value such that the area ratio between the third range 168 and the fourth range 170 becomes a predetermined ratio. That is, the second time value YY is a value such that the ratio of the customers who did not purchase the product to the customers who purchased the product becomes a predetermined ratio among the customers who stay longer than that value. The area ratio between the third range 168 and the fourth range 170 is, for example, 1:1 and is set to an optimum value based on experiments and verification.

(Example 8)
The customer behavior analysis system 10 according to the present embodiment is common to the other embodiments in that any one of the position detection methods according to the first to sixth embodiments is adopted as a basic configuration, but the customer behavior analysis system 10 is not available to customers other than customers at facilities such as stores. It differs from the other examples in that the behavior of the employee is analyzed. For example, when adopting the position detection method of the first to third embodiments, each employee carries a wireless communication device or an IC tag and recognizes the position by communicating with a base station or a reader/writer on the customer behavior analysis system 10 side. To be done. When the position detecting method of the fourth embodiment is adopted, the position of the employee is recognized through the image taken by the camera. When the position detecting method of the fifth embodiment is adopted, the position of the employee is recognized by transmitting/receiving ultrasonic waves. When the position detecting method of the sixth embodiment is adopted, the position of the employee is recognized based on the detection results of the geomagnetic sensor and the acceleration sensor. In order to identify the employee, it is preferable to adopt the position detecting method of the first to third, fifth and sixth embodiments. Hereinafter, the differences from the first to sixth embodiments will be mainly described, and the description of the common points will be omitted.

  In this embodiment, the pattern determination unit 42 can calculate the values of the following parameters based on the route information and display them on the screen. For example, the pattern determination unit 42, for each employee, entry time, exit time, flow line distance from entry to exit, number of stops from entry to exit, stop from entry to exit. Calculate time. The pattern determination unit 42 can also calculate the stay time, stop count, and stop time for each sales floor area for each employee. The values of these parameters are displayed on the screen through the operation of the administrator.

The pattern determination unit 42 “waits”, “out of the store”, “approach from customer”, “approach from employee”, “approach from employee”, and “business negotiation” based on the position, distance, speed, direction, etc. between the employee and the customer. It is determined whether it is “payment” or “move out of the sales area”. The pattern determination unit 42 calculates the time of the determined state for each employee, each month, each day of the week, each day, and each time zone. “Waiting” refers to the state of staying in the sales floor area in charge. “Outside the sales floor” refers to a state in which “customer approach”, “employee approach”, “business negotiation”, and “settlement” are excluded from the state of being outside the sales floor area in charge. “Customer approach” refers to a state in which the customer approaches a predetermined approach distance, for example, within 60 cm, of an employee who is “waiting” or “outside the sales floor”, and the distance continues for a predetermined time. The “employee approach” refers to a state in which an “waiting” or “outside the store” employee approaches a predetermined approach distance such as within 60 cm from a customer, and the approach distance continues for a predetermined time. The pattern determination unit 42 may determine which of the customer and the employee has approached by comparing the calculated values based on the motion vectors of the both. For example, let the employee vector be V _a and the customer vector be V _b . Assuming a straight line connecting the position of the employee and the position of the customer, the angle between the straight line and the vector V _a is θ _a, and the angle between the straight line and the vector V _b is θ _b . If V _a cos θ _a is greater than V _b cos θ _b , it is determined that the employee has approached the customer. Even if V _a cos θ _a and V _b cos θ _b are equal, it is determined that the employee has approached the customer, and V _a cos θ _a is If it is smaller than V _b cos θ _b, it is determined that the customer approaches the employee. It should be noted that V _a cos θ _a and V _b cos θ _b have positive values when they approach each other. The pattern determination unit 42 determines that transition to the state of “business negotiation” has been made if the approach of the employee and customer continues even after a predetermined time has passed, and if not, the state of “waiting” or “outside the sales floor” It is determined that the process has moved to. After shifting to the "business negotiation" state, at least one of the employee and the customer moves to the settlement area, and when it is recognized from the POS data that the settlement has been made at that time, "business negotiation" at that time Is completed and it is determined to be in the “payment” state. In addition, the pattern determination unit 42 determines that the “business talk” is completed when the employee and the customer in the “business talk” state are separated by a predetermined distance or more and the state has passed a predetermined time, and “wait” or “wait” or It is determined that the state has changed to "out of the store". Therefore, the employee in the “business negotiation” state only shifts to the “payment”, “waiting”, and “out of store” states next. After the state of “payment”, the pattern determination unit 42 determines that the “payment” has ended if the distance between the customer and the employee is more than a predetermined distance and the state continues for a predetermined time.

  The pattern determination unit 42 determines, for each employee, the number of times, the minimum time, the maximum time, and the average of the employee's action state such as “wait”, “approach from customer”, “approach from employee”, “business negotiation”, and “payment”. Calculate the time and the percentage of time in each state in all working hours. The pattern determination unit 42 calculates a contract probability for each employee for the number of negotiations. The number of negotiations is recognized from the route information of the employee and the relationship such as the position, distance, speed, direction between the employee and the customer, and the number of contracts is calculated from the route information of the employee or the customer and the POS data of the customer with whom the negotiation is made. recognize. The pattern determination unit 42 determines the staying time in each sales floor area for each customer by (1) when no employees are present, (2) when all employees are in service, and (3) one employee who is not in service. When the above exists, the calculation is performed separately for each of the three patterns. Each of (1) to (3) is totaled by active or frozen customers who pass through the sales area. For each sales floor area, the pattern determination unit 42 allows the employee to respond to the sales floor area with the number of customers whose stay time in the sales floor area corresponds to (1) or (2) and the total value of the stay time. It is calculated as the number and time that did not exist. This value is used as one indicator of loss of sales opportunity. As a result, it is possible to grasp the situation that the employee is absent or in service in the sales floor area forever.

  The customer behavior analysis system 10 can be applied to management of the entire store. The pattern determination unit 42, for example, the number of visitors, average staying time from entry to exit, sales, average speed or average speed excluding freezes, average flow distance, average number of stops, passing ratio, active ratio, freeze. Ratio, customer density in the settlement area, longest waiting time in the settlement area, average waiting time in the settlement area, active purchase probability and average purchase amount, purchasing probability and average purchase amount in passing, freeze purchase probability and average purchase Parameters such as the amount of money can be calculated for each store. Parameters such as the minimum value, average value, and maximum value of the ratio of the number of customers and the number of employees per unit area may be calculated for each store.

  FIG. 18 shows an example of a screen that displays a list of the length of time for each behavior state of employees. The screen 400 corresponds to each of the employee's action states such as “waiting”, “outside the sales floor”, “customer approach”, “employee approach”, “business negotiation”, “payment”, and “moving outside the sales floor area”. The length of time spent is indicated by the length of the bar graph. The output processing unit 48 can also display the time length of each of these states, for example, for each employee.

  In addition to the display contents of this figure, the output processing unit 48 causes the display unit 56 to display the time that the employee stays in each sales floor area on a monthly, weekly, weekday, daily, or time zone basis. .. Based on the operation of the administrator, the output processing unit 48 displays, on the display unit 56, a table in which the abscissa indicates the stay time for each sales floor area and the vertical axis indicates the type of sales floor area.

  The present invention has been described above based on the embodiments. It should be understood by those skilled in the art that this embodiment is merely an example, and that various modifications can be made to the combination of each constituent element and each processing process, and that such modifications are also within the scope of the present invention. Hereinafter, modified examples will be described.

  In the embodiment, the shopping cart used by the customer as the "moving body" is mainly illustrated. The “moving body” in the modification may be a shopping cart used by a customer, a communication device such as a mobile phone or a mobile terminal owned by the customer or an employee, an IC card or an IC chip. It may be a medium with a communication function such as an attached ticket. In the case of an IC tag such as an IC card or an IC chip, it may be an active IC tag having a built-in power supply or a passive IC tag supplied with power from the outside.

  In the embodiment, a system for calculating the position and speed of a customer or an employee who is a moving body in the space of the store in order to analyze the behavior thereof has been illustrated. In a modification, the system may be configured to analyze the behavior of a visitor who moves in a space in an entertainment facility such as a shopping district, an amusement park, a theme park, or a zoo. Further, it may be configured as a system that analyzes the movement of the player in an exercise facility such as a baseball field, a soccer field, or a rugby field.

  In the third embodiment, an IC tag is embedded in a shopping cart, a shopping cart, a membership card, a mobile phone, a mobile terminal, or the like carried by a customer, and a large number of reader/writers are installed in the store to communicate with each other. explained. In a modified example, a reader/writer is built in a shopping cart, a shopping cart, a mobile phone or the like carried by a customer, a large number of IC tags are installed in a store, and the position of the customer is detected by communicating with each other. It may be configured.

  The image pickup by the camera described in the fourth embodiment may be an image pickup using a CCD sensor or a CMOS sensor, or a special image pickup using a sensor such as ultrasonic waves, microwaves, infrared rays, or electromagnetic waves. In the fifth embodiment, the method of detecting the position by the angle of the radio wave in the communication between the ultrasonic transmitter and the receiver has been described. However, according to the time required for transmitting and receiving the ultrasonic wave between the mobile body and the receiver. It may be configured to detect the position of the moving body.

  According to the present invention, useful information can be provided by analyzing the behavior of a moving body in a space.

Claims (34)

A position acquisition unit that acquires position information indicating the position of the moving body that changes with time in a predetermined real space and time information indicating the positioning timing,
A space acquisition unit for acquiring space information representing the real space as a position on a virtual space,
A route acquisition unit that generates route information indicating a moving route of the moving body in the real space as a position on the virtual space based on the position information, the time information, and the space information;
A speed acquisition unit that generates speed information indicating the moving speed of the moving body based on the position information and the time information;
A pattern storage unit that stores a plurality of action pattern data in which different action patterns are defined as types of actions based on the relationship between the moving route and the moving speed of the moving body,
Based on the route information, the speed information, and the plurality of action pattern data, a pattern determination unit that determines which of the types of the action pattern the action of the mobile body is,
An output processing unit that outputs the result of the determination,
An action analysis device comprising: The space acquisition unit acquires, as the space information, respective coordinates for identifying the arrangement of a plurality of regions included in the real space on the virtual space,
The route acquisition unit generates route information indicating a moving route of the moving body in the real space at a position in the virtual space, together with information on which of the plurality of regions the moving body has traveled. ,
The speed acquisition unit, as the speed information, at least one of an average speed of the moving body in all of the plurality of areas, an average speed of the moving body in each of the plurality of areas, and an instantaneous speed of the moving body. To generate information that
The pattern determination unit determines the matching of the action patterns based on the route information including information regarding which of the plurality of areas the mobile body has traveled, the speed information, and a plurality of action pattern data. The behavior analysis device according to claim 1, wherein The position acquisition unit acquires position information indicating the position of the moving body in a space in a predetermined store as the real space and time information thereof,
The space acquisition unit acquires, as the space information, respective coordinates for specifying the arrangement of a plurality of sales floors included in the store on the virtual space,
The route acquisition unit, route information indicating a moving route of the moving body in the space in the store in a position on the virtual space, together with information about which of the plurality of sales floors the moving body has passed through. Generate,
The speed acquisition unit, as the speed information, an average speed of the moving body in all of the plurality of sales floor areas, an average speed of the moving body in each of the plurality of sales floor areas, and an instantaneous speed of the moving body Generate information indicating at least one of
The pattern determination unit determines the matching of the action patterns based on the route information including information regarding which one of the plurality of sales floors the mobile has traveled, the speed information, and a plurality of action pattern data. The behavior analysis apparatus according to claim 1 or 2, wherein Strength to acquire strength information indicating strength of a radio wave transmitted or received by a communication device carried by the mobile body via at least one of a plurality of radio communication devices installed at different positions in the real space. The acquisition part,
A position recognition unit that generates the position information and time information indicating the position of the moving body based on the intensity information,
The behavior analysis device according to claim 1, further comprising:   Between at least one of a plurality of wireless communication devices installed at different positions in the real space and a wireless communication medium carried by the moving body, one transmits a wireless wave and then receives the wireless wave. It further comprises a position recognition unit that generates the position information and the time information indicating the position of the moving body based on the length of time until the one receives a radio wave as a reply from the other. Item 5. The behavior analysis device according to any one of items 1 to 3.   Whether the IC tag or the wireless communication device carried by the moving body is communicating with which of a plurality of wireless communication devices or a plurality of IC tags installed at different positions in the real space. The behavior analysis device according to claim 1, further comprising a position recognition unit that generates the position information and the time information indicating the position of the moving body based on the position recognition unit. An image acquisition unit that acquires an image captured from a predetermined angle in the real space,
Position information and time information of the moving body by identifying an object that is moving between the plurality of images acquired by the image acquiring unit as the moving body and recognizing the position of the object in the space within the image. A position recognition unit that generates
The behavior analysis device according to claim 1, further comprising:   The position recognizing unit assumes that the moving body is replaced with a cylindrical body, and that a cylindrical body object in which the moving body is inscribed exists at the position of the moving body with its bottom surface in contact with the ground. The behavior analysis device according to claim 7, wherein the center coordinates of the bottom surface and the coordinate acquisition time thereof are set in the position information and time information of the moving body.   The position recognition unit is recognized from a plurality of images when one or more moving objects are recognized in each of a plurality of images captured at the same timing or in front of or behind the corresponding capturing range. It is possible to combine at least one of a distance, a speed difference, and a direction difference that are determined based on the position coordinates of a plurality of moving bodies, and determine which moving bodies are the same moving body according to the result of the combination. The behavioral analysis device according to claim 7, which is characterized in that. An ultrasonic acquisition device for receiving ultrasonic waves from the position of the moving body,
A position recognition unit that generates position information and time information of the mobile unit based on ultrasonic waves received from the position of the mobile unit,
The behavior analysis device according to claim 1, further comprising: A receiving device that receives, by wireless communication, direction information indicating the direction of movement of the moving body detected by a geomagnetic sensor and acceleration information indicating the acceleration of movement of the moving body detected by an acceleration sensor.
Based on the direction information and the acceleration information, a position recognition unit that generates the position information and the time information of the moving body based on the position of the start point of detection by the geomagnetic sensor and the acceleration sensor,
The behavior analysis device according to claim 1, further comprising:   The pattern determination unit acquires data indicating purchase details of the mobile customer, and estimates a factor of good or bad of purchase based on the data indicating the purchase details and the type of the matching action pattern. The behavior analysis device according to any one of claims 3 to 11.   The pattern determination unit determines the position of the settlement area where the customer stayed for purchasing the product and the date and time of the product purchase based on the route information, and corresponds to the position of the settlement region and the date and time of the product purchase. 13. The behavior analysis device according to claim 12, wherein data indicating the purchase content is acquired, and the acquired data and the route information are associated with each other.   The pattern determination unit is an attribute in which conditions of the route information, the speed information, and the type of the matching action pattern and the route information, the speed information, and the type of the action pattern are defined for each customer attribute. 14. The behavior analysis apparatus according to claim 3, wherein the attribute of the customer who is the moving body is estimated based on pattern data.   The pattern determination unit acquires an attribute of a customer who is the mobile body, stores the determined behavior pattern in the pattern storage unit in advance as the attribute pattern data associated with the attribute, and stores the pattern storage. 15. The behavior analysis apparatus according to claim 14, wherein the attribute of the customer whose attribute is not known is estimated by referring to the attribute pattern data stored in the section. In the plurality of behavior pattern data stored in the pattern storage unit, a first pattern showing a case of passing through the sales floor area as a type of the purchase behavior of the mobile customer, a relatively short stop in the sales floor area Of the stay time, presence/absence of stoppage, and acceleration/deceleration for each sales floor area for each of the second pattern showing the case of purchasing or considering purchasing the product and the third pattern showing the case of staying in one sales floor area for a long time. Standards for at least one are set in advance,
The pattern determination unit determines which one of the first pattern, the second pattern, and the third pattern the behavior of the customer who is the moving body in the sales floor area to be determined matches. The behavior analysis device according to any one of claims 3 to 15.   The pattern determination unit determines the determination result as to which of the first pattern, the second pattern, and the third pattern the behavior of the customer, who is the moving body, in the sales floor area to be determined matches the determination target. The behavior analysis apparatus according to claim 16, wherein the behavior analysis apparatus stores the behavior analysis area in a predetermined storage area in association with the sales floor area.   The pattern determination unit matches the combination or permutation of sales floor areas in which the behavior of the customer, which is the mobile object, matches the first pattern, the combination or permutation of sales floor areas in accordance with the second pattern, and the third pattern. The behavior analysis device according to claim 16, wherein at least one of a combination and a permutation of sales floor areas is stored in a predetermined storage area. The pattern determination unit,
Every time the mobile body enters each of a plurality of sales floor areas, the stay time from the entry to the exit is calculated,
Based on the route information and the data indicating the purchase content, the mobile unit that purchased the product and the sales area of the purchased product are specified, and the staying time of the mobile unit in the sales area is set to the staying time when the product is purchased. Identified as,
For each of the plurality of sales floor areas, specify the actual minimum time and maximum time for each day or time of the stay time when purchasing the product,
A first mark for distinguishing the first pattern from the second pattern is calculated by averaging the substantial shortest time over a plurality of days or a plurality of time zones for each sales floor area. Stored in the pattern storage unit as a threshold value in a form corresponding to the sales floor area,
A second maximum value for distinguishing the second pattern from the third pattern is calculated by averaging the substantial maximum time over a plurality of days or multiple time zones for each sales floor area. Stored in the pattern storage unit as a threshold value,
When the staying time in a certain sales floor area of the mobile object for which the behavior pattern is to be determined is less than the first threshold value associated with the sales floor area, the behavior pattern of the mobile object is the first with respect to the sales floor area. It is judged that it matches the pattern,
If the staying time in a certain sales floor area of the moving object that is the target of the action pattern determination is equal to or more than the first threshold value and less than the second threshold value associated with the sales floor area, the sales floor It is determined that the action pattern of the mobile body in the area matches the second pattern,
When the staying time in a certain sales floor area of the mobile object for which the action pattern is to be determined is equal to or more than the second threshold value associated with the sales floor area, the behavior pattern of the mobile object is the first with respect to the sales floor area. 19. The behavior analysis device according to claim 16, wherein it is determined that the three patterns match. The pattern determination unit,
When the relationship between the staying time and the number of customers staying in that area is expressed for each sales area, the sum of the customers who purchased products in the sales area among the customers whose staying time is shorter than the first time value and the sales area A time value such that the ratio with the sum of customers who did not purchase products in the area becomes a predetermined ratio is specified for each day or each time zone as the first time value, and a plurality of days or An average value of the first time values over a plurality of time zones is calculated and stored as the first threshold value for distinguishing the first pattern and the second pattern in the pattern storage unit,
When the relationship between the staying time and the number of customers staying in the area is expressed for each sales floor area, the sum of the customers who purchased the products in the sales floor area among the customers who stayed longer than the second time value and the sales floor. A time value such that the ratio with the total number of customers who did not purchase products in the area becomes a predetermined ratio is specified as the second time value for each day or each time zone, and a plurality of days for each sales area or An average value of the second time values over a plurality of time zones is calculated and stored as the second threshold value for distinguishing the second pattern and the third pattern in the pattern storage unit,
When the staying time in a certain sales floor area of the mobile object for which the behavior pattern is to be determined is less than the first threshold value associated with the sales floor area, the behavior pattern of the mobile object is the first with respect to the sales floor area. It is judged that it matches the pattern,
If the staying time in a certain sales floor area of the moving object that is the target of the action pattern determination is equal to or more than the first threshold value and less than the second threshold value associated with the sales floor area, the sales floor It is determined that the action pattern of the mobile body in the area matches the second pattern,
When the staying time in a certain sales floor area of the mobile object for which the action pattern is to be determined is equal to or more than the second threshold value associated with the sales floor area, the behavior pattern of the mobile object is the first with respect to the sales floor area. 19. The behavior analysis device according to claim 16, wherein it is determined that the three patterns match. The pattern determination unit,
A plurality of customers, which are the moving bodies, according to at least any one of a flow line distance from entry to exit in the entire space in the store, entry time, stay time, average speed, total number of stops Classifying into segments, calculating the first threshold value and the second threshold value for each of the segments, and storing them in the pattern storage unit,
Regarding the moving body that is the determination target of the action pattern, which of the plurality of segments is applicable, the flow line distance from the store entry to the store in the entire space in the store, the entry time, the stay time, the average speed , And the first pattern and the second threshold based on the first threshold value and the second threshold value corresponding to the corresponding segment The behavior analysis device according to claim 19 or 20, wherein it is determined whether the pattern or the third pattern matches.   The pattern determination unit positions that the customer, who is the moving body, does not have an interest in the sales floor when the behavior of the customer matches the first pattern, and the pattern determination unit when the behavior matches the second pattern and the third pattern. By classifying customers by positioning that they were interested in the sales floor, calculating the ratio of product purchases in each, the quality of the product according to the relationship between the interest in the sales floor and the ratio of product purchases, 22. A factor for purchasing or not purchasing a product including at least one of a price of a product, an advertisement of a product, a placement of a product, a method of promoting a product, and a potential demand is estimated. The behavior analysis device described in Crab.   The pattern determination unit purchases a product for any number of the first pattern, the second pattern, and the third pattern for each of the sales floor areas as a factor of quality of purchase by the customer who is the mobile body. 23. The behavior analysis device according to claim 16, wherein the number of times or purchase amount, or the number of times of the first pattern, the second pattern, and the third pattern is calculated.   The pattern determination unit determines the number of times of the first pattern or the third pattern, the ratio of the first pattern or the third pattern, the time of the first pattern or the third pattern, the flow line distance from the entry to the exit, the entry time. , Staying time, average speed or average speed excluding the third pattern, total number of stops, combination or permutation of sales floor areas determined to match the first pattern, sales floor area determined to match the second pattern Of the combination or permutation, and at least one of the combinations or permutations of the sales floor areas determined to match the third pattern is determined to be a suspicious person. The behavior analysis device according to claim 16. A product information storage unit that stores information that can be provided to a customer regarding a product to be sold in association with the sales floor of the product and the type of the action pattern,
At least one of a communication device, a customer terminal, a store terminal, and a store direct mail system, which the customer holds information that can be provided to the customer who is the mobile according to the type of the matching action pattern. An information distribution unit that sends to either
25. The behavior analysis device according to claim 3, further comprising:   The information delivery unit is determined to match the flow line distance from entry to exit, entry time, stay time, average speed or average speed excluding the third pattern, total stop count, and the first pattern. Combination or permutation of sales floor areas, combination or permutation of sales floor areas determined to match the second pattern, combination or permutation of sales floor areas determined to match the third pattern, customer identification information, and purchase history Of the communication device held by the customer, the terminal on the customer side, the terminal on the store side corresponding to the customer, and the direct mail system on the store side with respect to the customer in which at least one of the aspects corresponds to a predetermined criterion, The behavior analysis apparatus according to claim 25, wherein the information that can be provided is transmitted to at least one of them.   The information distribution unit determines that the customer, who is the mobile body, does not purchase the product in the sales floor even though the behavior of the customer matches the second pattern or the third pattern. 27, the behavior analysis device according to claim 25 or 26, characterized in that information for sales promotion regarding a product in the sales floor is transmitted.   The information distribution unit refers to the route information of the customer who is the mobile body, the speed information, and the type of the behavior pattern, and refers to a sales floor area that the customer does not pass through or a sales floor area that matches the first pattern. 27. The behavior analysis device according to claim 25, which transmits information for sales promotion regarding a product. The plurality of behavior pattern data stored in the pattern storage unit includes at least one of a position, a distance to a customer, a relative speed with a customer, and a duration of a state regarding a behavior type of an employee who is the moving body. Standards are set in advance,
The pattern determination unit, the behavior of the employee is a mobile body, the position, the distance to the customer, the relative speed with the customer, according to any of the plurality of behavior pattern data according to the duration of the state 29. The behavior analysis apparatus according to claim 3, wherein it is determined whether or not to perform. The plurality of behavior pattern data stored in the pattern storage unit, as a type of behavior of the employee who is the mobile body, a state of waiting for a customer, a state of being outside the sales floor, a state of approaching a customer, The standard is defined for at least one of the following types: a state in which a customer is approaching, a state in which a business negotiation is held, a state in which settlement is performed, and a state in which the customer moves out of the sales floor area.
The pattern determination unit, the behavior of the mobile employee, a state of waiting for a customer, a state of being outside the sales floor, a state of approaching a customer, a state of approaching a customer, a state of having a business negotiation. 30. The behavior analysis apparatus according to claim 29, wherein it is determined which of the two types, a state in which the payment is made and a state in which the user is moving outside the sales floor area, is matched.   The pattern determination unit, the behavior of the mobile employee, a state of waiting for a customer, a state of being outside the sales floor, a state of approaching a customer, a state of approaching a customer, a state of having a business negotiation. 31. For at least any one of the following types: the state of payment, the state of moving to the outside of the sales floor area, the total time corresponding to the type is calculated for each employee. The behavior analysis device described.   When the employee is absent in the sales floor area, all the employees in the sales floor area are waiting for the customer staying time corresponding to the second pattern and the third pattern in each sales floor area. If there is one or more employees who are not serving customers in the sales floor area, it is divided in each case, and the total time and the ratio in each case are calculated. Item 32. The behavior analysis device according to any one of items 29 to 31.   When the employee is absent in the sales floor area, all the employees in the sales floor area are waiting for the customer staying time corresponding to the second pattern and the third pattern in each sales floor area. If there is one or more employees who are not serving customers in the sales floor area, the situation is divided in each case and all the staying time in the sales floor area is absent for each of the sales floor areas. Or if all of the employees were serving customers, the total number of customers and the total stay time should be calculated as the number and hours of employees who missed sales opportunities because they could not respond in the sales area. 32. The behavior analysis device according to claim 29, wherein A function of reading position information indicating the position of the moving body that changes with time in a predetermined real space and time information indicating the positioning timing from the storage unit,
A function of reading the space information representing the real space at a position on the virtual space from a storage unit;
A function of generating route information indicating a moving route of the moving body in the real space by a position on the virtual space based on the position information, the time information, and the space information,
A function of generating speed information indicating the moving speed of the moving body based on the position information and the time information,
A function of storing a plurality of action pattern data in which different action patterns are respectively defined as types of actions based on the relationship between the moving path and the moving speed of the moving body,
Based on the route information, the speed information, and the plurality of action pattern data, a function of determining which of the types of the action pattern the action of the mobile body is,
A function for outputting the result of the determination,
A computer program characterized by causing a computer to perform.

Images