CN112101865A - Waiting time obtaining method and device, computer equipment and readable storage medium - Google Patents

Abstract

The invention discloses a waiting time acquisition method, device, computer equipment and readable storage medium, which relate to the field of Internet technology, acquire environmental data and behavior data, calculate the reliability of a time window corresponding to the environmental data and behavior data, and then determine Waiting time, using multivariate data to calculate the waiting time, solves the problem that the coverage of the geofence based on a single signal is too large or too small, and improves the accuracy of the calculation. The method includes: acquiring multiple environmental data and multiple behavior data of the target distribution facility in the historical distribution process; mapping the multiple environmental data and multiple behavior data to multiple time windows of the historical distribution process, and calculating multiple times Multiple reliability of the window; according to the multiple reliability, determine the arrival time window and the departure time window in multiple time windows; take the time interval between the arrival time window and the departure time window as the waiting time.

Description

Method, apparatus, computer device and readable storage medium for obtaining waiting time

technical field

The present invention relates to the field of Internet technologies, and in particular, to a method, an apparatus, a computer device and a readable storage medium for obtaining a waiting time.

Background technique

With the continuous development of Internet technology, in most service industries, users have higher and higher requirements for experience, so that terminals can provide users with more and more services. For example, the food delivery industry has become an important part of people's daily life. In part, the food delivery industry allows people to enjoy consumption and services without leaving their homes. In today's food delivery industry, many platforms that provide food delivery services are equipped with delivery personnel. During peak hours of delivery, the pressure on merchants to deliver meals is soaring, and the delivery speed is much lower than that in leisure time. Wait for a meal. If the waiting time for a meal is long, it is easy to cause the delivery staff's delivery efficiency to drop, and even cause the delivery time to be overtime. In the end, it is very likely that the delivery staff and the merchant will lead to a series of contradictions due to accountability. Therefore, many platforms are currently in use. A series of methods will be used to judge whether the delivery staff has arrived at the store or the location of the customer, obtain the waiting time of the delivery staff in the store or wait for the customer, and use this waiting time as an important certificate for subsequent dispute resolution.

In the related art, when the waiting time of the delivery personnel is currently counted, the arrival time and the departure time of the delivery personnel need to be obtained, and the time interval between the two is used as the waiting time of the delivery personnel. Among them, the same method can be used to judge whether the delivery personnel have arrived and whether they have left. Taking judging whether the delivery personnel have arrived as an example, it can be realized through the GPS (Global Positioning System, global positioning system) equipped by the delivery personnel and the reporting of the delivery personnel. . The specific implementation logic is that if the delivery staff reports to the platform that they have arrived and the GPS shows that the distance between the delivery staff and the store or customer is less than the threshold, it is determined that the delivery staff has arrived, and the delivery staff's arrival time can continue to be output based on the current time. In the follow-up, it is judged whether the delivery person has left and output the departure time in the same way, so as to obtain the waiting time of the delivery person.

In the process of realizing the present invention, the inventor found that the related art has at least the following problems:

GPS will be affected by the signal. The accuracy of GPS is low indoors and in locations with poor signal, and it will keep beating. It is difficult to accurately judge the distance between the delivery staff and the store or customer. Moreover, the reporting of delivery personnel is a personal behavior, there is room for cheating, and delivery personnel may make false reports. Under the influence of two factors, the accuracy of the waiting time obtained by the platform is very low, and the waiting time is used as a basis for dispute resolution. Unconvincing, lack of effective informatization means has affected the intelligent process of catering services.

SUMMARY OF THE INVENTION

In view of this, the present invention provides a waiting time acquisition method, device, computer equipment and readable storage medium, the main purpose of which is to solve the problem that the accuracy of the waiting time obtained by the current platform is very low, and the waiting time is difficult to be used as a basis for resolving disputes. It is convincing that the lack of effective informatization means has affected the intelligent process of catering services.

According to the first aspect of the present invention, there is provided a method for obtaining a waiting time, the method comprising:

Obtain multiple environmental data and multiple behavior data of the target distribution facility in the historical distribution process;

Mapping the plurality of environmental data and the plurality of behavior data to a plurality of time windows of the historical distribution process, and calculating a plurality of confidence levels of the plurality of time windows, the plurality of confidence levels indicating the occurrence probability of environmental data and behavior data corresponding to the multiple time windows;

determining an arrival time window and a departure time window in the plurality of time windows according to the plurality of confidence levels;

The time interval between the arrival time window and the departure time window is taken as the waiting time.

In another embodiment, the acquiring a plurality of environmental data and a plurality of behavior data of the target distribution facility in the historical distribution process includes:

Extracting multiple terminal signals uploaded by the terminal equipment of the target distribution facility during the historical distribution process, the multiple terminal signals including at least inertial signals, Bluetooth signals, wireless signals, magnetic induction signals and cellular signals;

dividing the plurality of terminal signals into a plurality of signal groups based on the upload time of the plurality of terminal signals, and the upload time of at least one terminal signal included in each signal group in the plurality of signal groups is consistent;

Signal analysis is performed on the plurality of signal groups, respectively, to obtain the plurality of environmental data and the plurality of behavior data.

In another embodiment, the signal analysis is performed on the multiple signal groups to obtain the multiple environmental data and the multiple behavior data, including:

For each signal group in the plurality of signal groups, read a plurality of target signals included in the signal group;

Based on the plurality of target signals, determine the environment where the target distribution device is located at the upload time corresponding to the signal group, and obtain environmental data, where the environmental data is at least any one of an indoor environment and an outdoor environment;

According to the plurality of target signals, the use state of the terminal device at the upload time corresponding to the signal group and the motion state of the target distribution facility at the upload time corresponding to the signal group are determined, based on the use state and The motion state obtains behavior data, and the behavior data is at least any one of waiting behavior and delivery behavior;

Each signal group in the plurality of signal groups is respectively read, the environmental data and behavior data of each signal group are output, and the plurality of environmental data and the plurality of behavior data are obtained.

In another embodiment, the obtaining behavior data based on the usage state and the motion state includes:

When the use state indicates that the terminal device executes a user instruction and the movement state indicates that the target distribution facility is stationary or the movement speed of the movement state is less than a speed threshold, set the waiting behavior as the behavior data;

When the use state indicates that the terminal device is in standby and the movement state indicates that the delivery facility is moving or the movement speed of the movement state is greater than or equal to the speed threshold, set the delivery behavior as the behavior data .

In another embodiment, the mapping of the plurality of environmental data and the plurality of behavioral data to a plurality of time windows of the historical distribution process includes:

Based on the preset division duration, time division is performed on the historical distribution process to obtain the multiple time windows, and the duration of each time window in the multiple time windows is equal to the preset division duration;

For each of the plurality of environmental data and each of the plurality of behavioral data or each of the environmental data or each of the behavioral data, determining a target upload time corresponding to the each of the environmental data or each of the behavioral data;

Query the target time window where the target upload time is located in the multiple time windows, map each environmental data or each behavior data to the target time window, and complete the multiple environmental data and all mapping between the plurality of behavioral data and the plurality of time windows.

In another embodiment, the calculating multiple reliability levels of the multiple time windows includes:

Using the plurality of time windows, the time series model is trained on the plurality of environmental data and the plurality of behavior data to generate a distribution process time series model, and the distribution process time series model includes each of the plurality of time windows. The estimated state corresponding to the time window;

For each time window in the multiple time windows, obtain the target estimated state corresponding to the time window in the time sequence model of the distribution process;

When the target environmental data and target behavior data corresponding to the time window match the target estimated state, set the environmental reliability of the target environmental data and the behavioral reliability of the target behavior data as the first A default value, based on the reliability of the environment and the reliability of the behavior to form the reliability of the time window;

When any of the target environment data or target behavior data corresponding to the time window does not match the target estimated state, calculate the difference between the standard value and the second default value, and use the difference as the The credibility of the data whose target estimated state does not match, and the second default value is taken as the credibility of the data that matches the target estimated state to obtain the credibility of the time window;

When the target environment data or target behavior data corresponding to the time window do not match the target estimated state, set the environmental credibility of the target environment data and the behavior credibility of the target behavior data as a third default value, based on the reliability of the environment and the reliability of the behavior to form the reliability of the time window;

The reliability of the multiple time windows is calculated respectively to obtain the multiple reliability.

In another embodiment, the determining an arrival time window and a departure time window in the multiple time windows according to the multiple reliability levels includes:

According to the multiple reliability levels, a window label is set for each time window in the multiple time windows, and the window label is at least any one of an arrival label or a departure label;

Acquire a plurality of consecutive candidate time windows with the window label as the arrival label in the plurality of time windows, and use the plurality of candidate time windows to form a window queue, and the window queues in the plurality of time windows are The window labels of the previous time window and the next time window are the leaving labels;

The candidate time window ranked first in the window queue is used as the arrival time window, and the candidate time window ranked at the end of the window queue is used as the departure time window.

In another embodiment, the setting a window label for each of the multiple time windows according to the multiple reliability levels includes:

For each time window in the multiple time windows, obtain the target credibility corresponding to the time window, and compare the target environment credibility and target behavior credibility included in the target credibility;

If the reliability of the target environment and the reliability of the target behavior are consistent, setting a window label for the time window that matches the environmental data and behavior data of the time window;

If the credibility of the target environment and the credibility of the target behavior are inconsistent, extract the specified credibility from the credibility of the target environment and the credibility of the target behavior, and set the time window to be the same as that of the target behavior. The window label of the data matching corresponding to the specified credibility, the specified credibility is greater than another credibility except the specified credibility in the target environment credibility and the target behavior credibility .

In another embodiment, before setting a window label for each of the multiple time windows according to the multiple reliability levels, the method further includes:

extracting the credibility to be adjusted from the multiple credibility, the environment credibility and behavior credibility included in the credibility to be adjusted are both lower than the credibility threshold;

determining the to-be-adjusted time window corresponding to the to-be-adjusted reliability, and extracting the previous time window and the next time window of the to-be-adjusted time window;

determining a standard time window in the previous time window and the next time window, the environmental data and behavior data included in the standard time window are inconsistent with the environment data and behavior data included in the to-be-adjusted time window;

Assign the environmental data, behavior data and the reliability corresponding to the standard time window included in the standard time window to the to-be-adjusted time window to obtain the adjusted to-be-adjusted time window, and based on the adjusted Set the window label for the time window to be adjusted.

According to a second aspect of the present invention, there is provided a waiting time acquisition device, the device comprising:

The acquisition module is used to acquire multiple environmental data and multiple behavior data of the target distribution facility in the historical distribution process;

The calculation module is configured to map the plurality of environmental data and the plurality of behavior data to a plurality of time windows of the historical distribution process, and calculate a plurality of reliability of the plurality of time windows, and the plurality of The reliability indicates the occurrence probability of the environmental data and behavior data corresponding to the multiple time windows;

a determining module, configured to determine an arrival time window and a departure time window in the multiple time windows according to the multiple reliability degrees;

The determining module is further configured to use the time interval between the arrival time window and the departure time window as the waiting time.

In another embodiment, the obtaining module includes:

an extraction unit, configured to extract multiple terminal signals uploaded by the terminal equipment of the target distribution facility during the historical distribution process, the multiple terminal signals at least include inertial signals, bluetooth signals, wireless signals, magnetic induction signals and cellular signals Signal;

a dividing unit, configured to divide the multiple terminal signals into multiple signal groups based on the upload time of the multiple terminal signals, and the upload time of at least one terminal signal included in each signal group in the multiple signal groups consistent;

An analysis unit, configured to perform signal analysis on the multiple signal groups respectively to obtain the multiple environmental data and the multiple behavior data.

In another embodiment, the analyzing unit is configured to, for each signal group in the plurality of signal groups, read a plurality of target signals included in the signal group; and determine the target signal based on the plurality of target signals. The environment in which the target distribution device is located at the upload time corresponding to the signal group, obtains environmental data, and the environmental data is at least any one of an indoor environment and an outdoor environment; according to the multiple target signals, determine the The use state of the terminal device at the upload time corresponding to the signal group and the motion state of the target distribution facility at the upload time corresponding to the signal group, based on the use state and the motion state, to obtain behavior data, the The behavior data is at least any one of waiting behavior and delivery behavior; read each signal group in the multiple signal groups respectively, output the environment data and behavior data of each signal group, and obtain the multiple environments data and the plurality of behavioral data.

In another embodiment, the parsing unit is configured to, when the use state indicates that the terminal device executes a user instruction and the movement state indicates that the target distribution facility is stationary or the movement speed of the movement state is less than a speed threshold When the waiting behavior is set as the behavior data; when the use state indicates that the terminal device is in standby and the movement state indicates that the distribution facility is moving or the movement speed of the movement state is greater than or equal to the speed When the threshold is reached, the delivery behavior is set as the behavior data.

In another embodiment, the computing module includes:

A division unit, configured to perform time division on the historical distribution process based on a preset division duration to obtain the multiple time windows, and the duration of each time window in the multiple time windows is equal to the preset division duration ;

a determining unit, configured to determine, for each of the plurality of environmental data and each of the plurality of behavioral data or each of the behavioral data, a target upload time corresponding to the each of the environmental data or each of the behavioral data;

The mapping unit is configured to query the target time window where the target upload time is located in the multiple time windows, map each environmental data or each behavior data to the target time window, and complete the multiple time windows. environmental data and a mapping between the plurality of behavioral data and the plurality of time windows.

In another embodiment, the computing module includes:

A training unit, configured to use the plurality of time windows to train the time series model on the plurality of environmental data and the plurality of behavior data to generate a distribution process time series model, where the distribution process time series model includes the plurality of The estimated state corresponding to each time window in the time window;

an acquisition unit, for each time window in the multiple time windows, acquiring the target estimated state corresponding to the time window in the distribution process sequence model;

A setting unit for setting the environmental credibility of the target environment data and the behavior credibility of the target behavior data when the target environment data and the target behavior data corresponding to the time window are matched with the target estimated state The degree is set as the first default value, and the credibility of the time window is formed based on the environmental credibility and the behavior credibility;

The setting unit is also used to calculate the difference between the standard value and the second default value when any data in the target environment data or target behavior data corresponding to the time window does not match the target estimated state, and The difference is used as the reliability of the data that does not match the target estimated state, and the second default value is used as the reliability of the data matched with the target estimated state to obtain the time window. credibility;

The setting unit is further configured to set the environmental credibility of the target environment data and the target behavior when the target environment data or target behavior data corresponding to the time window do not match the target estimated state. The behavior credibility of the data is set to a third default value, and the credibility of the time window is formed based on the environmental credibility and the behavior credibility;

The setting unit is further configured to calculate the reliability of the multiple time windows respectively to obtain the multiple reliability.

In another embodiment, the determining module includes:

a setting unit, configured to set a window label for each time window in the plurality of time windows according to the plurality of reliability degrees, and the window label is at least any one of an arrival label or a departure label;

The obtaining unit is configured to obtain a plurality of consecutive candidate time windows whose window label is the arrival label in the plurality of time windows, and use the plurality of candidate time windows to form a window queue, and the window queue is in the plurality of time windows. The previous time window and the window label of the next time window in each time window are the leaving labels;

A determination unit, configured to use the candidate time window at the top of the window queue as the arrival time window, and the candidate time window at the end of the window queue as the departure time window.

In another embodiment, the setting unit is configured to, for each time window in the multiple time windows, obtain a target reliability corresponding to the time window, and set the target environment included in the target reliability The credibility and the credibility of the target behavior are compared; if the credibility of the target environment and the credibility of the target behavior are consistent, set the time window to match the environment data and behavior data of the time window window label; if the credibility of the target environment and the credibility of the target behavior are inconsistent, the specified credibility is extracted from the credibility of the target environment and the credibility of the target behavior, which is the time The window is set with a window label matching the data corresponding to the specified confidence level, and the specified confidence level is greater than the target environment confidence level and the target behavior confidence level except for the specified confidence level. a credibility.

In another embodiment, the determining module further includes:

a first extracting unit, configured to extract the credibility to be adjusted from the multiple credibility, and the credibility to be adjusted includes both the environmental credibility and the behavior credibility lower than the credibility threshold;

a second extraction unit, configured to determine the to-be-adjusted time window corresponding to the to-be-adjusted reliability, and to extract the previous time window and the next time window of the to-be-adjusted time window;

The determining unit is further configured to determine a standard time window in the previous time window and the next time window, where the environment data and behavior data included in the standard time window and the environment included in the to-be-adjusted time window inconsistent data and behavioral data;

an adjustment unit, configured to assign the environmental data, behavior data included in the standard time window, and the reliability corresponding to the standard time window to the time window to be adjusted, to obtain the adjusted time window to be adjusted, and A window label is set based on the adjusted time window to be adjusted.

According to a third aspect of the present invention, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in the first aspect when the processor executes the computer program.

According to a fourth aspect of the present invention, there is provided a readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the steps of the method described in the above-mentioned first aspect.

With the above technical solutions, the present invention provides a method, device, computer equipment and readable storage medium for obtaining waiting time. The present invention obtains the environmental data and behavior data of the target distribution facility in the historical distribution process, and converts the environmental data and the The behavior data is mapped to the time window of the historical distribution process, the reliability of the time window is calculated, and according to the reliability, the arrival time window and the departure time window are determined in the time window, and the time interval between the arrival time window and the departure time window is taken as Waiting time, using the multi-dimensional data of the delivery personnel in the delivery process to calculate the waiting time of the delivery personnel, solves the problem of too large or too small coverage of the geofence based on a single signal, and ensures that effective information methods are used to obtain the waiting time. Time, improve the accuracy of waiting time calculation, make the waiting time convincing, and promote the intelligent process of catering services.

The above description is only an overview of the technical solutions of the present invention, in order to be able to understand the technical means of the present invention more clearly, it can be implemented according to the content of the description, and in order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand , the following specific embodiments of the present invention are given.

Description of drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are for the purpose of illustrating preferred embodiments only and are not to be considered limiting of the invention. Also, the same components are denoted by the same reference numerals throughout the drawings. In the attached image:

FIG. 1 shows a schematic flowchart of a method for obtaining a waiting time provided by an embodiment of the present invention;

2A shows a schematic flowchart of a method for obtaining a waiting time provided by an embodiment of the present invention;

2B shows a schematic flowchart of a method for obtaining a waiting time provided by an embodiment of the present invention;

FIG. 3A shows a schematic structural diagram of an apparatus for obtaining a waiting time provided by an embodiment of the present invention;

FIG. 3B shows a schematic structural diagram of an apparatus for obtaining a waiting time provided by an embodiment of the present invention;

FIG. 3C shows a schematic structural diagram of an apparatus for obtaining a waiting time provided by an embodiment of the present invention;

FIG. 3D shows a schematic structural diagram of an apparatus for obtaining a waiting time provided by an embodiment of the present invention;

3E shows a schematic structural diagram of an apparatus for obtaining a waiting time provided by an embodiment of the present invention;

3F shows a schematic structural diagram of an apparatus for obtaining a waiting time provided by an embodiment of the present invention;

FIG. 4 shows a schematic diagram of an apparatus structure of a computer device provided by an embodiment of the present invention.

Detailed ways

Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention may be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present invention will be more thoroughly understood, and will fully convey the scope of the present invention to those skilled in the art.

An embodiment of the present invention provides a method for obtaining a waiting time, as shown in FIG. 1 , the method includes:

101. Acquire multiple environmental data and multiple behavior data of the target distribution facility in the historical distribution process.

102. Map multiple environmental data and multiple behavior data to multiple time windows of the historical distribution process, and calculate multiple reliability levels of the multiple time windows, where the multiple reliability levels indicate the environmental data corresponding to the multiple time windows. and the probability of occurrence of behavioral data.

103. Determine the arrival time window and the departure time window in multiple time windows according to the multiple reliability degrees.

104. Use the time interval between the arrival time window and the departure time window as the waiting time.

In the method provided by the embodiment of the present invention, the environmental data and behavior data of the target distribution facility in the historical distribution process are acquired, the environmental data and behavior data are mapped to the time window of the historical distribution process, the reliability of the time window is calculated, and the reliability of the time window is calculated according to Credibility, determine the arrival time window and the departure time window in the time window, take the time interval between the arrival time window and the departure time window as the waiting time, and use the multivariate data of the delivery personnel in the delivery process to calculate the waiting time of the delivery personnel, It solves the problem that the coverage of the geofence based on a single signal is too large or too small, ensures that the waiting time is obtained by effective information means, improves the accuracy of the calculation of the waiting time, makes the waiting time convincing, and promotes the catering industry. The intelligent process of service.

An embodiment of the present invention provides a method for obtaining a waiting time, as shown in FIG. 2A , the method includes:

201. Acquire multiple environmental data and multiple behavior data of the target distribution facility in the historical distribution process.

At present, there has always been a problem in the food delivery industry, that is, how to effectively judge whether the delivery personnel have waited at the store of the merchant or where the customer is during the process of product delivery, and for how long. During the peak hours of take-out, due to the increasing pressure of the merchants to serve meals, the speed of meal delivery is much lower than that in idle time, and there may even be a burst of orders, and the delivery staff will be forced to wait in the store. Since there is no effective method to judge whether the delivery personnel have been waiting, the platform cannot effectively obtain the waiting time of the delivery personnel, and the subsequent waiting time cannot be predicted based on the previous waiting time, nor can it be based on the waiting time. Provide effective advice to delivery personnel, therefore, this may cause a series of problems in the process of practical application. First of all, the delivery staff waiting in the store will reduce the efficiency of food delivery and even lead to overtime. Secondly, the long waiting time of the delivery personnel will cause conflicts between the delivery personnel and the merchants due to authentication, which requires the platform to conduct arbitration, but the platform lacks strong evidence of arbitration. Therefore, there are currently two methods to obtain the waiting time of the delivery personnel. One method is to determine whether the delivery personnel has arrived at the store through GPS and the reporting of the delivery personnel. The specific judgment logic is that if the delivery personnel report to the platform that they have arrived at the store, And the platform detects that the distance between the delivery staff and the store is less than the preset distance threshold based on GPS, it can determine that the delivery staff has arrived at the store, start timing, and use the same method to detect that the delivery staff leaves the store and stop timing, that is, The waiting time of the delivery person can be obtained. Another method is to determine when the delivery staff arrives and leaves the store through Bluetooth and WiFi (Wireless-Fidelity, wireless network). If the terminal set up is successfully connected with the merchant's Bluetooth and WiFi, it can be determined that the delivery person has arrived at the store and the timing is started. Similarly, if the connection is disconnected, it can be determined that the delivery person has left the store, the timing will be stopped, and the waiting time will be obtained.

However, the inventor realized that there are certain drawbacks in obtaining the waiting time of the delivery personnel by using the above two methods. The disadvantage of the first method is that the accuracy of GPS will be affected by the environment. In confined spaces such as indoors, elevators, stairs, and corridors, the accuracy of GPS is poor, and there will be constant beating. Simply use GPS to determine the distribution personnel and stores. There is a large error in the distance between them. Moreover, reporting is the active behavior of the delivery personnel, and there is room for cheating. It is very likely that the delivery personnel have not arrived but report to the platform that they have arrived, which is unfair to the merchant. The disadvantage of the second method is that the distance of Bluetooth communication is too short, and the compatibility of Bluetooth is poor. It is very likely that the delivery personnel have arrived but cannot connect to the Bluetooth of the store. However, WiFi has a long communication distance and has a certain penetrating power. Sometimes the delivery personnel are still in the downstairs of the store or have been connected to the store's WiFi when there is a certain distance from the store, resulting in the combination of Bluetooth and WiFi. Certain errors and failures are possible. Therefore, the present invention proposes a method for obtaining the waiting time. The environment data and behavior data are obtained by collecting data on the environment of the delivery personnel and the behavior of the delivery personnel in the historical delivery process, and the delivery personnel are calculated by establishing a time series model. Based on the reliability of the corresponding environmental data and behavior data in each time slice, determine when the delivery staff arrives at the store and when they leave based on the time slice with high reliability, and then obtains the waiting time of the delivery staff, and uses the delivery staff in the delivery The multivariate data modeling in the process is used to calculate the waiting time of the delivery personnel, which solves the problem that the coverage of the geofence based on a single signal is too large or too small, and improves the accuracy of the waiting time calculation. In the following, the present invention will be described as an example for determining whether the delivery personnel arrive at the store and the waiting time of the delivery personnel in the store. As for the waiting time at the location, the present invention does not specifically limit the waiting time in which process the delivery person is determined.

Therefore, in order to realize the present invention, it is necessary to collect the environmental data and behavior data of the delivery personnel during the delivery process. It should be noted that, in the embodiment of the present invention, a target distribution facility is used as an example for description, wherein, the target distribution facility may actually be a selected takeaway, courier, distribution robot, distribution box, etc., as long as the facility has a distribution function The waiting time can be obtained for all of them, and the present invention does not limit the specific style of the distribution facility. When acquiring multiple environmental data and multiple behavior data of the target distribution facility in the historical distribution process, each distribution facility will carry a terminal device for communication, and the current terminal device is usually equipped with multiple different functions Sensors, such as IMU sensors for sensing the acceleration or angular velocity of distribution facilities, magnetic sensors for sensing changes in the magnetic properties of sensitive components caused by external factors, etc., are also equipped with Bluetooth modules, wireless modules, cellular modules, etc. Therefore, various types of signals uploaded to the platform by the terminal equipment in the historical distribution process can be used to summarize and count the multiple environmental data and behavior data of the target distribution facility in the historical distribution process. The specific process of acquiring multiple environmental data and multiple behavior data is as follows: First, extract multiple terminal signals uploaded by the terminal equipment of the target distribution facility in the historical distribution process, wherein the multiple terminal signals include at least inertial signals, Bluetooth signals, Wireless, magnetic and cellular signals. It should be noted that, after extracting multiple terminal signals, considering the possibility of errors in the terminal signals, there may be terminal signals that are completely opposite to most terminal signals in the continuously uploaded terminal signals. For the error of waiting time, after multiple terminal signals are acquired, abnormal value filtering, denoising and filtering processing can be performed on the multiple terminal signals, that is, data preprocessing is performed on the obtained terminal signals, so as to discard abnormal terminal signals. . Then, based on the uploading time of the multiple terminal signals, the multiple terminal signals are divided into multiple signal groups, and the uploading time of at least one terminal signal included in each signal group in the multiple signal groups is consistent, that is, the uploading time of the multiple signal groups is the same. The signals are divided into the same signal group, and then through the signal analysis of the signal group, the environment and behavior of the target distribution facility at the upload time of the signal group can be obtained. Whether the upload time has arrived or left. Finally, signal analysis is performed on multiple signal groups to obtain multiple environmental data and multiple behavior data.

Wherein, taking any one of the multiple signal groups as an example to illustrate the process of signal analysis by the signal group, the specific process is as follows: First, for each signal group in the multiple signal groups, read multiple targets included in the signal group Signal. Secondly, based on multiple target signals, determine the environment where the target distribution device is located at the upload time corresponding to the signal group, and obtain environment data, where the environment data is at least any one of an indoor environment and an outdoor environment. Since target signals include inertial signals, Bluetooth signals, wireless signals, magnetic induction signals and cellular signals, different types of terminal signals have different characteristics. For example, Bluetooth signals, wireless signals and magnetic induction signals are prone to beating indoors. , simply using one of these signals to determine whether the target distribution facility is indoors is likely to cause false positives and false negatives. In view of this, when determining environmental data, it is necessary to combine different types of terminal signals to determine which type of environment the distribution facility is currently in. For example, since the humidity and temperature will change after entering the room, the speed of the target distribution facility will increase. The terminal equipment is likely to be successfully connected to the WiFi in the store, and the communication quality in the store will be worse than that in the outdoor. Therefore, it can be judged whether the distribution facility is indoors based on the magnetic induction signal, inertial signal, wireless signal and cellular signal. . Furthermore, when the target distribution facility takes the elevator, it will rise or fall with the elevator, the speed of the elevator is fast, and the environment in the elevator is different from indoor and outdoor. Therefore, the distribution can be judged based on the magnetic induction signal and inertial signal. Whether the facility is inside the elevator car. Also, when the target distribution facility is around the store, only the strong penetration of WiFi will change the wireless signal. Therefore, it can be judged whether the distribution facility is around the store based on the wireless signal.

After determining the environmental data, it is also necessary to obtain behavior data of the target distribution facility based on multiple target signals. The movement state of distribution facilities in the distribution process can be roughly divided into two types: macro-state and micro-state. Micro-states include behavioral states of distribution facilities based on terminal equipment, such as making a call, placing a phone in a pocket, using a phone by hand, and swinging a phone by hand. Whether it is a macro state or a micro state, it is directly related to whether the distribution facility has reached the store. Generally speaking, when a distribution facility is in a store, there are only two macro-states, stationary and slow-moving, and there may be micro-states such as making a phone call and using a hand-held phone. On the contrary, when the distribution facility is outdoors, there will be macro-states such as cycling and walking at extreme speeds, and there will be micro-states such as the phone in the pocket and the phone swinging freely. Whether it is a macro-state or a micro-state is obviously mutually exclusive, so the behavior data of the target distribution facility can be obtained based on multiple target signals. The specific process of obtaining the behavior data is as follows: according to multiple target signals, determine the use state of the terminal device at the upload time corresponding to the signal group and the movement state of the target distribution facility at the upload time corresponding to the signal group, and based on the use state and movement state, obtain Behavior data, the behavior data is at least any one of waiting behavior and delivery behavior. Specifically, when the usage state indicates that the terminal device executes the user instruction and the movement state indicates that the target distribution facility is stationary or the movement speed of the movement state is less than the speed threshold, the waiting behavior is set as the behavior data. That is to say, when the terminal device executes the user instruction of the distribution facility, it can be determined that the distribution facility is using the terminal device, and the distribution facility is stationary or slow. The combination of the two can directly indicate that the distribution facility is waiting in the store, and the waiting behavior can be set as behavior data. . When the use state indicates that the terminal device is in standby and the movement state indicates that the distribution facility is moving or the movement speed of the movement state is greater than or equal to the speed threshold, the distribution behavior is set as the behavior data. That is to say, the terminal device is not used by the distribution facility, it may be in the pocket of the distribution facility, and the distribution facility is riding or walking at high speed. The combination of the two can directly indicate that the distribution facility is delivering outside, not waiting in the store. So, set the shipping behavior as behavior data.

Through the above process, the signal analysis of the target signal included in the signal group is realized, and the environmental data and behavior data of the signal group are generated according to the actual situation in the signal group. By repeating the above process, each signal group in the multiple signal groups can be read separately, the environmental data and behavior data of each signal group can be output, and a plurality of environmental data and a plurality of behavior data can be obtained. It should be noted that, the above process of outputting environment data and behavior data does not have an obvious sequence, and may be performed simultaneously or sequentially, which is not specifically limited in the present invention.

202. Map multiple environmental data and multiple behavior data to multiple time windows of the historical distribution process.

In the embodiment of the present invention, after multiple environmental data and multiple behavior data are determined, in order to integrate and study multiple environment data and multiple behavior data, it is necessary to establish a time series model based on multiple environment data and multiple behavior data , so that reliability analysis can be performed on multiple environmental data and multiple behavior data in the future. Among them, the establishment of a time series model requires data modeling according to time slices. Therefore, the historical distribution process needs to be divided into multiple time slices. A time slice is also a time window. Multiple environmental data and multiple behavior data They are respectively mapped to different time windows, so that the time series model is subsequently established according to different data of the time windows. Specifically, when mapping multiple environmental data and multiple behavior data to multiple time windows of the historical distribution process, first, the historical distribution process needs to be time-divided based on the preset division duration to obtain multiple time windows, wherein , the duration of each time window in the multiple time windows is equal to the preset division duration. Then, for each environmental data or each behavior data in the multiple environmental data and the multiple behavior data, determine the target upload time corresponding to each environmental data or each behavior data, and query the target upload time in the multiple time windows. At the target time window, each environment data or each behavior data is mapped to the target time window, and the mapping between multiple environment data and multiple behavior data and multiple time windows is completed.

203. Calculate multiple reliability levels of multiple time windows.

In the embodiment of the present invention, after the mapping between the data and the time windows is completed, the multiple reliability levels of the multiple time windows can be calculated according to the data corresponding to each time window. Among them, the multiple reliability levels indicate the occurrence probability of environmental data and behavior data corresponding to multiple time windows, that is to say, the reliability level indicates the probability that the distribution facility is in the time window in the current time window. The environmental data and behavior data indicate the status, through this reliability, each time window can be marked later, and the status of the target distribution facility is marked on the time window, so as to determine which time window the target distribution facility is in. Arrival, which time window the target distribution facility leaves. Therefore, firstly, multiple time windows are used to train a time series model for multiple environmental data and multiple behavior data, and a time series model for the distribution process is generated. Among them, the time series model of the distribution process includes the estimated state corresponding to each time window in multiple time windows, that is to say, the time series model of the distribution process is actually based on the global environmental data and behavior data, combined with the experience experienced in the entire historical distribution process It can predict the actual state of the target distribution facility in each time window for the entire duration, so as to compare the estimated state with the actual environmental data and behavior data corresponding to the time window in the future, and determine that there are hundreds of time windows. What percentage of the probability will actually generate the corresponding environmental data and behavioral data.

After the time series model of the distribution process is constructed, for each time window in the multiple time windows, the target estimated state corresponding to the time window in the time series model of the distribution process is obtained. When the target environmental data and target behavior data corresponding to the time window match the target estimated state, it means that the state estimated based on the time series model of the distribution process is consistent with the state corresponding to the environmental data and behavior data corresponding to the time window. When they are consistent, the credibility of the environmental data and behavioral data is very high. Therefore, the environmental credibility of the target environmental data and the behavioral credibility of the target behavioral data are set as the first default values, based on the environmental credibility. and behavioral credibility to form the credibility of the time window. Among them, the first default value may actually be 100%, and 100% is used to prove the consistency of the three. Further, when any data in the target environment data or target behavior data corresponding to the time window does not match the target estimated state, it means that there may be an error in data collection, and the reliability of the data is questioned. The second default value is calculated as the difference between the standard value and the second default value, the difference is used as the reliability of the data that does not match the target estimated state, and the second default value is used as the data matching the target estimated state. Credibility, get the credibility of the time window. Among them, the second default value may actually be 66%, and the standard value may actually be 1. For example, assuming that the target estimated state is a waiting state, the environmental data indicates that the store is waiting, and the behavior data indicates that the delivery is in progress, then the environmental data The set environmental confidence level may be 66%, and the behavior confidence level set for the behavior data may be 1-66%=34%. Further, when the target environment data or target behavior data corresponding to the time window do not match the target estimated state, considering that the target estimated state is obtained for the global big data, the current time window before and after will be comprehensively considered. The data of the time window is more persuasive than the environmental data and behavioral data. Therefore, a third default value can be obtained, and the environmental reliability of the target environmental data and the behavioral reliability of the target behavioral data can be set to the first value. Three default values, based on environmental reliability and behavioral reliability to form the reliability of the time window, wherein the third default value can actually be a lower value, such as 1%, with 1% to emphasize environmental data, Behavioral data is diametrically opposed to the target estimated state. Alternatively, it is also possible to consider the consistency of the target environment data and the target behavior data, and set the environmental credibility of the target environment data and the behavior credibility of the target behavior data as the difference between the above-mentioned standard value and the second default value. value, taking into account the possibility of environmental and behavioral data. It should be noted that the data set above are all examples. In the process of practical application, big data can also be used to calculate each value, or the environmental data of each time window can be directly calculated based on the generated time series model of the distribution process. Evaluate with behavior data, and directly output environmental credibility and behavior credibility from the distribution process timing model, and the present invention does not specifically limit the method of generating credibility.

By repeating the above process, the reliability of multiple time windows can be calculated respectively, and multiple reliability levels can be obtained.

204. Determine the arrival time window and the departure time window in multiple time windows according to the multiple reliability levels.

In this embodiment of the present invention, after multiple reliability levels are determined, the arrival time window and the departure time window may be determined in multiple time windows according to the multiple reliability levels. Among them, referring to the content in the above step 203, it can be seen that the environmental reliability and behavioral reliability corresponding to some time windows are very low, which may reach 1%. This reliability is probably due to the environmental data and behavioral data. It is necessary to adjust the environmental data and behavior data to improve the reliability and ensure that the subsequent marking of the time window is accurate. Therefore, after obtaining multiple credibility, it is necessary to determine To-be-adjusted credibility, adjust the to-be-adjusted credibility to ensure that all credibility can be at a higher value. Among them, when adjusting the credibility, first, the credibility to be adjusted needs to be extracted from the multiple credibility, and the credibility to be adjusted actually means that both the environmental credibility and the behavior credibility are included. Below the reliability threshold, the reliability threshold may specifically be 2%, 50%, etc., which is not specifically limited in the present invention. Subsequently, data adjustment is performed on the to-be-adjusted time window corresponding to the to-be-adjusted reliability. Among them, the adjustment mechanism is adjusted according to the combination of the previous time window and the next time window of the time window to be adjusted. Therefore, it is necessary to extract the previous time window and the next time window of the time window to be adjusted. A standard time window is determined in the following time window, and the environmental data and behavior data included in the standard time window are inconsistent with the environment data and behavior data included in the time window to be adjusted. That is to say, if the previous time window or the next time window is consistent with the environmental data and behavior data included in the time window to be adjusted, and the time window consistent with the time window to be adjusted is not extracted because the credibility is too low If it comes out, it means that this time window is credible, and the data included in the time window to be adjusted needs to be adjusted according to the time window that is inconsistent with its data, so the time window that is inconsistent with the data of the time window to be adjusted will be used as the Standard time window. Finally, assign the environmental data, behavior data included in the standard time window, and the reliability corresponding to the standard time window to the time window to be adjusted, to obtain the adjusted time window to be adjusted, and subsequently set based on the adjusted time window to be adjusted window label. For example, it is assumed that the environmental data in the time window to be adjusted indicates indoors and the behavior data indicates waiting, while the environmental data in the previous time window indicates outdoor, the behavior data indicates delivery, the environmental data in the next time window indicates indoors, and the behavior data indicates Wait, the previous time window is the standard time window, adjust the time window to be adjusted according to the previous time window to indicate that the environment data is outdoors, and the behavior data indicates that the delivery is sufficient, and also set the credibility of the previous time window. in the window to be adjusted. It should be noted that it is very likely that the data of the previous time window and the next time window are inconsistent with the data of the window to be adjusted, which means that the data of the previous time window and the next time window are consistent. Therefore, at the previous time window The window and the next time window can be adjusted by selecting any time window to be adjusted.

In the process of practical application, the above-mentioned adjustment process of the time window can be performed based on the recursive model, that is, the environmental data, behavior data and credibility corresponding to each time window obtained in the above process are all input into the recursive model, by The recursive model judges and adjusts the time window with low reliability. Specifically, the recursive model may be LSTM (Long Short-Term Memory, long short-term memory network), GRU (Gate Recurrent Unit, recurrent neural network), etc. The present invention does not limit the specific content of the recursive model.

After the credibility adjustment is completed, the credibility and data corresponding to all current time windows are relatively accurate. Therefore, a window label can be set for each time window in multiple time windows according to multiple credibility degrees. , the window label is used to indicate the state of the target distribution facility in each time window. Specifically, the window label is at least any one of the arrival label or the departure label. When setting the window label, for each time window in multiple time windows, considering that the environmental credibility and behavior credibility in some time windows are different, there are high and low, and the lower one does not reach the above Therefore, it is necessary to obtain the target credibility corresponding to the time window, and compare the target environment credibility and target behavior credibility included in the target credibility. If the credibility of the target environment and the credibility of the target behavior are consistent, it means that the state indicated by the environment data and behavior data in the time window is consistent and more reliable. Therefore, the environment data and behavior of the time window are set for the time window. Window label for data matching. If the credibility of the target environment and the credibility of the target behavior are inconsistent, it is necessary to extract the specified credibility from the credibility of the target environment and the credibility of the target behavior, and set a window that matches the data corresponding to the specified credibility for the time window. label, in which the specified credibility is greater than the credibility of the target environment and the credibility of the target behavior except the specified credibility, that is, the larger of the two credibility is determined. Reliability, set the window label of the data corresponding to the larger reliability for the time window. For example, assuming that the environmental reliability corresponding to the environmental data in the time window is 66%, and the behavioral reliability corresponding to the behavior data is 34%, the window label set for the time window matches the environmental data.

After setting the window label for each time window in the multiple time windows according to the multiple reliability degrees, obtain multiple consecutive candidate time windows with the window label as the arrival label in the multiple time windows, and use multiple candidate time windows to form A window queue, wherein the window labels of the previous time window and the next time window of the window queue in the multiple time windows are both leave labels. That is to say, assuming that the window label of time window A is the departure label, the window labels of time windows B to E are the arrival label, the next time window of time window E is F, and the window label of time window F is the departure label, then the window The queues are time windows B to E. Then, the candidate time window ranked first in the window queue is used as the arrival time window, and the candidate time window ranked at the end of the window queue is used as the departure time window. Continuing with the above example, assuming that the window queues are time windows B to E, the arrival time window is B, and the departure time window is E. It should be noted that, since the recursive model has the characteristic of recursively processing data based on time, when determining the arrival time window and the departure time window, it can also be implemented based on the recursive model.

205. Use the time interval between the arrival time window and the departure time window as the waiting time.

In this embodiment of the present invention, after the arrival time window and the departure time window are determined, the time interval between the arrival time window and the departure time window can be used as the waiting time. Specifically, the start time point of the arrival time window may be determined, the end time point of the departure time window may be determined, and the time interval between the start time point and the end time point may be used as the waiting time.

In summary, the whole process can be summarized as follows:

Referring to FIG. 2B, a plurality of terminal signals collected by a terminal device are acquired, outlier filtering, denoising and filtering are performed on the multiple terminal signals, and based on the processed multiple terminal signals, the environmental data of the distribution facility during the distribution process and the behavioral data. By modeling the environmental data and behavioral data, a time series model is obtained, and the time series model is used to determine the credibility of the environmental data and behavioral data. Get wait times for delivery facilities.

In the method provided by the embodiment of the present invention, the environmental data and behavior data of the target distribution facility in the historical distribution process are acquired, the environmental data and behavior data are mapped to the time window of the historical distribution process, the reliability of the time window is calculated, and the reliability of the time window is calculated according to Credibility, determine the arrival time window and the departure time window in the time window, take the time interval between the arrival time window and the departure time window as the waiting time, and use the multivariate data of the delivery personnel in the delivery process to calculate the waiting time of the delivery personnel, It solves the problem that the coverage of the geofence based on a single signal is too large or too small, ensures that the waiting time is obtained by effective information means, improves the accuracy of the calculation of the waiting time, makes the waiting time convincing, and promotes the catering industry. The intelligent process of service.

Further, as a specific implementation of the method described in FIG. 1 , an embodiment of the present invention provides an apparatus for obtaining a waiting time. As shown in FIG. 3A , the apparatus includes: an obtaining module 301 , a calculation module 302 and a determination module 303 .

The obtaining module 301 is used to obtain a plurality of environmental data and a plurality of behavior data of the target distribution facility in the historical distribution process;

The calculation module 302 is configured to map the multiple environmental data and the multiple behavior data to multiple time windows of the historical distribution process, and calculate multiple reliability levels of the multiple time windows, the The multiple reliability levels indicate the occurrence probability of the environmental data and the behavior data corresponding to the multiple time windows;

The determining module 303 is configured to determine an arrival time window and a departure time window in the multiple time windows according to the multiple reliability levels;

The determining module 303 is further configured to use the time interval between the arrival time window and the departure time window as the waiting time.

In a specific application scenario, as shown in FIG. 3B , the obtaining module 301 includes: an extracting unit 3011 , a dividing unit 3012 and a parsing unit 3013 .

The extraction unit 3011 is configured to extract multiple terminal signals uploaded by the terminal equipment of the target distribution facility during the historical distribution process, the multiple terminal signals at least include inertial signals, Bluetooth signals, wireless signals, and magnetic induction signals and cellular signals;

The dividing unit 3012 is configured to divide the plurality of terminal signals into a plurality of signal groups based on the upload time of the plurality of terminal signals, and each signal group in the plurality of signal groups includes at least one terminal signal. The upload time is the same;

The analyzing unit 3013 is configured to perform signal analysis on the multiple signal groups respectively to obtain the multiple environmental data and the multiple behavior data.

In a specific application scenario, the parsing unit 3013 is configured to, for each signal group in the multiple signal groups, read multiple target signals included in the signal group; based on the multiple target signals, determine the The environment in which the target distribution device is located at the upload time corresponding to the signal group, obtains environmental data, and the environmental data is at least any one of an indoor environment and an outdoor environment; according to the multiple target signals, determine the The use state of the terminal device at the upload time corresponding to the signal group and the motion state of the target distribution facility at the upload time corresponding to the signal group, based on the use state and the motion state, to obtain behavior data, the The behavior data is at least any one of waiting behavior and delivery behavior; read each signal group in the multiple signal groups respectively, output the environment data and behavior data of each signal group, and obtain the multiple environments data and the plurality of behavioral data.

In a specific application scenario, the parsing unit 3013 is configured to, when the use state indicates that the terminal device executes user instructions and the movement state indicates that the target distribution facility is stationary or the movement speed of the movement state is less than a speed threshold When the waiting behavior is set as the behavior data; when the use state indicates that the terminal device is in standby and the movement state indicates that the distribution facility is moving or the movement speed of the movement state is greater than or equal to the speed When the threshold is reached, the delivery behavior is set as the behavior data.

In a specific application scenario, as shown in FIG. 3C , the computing module 302 includes: a dividing unit 3021 , a determining unit 3022 and a mapping unit 3023 .

The division unit 3021 is configured to perform time division on the historical distribution process based on a preset division duration to obtain the multiple time windows, and the duration of each time window in the multiple time windows is equal to the preset duration dividing time;

The determining unit 3022 is configured to, for the plurality of environmental data and each environmental data or each behavior data in the plurality of behavior data, determine the target upload time corresponding to each environmental data or each behavior data;

The mapping unit 3023 is configured to query the target time window where the target upload time is located in the multiple time windows, map each environmental data or each behavior data to the target time window, and complete all and a mapping between the plurality of environmental data and the plurality of behavioral data and the plurality of time windows.

In a specific application scenario, as shown in FIG. 3D , the calculation module 302 includes: a training unit 3024 , an acquisition unit 3025 and a setting unit 3026 .

The training unit 3024 is configured to use the multiple time windows to perform time series model training on the multiple environmental data and the multiple behavior data to generate a distribution process time sequence model, where the distribution process time sequence model includes the The estimated state corresponding to each time window in multiple time windows;

The obtaining unit 3025 is configured to obtain, for each time window in the plurality of time windows, the target estimated state corresponding to the time window in the time sequence model of the distribution process;

The setting unit 3026 is configured to set the environmental credibility of the target environment data and the behavior of the target behavior data when the target environment data and target behavior data corresponding to the time window match the target estimated state. The credibility is set as a first default value, and the credibility of the time window is formed based on the environmental credibility and the behavior credibility;

The setting unit 3026 is further configured to calculate the difference between the standard value and the second default value when any data in the target environment data or target behavior data corresponding to the time window does not match the target estimated state, and calculate the difference between the standard value and the second default value. The difference is used as the reliability of the data that does not match the target estimated state, and the second default value is used as the reliability of the data matched with the target estimated state to obtain the time window. credibility;

The setting unit 3026 is further configured to set the environmental reliability of the target environment data and the target behavior when neither the target environment data or the target behavior data corresponding to the time window matches the target estimated state. The behavior credibility of the data is set to a third default value, and the credibility of the time window is formed based on the environmental credibility and the behavior credibility;

The setting unit 3026 is further configured to calculate the reliability of the multiple time windows respectively to obtain the multiple reliability.

In a specific application scenario, as shown in FIG. 3E , the determining module 303 includes: a setting unit 3031 , an obtaining unit 3032 and a determining unit 3033 .

The setting unit 3031 is configured to set a window label for each time window in the plurality of time windows according to the plurality of reliability degrees, and the window label is at least any one of an arrival label or a departure label;

The obtaining unit 3032 is configured to obtain, from the multiple time windows, multiple consecutive candidate time windows whose window label is the arrival label, and use the multiple candidate time windows to form a window queue, and the window queue is located in all the time windows. The previous time window and the window label of the next time window in the multiple time windows are the leaving labels;

The determining unit 3033 is configured to use the candidate time window at the top of the window queue as the arrival time window, and the candidate time window at the end of the window queue as the departure time window.

In a specific application scenario, the setting unit 3031 is configured to, for each time window in the multiple time windows, obtain the target reliability corresponding to the time window, and set the target environment included in the target reliability The credibility and the credibility of the target behavior are compared; if the credibility of the target environment and the credibility of the target behavior are consistent, set the time window to match the environment data and behavior data of the time window window label; if the credibility of the target environment and the credibility of the target behavior are inconsistent, the specified credibility is extracted from the credibility of the target environment and the credibility of the target behavior, which is the time The window is set with a window label matching the data corresponding to the specified confidence level, and the specified confidence level is greater than the target environment confidence level and the target behavior confidence level except for the specified confidence level. a credibility.

In a specific application scenario, as shown in FIG. 3F , the determination module 303 further includes: a first extraction unit 3034 , a second extraction unit 3035 and an adjustment unit 3036 .

The first extracting unit 3034 is configured to extract the credibility to be adjusted from the multiple credibility degrees, and the credibility to be adjusted includes both the environmental credibility and the behavior credibility that are lower than the credibility threshold ;

The second extraction unit 3035 is configured to determine the to-be-adjusted time window corresponding to the to-be-adjusted reliability, and to extract the previous time window and the next time window of the to-be-adjusted time window;

The determining unit 3033 is further configured to determine a standard time window in the previous time window and the next time window, the environment data and behavior data included in the standard time window and the environment included in the to-be-adjusted time window inconsistent data and behavioral data;

The adjustment unit 3036 is configured to assign the environmental data, behavior data included in the standard time window, and the reliability corresponding to the standard time window to the time window to be adjusted, to obtain the adjusted time window to be adjusted , and set a window label based on the adjusted time window to be adjusted.

The device provided by the embodiment of the present invention maps the environmental data and behavior data to the time window of the historical distribution process by acquiring the environmental data and behavior data of the target distribution facility in the historical distribution process, calculates the reliability of the time window, and calculates the reliability of the time window according to Credibility, determine the arrival time window and the departure time window in the time window, take the time interval between the arrival time window and the departure time window as the waiting time, and use the multivariate data of the delivery personnel in the delivery process to calculate the waiting time of the delivery personnel, It solves the problem that the coverage of the geofence based on a single signal is too large or too small, ensures that the waiting time is obtained by effective information means, improves the accuracy of the calculation of the waiting time, makes the waiting time convincing, and promotes the catering industry. The intelligent process of service.

It should be noted that, for other corresponding descriptions of the functional units involved in the apparatus for obtaining a waiting time provided in this embodiment of the present invention, reference may be made to the corresponding descriptions in FIG. 1 and FIG. 2A to FIG. 2B , which will not be repeated here.

In an exemplary embodiment, referring to FIG. 4 , a device is also provided, and the device 400 includes a communication bus, a processor, a memory and a communication interface, and may also include an input and output interface and a display device, wherein among the various functional units Communication between them can be accomplished through the bus. The memory stores a computer program and a processor for executing the program stored in the memory and executing the method for obtaining the waiting time in the above embodiment.

A computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the steps of the method for obtaining the waiting time.

From the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, and the software product can be stored in a non-volatile storage medium (which may be CD-ROM, U disk, mobile hard disk, etc.), including several The instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of this application.

Those skilled in the art can understand that the accompanying drawing is only a schematic diagram of a preferred implementation scenario, and the modules or processes in the accompanying drawing are not necessarily necessary to implement the present application.

Those skilled in the art can understand that the modules in the device in the implementation scenario may be distributed in the device in the implementation scenario according to the description of the implementation scenario, or may be located in one or more devices different from the implementation scenario with corresponding changes. The modules of the above implementation scenarios may be combined into one module, or may be further split into multiple sub-modules.

The above serial numbers in the present application are only for description, and do not represent the pros and cons of the implementation scenarios.

The above disclosures are only a few specific implementation scenarios of the present application, however, the present application is not limited thereto, and any changes that can be conceived by those skilled in the art should fall within the protection scope of the present application.

Claims (10)

1. A latency acquisition method, comprising:

acquiring a plurality of environmental data and a plurality of behavior data of a target distribution facility in a historical distribution process;

mapping the plurality of environment data and the plurality of behavior data to a plurality of time windows of the historical distribution process, and calculating a plurality of credibility of the plurality of time windows, wherein the plurality of credibility indicates occurrence probability of the environment data and behavior data corresponding to the plurality of time windows;

determining an arrival time window and a departure time window in the plurality of time windows according to the plurality of degrees of confidence;

and taking the time interval between the arrival time window and the departure time window as the waiting time.

2. The method of claim 1, wherein the obtaining a plurality of environmental data and a plurality of behavioral data of the target distribution facility during historical distribution comprises:

extracting a plurality of terminal signals uploaded by terminal equipment of the target distribution facility in the historical distribution process, wherein the plurality of terminal signals at least comprise inertia signals, Bluetooth signals, wireless signals, magnetic induction signals and cellular signals;

dividing the plurality of terminal signals into a plurality of signal groups based on the uploading time of the plurality of terminal signals, wherein the uploading time of at least one terminal signal included in each of the plurality of signal groups is consistent;

and respectively carrying out signal analysis on the plurality of signal groups to obtain the plurality of environment data and the plurality of behavior data.

3. The method of claim 2, wherein the performing signal analysis on the plurality of signal groups to obtain the plurality of environmental data and the plurality of behavior data comprises:

for each of the plurality of signal groups, reading a plurality of target signals included in the signal group;

determining the environment of the target distribution equipment at the uploading time corresponding to the signal group based on the plurality of target signals to obtain environment data, wherein the environment data is at least any one of indoor environment and outdoor environment;

determining the use state of the terminal equipment at the uploading time corresponding to the signal group and the motion state of the target distribution facility at the uploading time corresponding to the signal group according to the target signals, and obtaining behavior data based on the use state and the motion state, wherein the behavior data is at least any one of waiting behavior and distribution behavior;

and respectively reading each signal group in the plurality of signal groups, and outputting the environmental data and the behavior data of each signal group to obtain the plurality of environmental data and the plurality of behavior data.

4. The method of claim 3, wherein deriving behavioral data based on the usage state and the motion state comprises:

setting the waiting behavior as the behavior data when the use status indicates that the terminal device executes a user instruction and the motion status indicates that the target distribution facility is stationary or the motion speed of the motion status is less than a speed threshold;

setting the delivery behavior as the behavior data when the usage state indicates that the terminal device is in standby and the motion state indicates that the delivery facility moves or a motion speed of the motion state is greater than or equal to the speed threshold.

5. The method of claim 1, wherein mapping the plurality of environmental data and the plurality of behavioral data to a plurality of time windows of the historical delivery process comprises:

based on preset dividing time, time division is carried out on the historical distribution process to obtain a plurality of time windows, and the time length of each time window in the time windows is equal to the preset dividing time length;

determining target uploading time corresponding to each environmental data or each behavior data for the plurality of environmental data and each environmental data or each behavior data in the plurality of behavior data;

inquiring a target time window of the target uploading time in the plurality of time windows, mapping each environmental data or each behavior data to the target time window, and completing the mapping between the plurality of environmental data and the plurality of behavior data and the plurality of time windows.

6. The method of claim 1, wherein said calculating a plurality of degrees of confidence for said plurality of time windows comprises:

training a time sequence model on the plurality of environment data and the plurality of behavior data by adopting the plurality of time windows to generate a distribution process time sequence model, wherein the distribution process time sequence model comprises an estimated state corresponding to each time window in the plurality of time windows;

for each time window in the plurality of time windows, acquiring a corresponding target estimation state of the time window in the distribution process time sequence model;

when the target environment data and the target behavior data corresponding to the time window are matched with the target estimation state, setting the environment credibility of the target environment data and the behavior credibility of the target behavior data as first default values, and forming the credibility of the time window based on the environment credibility and the behavior credibility;

when any one of the target environment data or the target behavior data corresponding to the time window is not matched with the target estimated state, calculating a difference value between a standard value and a second default value, taking the difference value as the credibility of the data which are not matched with the target estimated state, and taking the second default value as the credibility of the data which are matched with the target estimated state to obtain the credibility of the time window;

when the target environment data or the target behavior data corresponding to the time window are not matched with the target estimation state, setting the environment credibility of the target environment data and the behavior credibility of the target behavior data as third default values, and forming the credibility of the time window based on the environment credibility and the behavior credibility;

and respectively calculating the credibility of the plurality of time windows to obtain the plurality of credibility.

7. The method of claim 1, wherein said determining an arrival time window and a departure time window among said plurality of time windows based on said plurality of degrees of confidence comprises:

setting a window label for each time window in the plurality of time windows according to the plurality of credibility, wherein the window label is at least any one of an arrival label or a departure label;

obtaining a plurality of continuous candidate time windows with window labels as the arrival labels from the plurality of time windows, and adopting the plurality of candidate time windows to form a window queue, wherein the window labels of the window queue in the previous time window and the next time window of the plurality of time windows are the departure labels;

and taking the candidate time window arranged at the head of the window queue as the arrival time window, and taking the candidate time window arranged at the tail of the window queue as the departure time window.

8. A latency acquisition apparatus, comprising:

the acquisition module is used for acquiring a plurality of environmental data and a plurality of behavior data of the target distribution facility in the historical distribution process;

a calculation module, configured to map the plurality of environmental data and the plurality of behavior data to a plurality of time windows of the historical distribution process, and calculate a plurality of reliabilities of the plurality of time windows, where the plurality of reliabilities indicate occurrence probabilities of the environmental data and the behavior data corresponding to the plurality of time windows;

a determining module configured to determine an arrival time window and a departure time window in the plurality of time windows according to the plurality of degrees of confidence;

the determining module is further configured to use a time interval between the arrival time window and the departure time window as a waiting time.

9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.

10. A readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.

Images