KR100982439B1 - Apparatus and method for the prediction of group mobility in wireless sensor network - Google Patents

Abstract

The present invention relates to an apparatus and method for predicting mobility of a group in a wireless sensor network. The present invention relates to a group of sensing information of a group member and to generate and manage a tree structure at a predetermined level. An apparatus and method for predicting mobility of a group in a wireless sensor network for searching a tree structure corresponding to a 'current position' and predicting a 'next position' to move a group according to the corresponding tree structure having the largest weight are provided.

To this end, the present invention provides a mobility prediction apparatus, comprising: sensing information collection means for generating group movement information by collecting sensing information corresponding to the number of group members and combining them into one; Tree structure management means for securing the group movement information to generate a new tree structure at a predetermined level, and managing the entire tree structure according to the comparison result between the new tree structure and the existing tree structure; And searching the corresponding tree structure in the entire tree structure by using the group movement information of the previous position and the group movement information of the current position collected by the sensing information collecting means when the mobility of the group is predicted. Mobility prediction means for predicting the mobility of the group according to the weight of the structure.

Wireless sensor network, group mobility, tree structure, mobility prediction, weight

Description

Apparatus and method for the prediction of group mobility in wireless sensor network

The present invention relates to an apparatus and method for predicting mobility of a group in a wireless sensor network. More particularly, the present invention relates to a group of sensing information of group members, and generates and manages a tree structure at a predetermined level. Apparatus and method for predicting mobility of a group in a wireless sensor network for searching a tree structure corresponding to a previous position 'and a' current position 'and predicting a' next position 'to move the group according to the corresponding tree structure having the largest weight. It is about.

Recently, the wireless sensor network, which has been in the spotlight, has been applied to a technology for predicting a member's movement in advance by detecting mobility of a member who is active indoors in a home network.

In the past, researches on how to predict the mobility of a single member have been mainly conducted. A typical representative method is a method of predicting the mobility of a single member by generating a tree structure using Hybrid Markov Models or LeZi-Update algorithm. That is, in the related art, a tree structure is generated by connecting continuous information around a route, and predicting mobility of a single member through a method in which weights are assigned when the same continuous information is generated.

As such, a method of predicting the mobility of a single member through a Markov model or a reg algorithm may have accuracy in predicting the mobility of a member, but the processing thereof is complicated.

Meanwhile, in order to predict the mobility of at least two or more members, that is, a group, each member is distinguished to predict mobility of each member, or an RFID tag is attached to each member to predict mobility of each member. .

However, the method of predicting the mobility of the group tracks the movement path of each member and predicts the mobility of the member according to the result, so that the system construction cost or data processing cost for this can be increased.

In the past, a technique for tracking the mobility of a single member through opening and closing the door and using electronic devices has been proposed. That is, it tracks the position of a single member through a sensor for detecting the opening and closing of the door, a sensor for detecting the use of electronic devices. For example, if the TV is turned on after the door is detected, the occupant is determined to have moved to the room. If the door is detected after the TV is turned off, the occupant is determined to have moved out of the room. By the way, the technology for tracking the mobility of members through the opening and closing of the door and the use of electronic devices, because it predicts the mobility of a single member, there is a limit that is difficult to predict the mobility of the group.

Therefore, in the conventional method, when predicting mobility of a group as well as mobility of a single member, there is a need to predict a moving direction according to an algorithm that is simple to construct and simple to process.

In particular, when the mobility of the group is predicted, a technique for predicting the direction of movement based on the minimum information about the group is required.

Therefore, the prior art as described above has a problem of estimating the moving direction according to an algorithm that is difficult to construct a system and a complicated processing method when predicting the mobility of a group as well as the mobility of a single member. to be.

Therefore, the present invention combines the sensing information of group members into one, creates and manages a tree structure at a predetermined level, and searches for a tree structure corresponding to the 'previous position' and 'current position' of the group when the group mobility is predicted to have the most weight. An object of the present invention is to provide an apparatus and method for predicting mobility of a group in a wireless sensor network for predicting a 'next position' to which a group moves according to a large corresponding tree structure.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention which are not mentioned can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. Also, it will be readily appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

According to an aspect of the present invention, there is provided a mobility prediction device, comprising: sensing information collection means for generating group movement information by collecting sensing information corresponding to the number of group members and combining the same into one; Tree structure management means for securing the group movement information to generate a new tree structure at a predetermined level, and managing the entire tree structure according to the comparison result between the new tree structure and the existing tree structure; And searching the corresponding tree structure in the entire tree structure by using the group movement information of the previous position and the group movement information of the current position collected by the sensing information collecting means when the mobility of the group is predicted. Mobility prediction means for predicting the mobility of the group according to the weight of the structure.

According to another aspect of the present invention, there is provided a mobility prediction method comprising: generating group movement information by collecting sensing information corresponding to the number of group members and combining the group movement information into one; A management step of creating a new tree structure at a predetermined level by securing the group movement information, and managing the entire tree structure according to a comparison result between the new tree structure and the existing tree structure; And searching for the tree structure in the entire tree structure by using the group movement information of the previous position and the group movement information of the current position, which are collected for the mobility estimation of the group, and according to the weight of the tree structure. A prediction step of predicting mobility.

According to another aspect of the present invention, there is provided a mobility prediction apparatus, comprising: sensing information collection means for generating group movement information by collecting sensing information corresponding to the number of group members and combining the same into one; And mobility prediction means for predicting mobility of the group through calculation through Bayesian estimation when the group movement information is input, wherein the mobility prediction means includes a probability of occurrence of sequences of positions generated so far when Bayesian estimation is performed. 'Probability distribution for the probability that the next position occurs' and 'Probability distribution for the probability indicated by the current position and previous position' when the next position is determined. When the sequences of are determined, the mobility of the group is predicted according to the probability distribution of what probability the next position will occur.

In addition, the present invention provides a mobility prediction method comprising: generating group movement information by collecting sensing information corresponding to the number of group members and combining them into one; And a group mobility prediction step of predicting mobility of the group through calculation through Bayesian estimation when the group movement information is input, wherein the group mobility prediction step includes' sequences of positions generated so far when Bayesian estimation occurs. 'Probability Distribution for Probability of Probability' and 'Probability Distribution for Probability of Next Position' and 'Probability Distribution for Probability of Current Position and Previous Position' when Next Position is Determined, When sequences of positions are determined, the mobility of the group is predicted according to the probability distribution of what probability the next position will occur.

As described above, the present invention has the effect of predicting the mobility of a group as well as the mobility of a single member.

In addition, the present invention has an effect that can be adapted to the environment in which the mobility of members by changing the weight of the tree structure for the sensing information continuously.

In addition, the present invention has an effect that can easily process the mobility prediction of the member.

In addition, the present invention has the effect of predicting mobility on the basis of minimum information on the movement of the group active in the room.

BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings, It can be easily carried out. In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Prior to describing the present invention, the term 'group member' in the present invention refers to a single member as well as at least two or more members who are active in a specific building (limited space) such as a home or an office. The meaning thereof will be described in detail, and a person skilled in the art will readily understand that the present invention may be applied to a single member (see FIG. 2 to be described later).

In particular, in the present invention, after the number of group members active indoors is determined, the mobility of the group members according to the present invention is predicted. At this time, in the present invention, the number of group members in the building is determined by detecting the group members entering and leaving the building (for example, attaching a sensor to the front and rear of the exit and detecting the access of the group members according to the detection order).

1 is a configuration diagram of an apparatus for predicting mobility of a group in a wireless sensor network according to the present invention.

As shown in FIG. 1, in the wireless sensor network according to the present invention, the mobility prediction apparatus of a group (hereinafter referred to as “mobility prediction apparatus”) 120 may track the position of group members through a wireless sensor network formed indoors. It also predicts where the group is going to move. That is, the mobility prediction apparatus 120 tracks the position of the group by using the 'sensing information' transmitted from the sensor 110 capable of sensing each group member, and predicts the position to which the group is to move in advance.

In general, the sensor 110 includes a processor module, a communication module, a low capacity storage module, and a sensing module. Here, the sensing module may be a conventional sensing method, for example, infrared, light, ultrasonic, etc. may be applied.

The sensor 110 transmits the sensing information sensed as the group member falls within the corresponding sensing range to the mobility prediction device 120 through another sensor, or relays the sensing information transmitted from the other sensor.

In addition, the sensor 110 is pre-installed indoors, and detects only a state in which the group member is present or absent. In this case, the sensor 110 may periodically notify the mobility prediction device 120 of the state of the 110, but transmits the sensing information to the mobility prediction device 120 when the group member is detected in the corresponding area. It is desirable to.

In particular, the sensor 110 is preferably installed so that the sensing ranges do not overlap each other, and provides the sensed sensing information to the mobility prediction apparatus 120. This is because the mobility prediction apparatus 120 predicts the mobility of the group member through the sensing information, so that the mobility of the group member can be reliably predicted according to the accuracy of the 'sensing information corresponding to the movement path of the group member'.

In addition, the sensor 110, when installed on the ceiling of the building, adjusts so that the group members are detected only when a predetermined distance (for example, 1m) or more from the ground, thereby detecting errors detected by pets such as dogs or cats Prevent it in advance.

The wireless sensor network shown in FIG. 1 is represented by a graph structured by connecting unit nodes by a link, and an object or an object of the graph is called 'vertex' and a link of the graph is called an 'edge'. Here, 'vertex' means each sensor 110 that is a unit node, 'edge' means a connection relationship between the sensor (110).

In detail, the mobility prediction apparatus 120 includes a sensing information collecting unit 121, a tree structure managing unit 122, a mobility predicting unit 123, and a database 124.

The sensing information collecting unit 121 collects the sensing information transmitted from the sensor 110 according to the movement of the group member who is active indoors and provides the tree structure managing unit 122.

In particular, the sensing information collecting unit 121 focuses on 'number of group members' corresponding to 'number of group members' in order to predict mobility of the group when providing sensing information. Each of the 'sensing information' corresponding to each other is provided to the tree structure manager 122 in the form of a bundle of information (hereinafter referred to as "group movement information"). For example, when two group members are active, the sensing information collecting unit 121 collects two pieces of sensing information corresponding to the number of group members from the sensor 110 and then tree-groups the group movement information. Provided to the structure management unit 122.

This is to predict the mobility of the group by considering each sensing information indicative of the position of each group member under the assumption that each group member influences each other's behavior.

In other words, the group tends to repeat the behavior patterns that are scattered in the common living spaces (living room, kitchen, dining room, etc.) and scattered into individual living spaces (bedroom, bathroom, etc.) while influencing each other's behaviors in limited spaces. There is this. For example, group members in a family may watch TV together in the living room and gather at the dining room at mealtime, or they may consistently repeat the pattern of moving to their bedroom after watching their work in the bathroom. .

As described above, the mobility prediction apparatus 120 predicts mobility of group members in a state in which the number of group members who are active indoors is determined in advance. Accordingly, the sensing information collecting unit 121 matches the elements of the group movement information with the number of group members and substitutes the respective sensing information transmitted from the sensor 110.

Meanwhile, in the present invention, when N group members exist in a certain space, for each of the 'sensing information', the group movement information is expressed as "(sensing information 1, sensing information 2, ..., sensing information N)". Let's explain.

For example, if two group members exist in a certain space, the sensor A (in this case, the sensing information is called "a") and the sensor B (the sensing information is called "b") are detected at the same time. , Group movement information is expressed as "(a, b)". This means that one member passes through Sensor A and the other member passes through Sensor B.

In addition, in the present invention, since the behavior characteristics of each group member are not defined, if each element constituting each other in the two group movement information is the same, it is treated the same when the tree structure is generated by the tree structure manager 122 which will be described later. do.

For example, group movement information "(a, b)" and "(b, a)" are treated the same. However, the sensing information a and b represent the front and rear that are transmitted from the sensor 110 to the sensing information collecting unit 121.

In particular, when the sensing information is continuously transmitted from any one sensor 110, the sensing information collection unit 121 determines that the group members are located within the same sensing range. At this time, the sensing information collecting unit 121 displays each element of the group movement information corresponding to the number of group members as the same sensing information (that is, sensing information transmitted from any one sensor 110) and displays the tree structure managing unit ( 122).

For example, in the case where two group members exist, when the sensing information "a" is continuously transmitted from one sensor 110, the sensing information collecting unit 121 displays the group movement information as "(a, a" ) "To the tree structure management unit 122.

In addition, the sensing information collecting unit 121 is located within the same sensing range for a predetermined time (that is, the re-detection time set in the sensor 110), the group member by any one sensor 110 As this is detected again, the same sensing information may be retransmitted. For example, the group members may be eating in a restaurant, watching TV in a living room, sleeping in a bedroom, or studying in a bedroom.

In this case, the sensing information collecting unit 121 filters the same sensing information transmitted from the sensor 110 after excluding the sensing information initially transmitted from the sensor 110 to the tree structure manager 122. Improve the transmission efficiency of the transmitted sensing information. This is because the mobility of group members does not appear clearly from the tree structure generated from the same sensing information transmitted in succession, that is, duplicated sensing information.

The tree structure manager 122 may generate a tree structure having an arbitrary level (that is, a K = k tree structure) by sequentially listing the group movement information provided from the sensing information collecting unit 121. That is, assuming that the number of sensors 110 installed in the room is X and the number of group members active in the room is N, the tree structure manager 122 totally " _X C _N (ie, X combination N)" You can create a tree structure with any level.

However, in the present invention, it is preferable that the tree structure manager 122 generates a tree structure having a 'predetermined level having the highest predicted success rate for mobility of the group' determined through experiments. This is to increase the predicted success rate of group mobility through the tree structure having the level within 1 to 3 levels because the short behavior pattern of each group member who is active in a limited room repeatedly appears within 1 to 3 levels. to be.

In the present invention, when the level is determined to be '3' through the experiment shown in FIG. 5 to be described later (that is, when the tree structure manager 122 generates a three-level tree structure (that is, K = 3 tree structure). ] Will be described as an example. However, it will be apparent to those skilled in the art that unlike the three-level tree structure, the mobility of the group can be predicted by using any level tree structure.

The tree structure manager 122 manages the three-level tree structure by using the group movement information continuously transmitted by the sensing information collector 121.

First, the tree structure manager 122 secures three group movement information sequentially from a sequence of group movement information continuously transmitted from the sensing information collecting unit 121 to generate a three-level tree structure. In other words, when three group movement information is secured sequentially, the tree structure manager 122 starts to generate a three-level tree structure.

At this time, the tree structure manager 122 checks how much the new three-level tree structure corresponds to the existing three-level tree structure in preparation for the new three-level tree structure and the existing three-level tree structure.

That is, when the new three-level tree structure corresponds to the existing three-level tree structure (that is, when the new three-level tree structure and the existing three-level tree structure are the same), the tree structure manager 122 adds a new three-level tree structure. Add weight to the existing three-level tree structure without saving it. This is used to estimate the mobility of the group member by the mobility prediction unit 123, which will be described later, by assigning a weight to a three-level tree structure in which the mobility of the group member is repeatedly displayed.

In addition, when the new three-level tree structure is part of the root of the existing three-level tree structure or less, the tree structure management unit 122 updates the portion of the tree structure manager 122 to share a portion corresponding to the root of the existing three-level tree structure.

If the new three-level tree structure does not correspond to the existing three-level tree structure (that is, when the new three-level tree structure and the existing three-level tree structure are different), the tree structure management unit 122 generates a new three-level tree structure. Save it.

Then, when one group movement information is added, the tree structure management unit 122 deletes the highest priority group movement information (that is, the oldest group movement information among the three group movement information) from the three-level tree structure previously created and the rest. After the two group movement information, the added group movement information is connected to create a three-level tree structure.

For example, when the group movement information is sequentially transmitted with (a, c), (a, b), and (b, b), the tree structure management unit 122 is "(a, c)-(a, b)-(b, b) "to create a three-level tree structure. Here, the tree structure manager 122 checks whether the new three-level tree structure (a, c)-(a, b)-(b, b) "corresponds to the existing three-level tree structure, and processes as described above. do

After that, if the group movement information (a, b) is added, the tree structure management unit 122 has the highest priority in the three-level tree structure "(a, c)-(a, b)-(b, b)" created above. Delete the group movement information "(a, c)" and connect the remaining two group movement information "(a, b)-(b, b)" followed by the group movement information "(a, b)" The tree structure "(a, b)-(b, b)-(a, b)" is newly created.

If the group movement information is first transmitted after (a, c), (a, b), and (b, b) sequentially, and further (a, b) is transmitted, the tree structure manager 122 has three levels. Create and store the tree structures "(a, c)-(a, b)-(b, b)" and "(a, b)-(b, b)-(a, b)" in the database 124 Done.

The mobility predicting unit 123 predicts the mobility of the group. That is, the mobility predicting unit 123 predicts the mobility of the group based on the weight given to the existing three-level tree structure stored in the database 124.

First, the mobility predictor 123 checks the group movement information of the previous position and the group movement information of the current position from the sensing information collection unit 121 when the group mobility is predicted. .

Thereafter, the mobility predicting unit 123 generates a tree structure (that is, a two-level tree structure) using 'group movement information of the previous position' and 'group movement information of the current position', and then 'group movement information of the current position'. By predicting 'group movement information of the next position' to be followed, 'predictably the next position to which the group will move.

That is, the mobility predicting unit 123 generates a tree structure using 'group movement information of the previous position' and 'group movement information of the current position', from the root to 2 levels in the entire tree structure previously stored in the database 124. Navigate through the same tree structure. In this case, the mobility prediction unit 123 searches a tree structure having the largest weight among the tree structures, checks group movement information of the last level (that is, the third level) following the second level of the tree structure, and moves the group. The location corresponding to the information (that is, each location corresponding to the sensing information constituting the group movement information) is checked. Here, the mobility predicting unit 123 directly predicts the position corresponding to the group movement information as the next position to which the group will move.

On the other hand, the mobility predictor 123 is a tree structure having an arbitrary level (ie, a tree structure of K = k) when predicting the mobility of the group, indicating the position where the group moved from the first to k-2th level. The group movement information of k-2 previous positions' and the group movement information of one current position in total of the k-1th level are checked from the sensing information collecting unit 121. Thereafter, the mobility predicting unit 123 generates a k-1 level tree structure to predict 'group movement information of the next position (kth)' followed by 'group movement information of the current position (k-1)'. Done. Detailed description thereof will be readily understood by those skilled in the art, so a detailed description thereof will be omitted.

Alternatively, the mobility predictor 123 may probabilistically predict the mobility of the group through Bayesian estimation of Equation 1 below (ie, when the tree structure manager 122 is unnecessary).

At this time, the mobility prediction unit 123 continuously performs Bayesian estimation through the following [Equation 1] with respect to the group movement information collected through the sensing information collecting unit 121, whenever new group movement information is input. The probability must be calculated. This has a limitation in that the amount of calculation increases because new probability is continuously calculated as new group movement information is collected.

Hereinafter, the meaning of [Equation 1] is as follows.

'는 다음에 발생할 위치, '

'는 현재까지 발생한 위치의 시퀀스들(또는 벡터)을 나타낸다. 즉, 본 발명과 같이 3 레벨 트리구조를 가정하는 경우에는,

=(현재위치 X_n, 이전위치 X_n-1)이고,

=(다음에 나타날 위치 X_n+1)라고 할 수 있다. '

'Is the next occurrence,'

'Represents the sequence (or vector) of the position that has occurred so far. That is, in the case of a three-level tree structure as in the present invention,

= (Current position X _n , previous position X _n-1 ),

= (Next position X _{n + 1} ).

'는 현재까지 발생한 위치의 시퀀스들이 발생할 확률에 대한 확률분포를 나타낸다. '

'는 다음위치가 발생할 확률에 대한 확률분포를 나타낸다.'

'Represents a probability distribution of the probability of occurrence of the sequence of the position occurring so far. '

'Represents the probability distribution for the probability that the next position will occur.

'는 다음위치가 결정되었을 때, 현재위치 및 이전위치의 값이 나타내는 확률에 대한 확률분포를 나타낸다.'

'Represents the probability distribution of the probability represented by the values of the current position and the previous position when the next position is determined.

'는 현재까지 발생한 위치의 시퀀스들이 결정되었을 때, 다음위치가 어떤 확률로 발생할지에 대한 확률분포이다.'

'Is the probability distribution of what probability the next position will occur when the sequences of the positions that occur so far have been determined.

'를 계산해야 한다. 이때, 이동성 예측부(123)는 '

', '

' 및 '

'를 계산한 후에

를 계산한다. 이후, 이동성 예측부(123)는 상기와 같이 확률을 계산하여 그룹의 이동성을 예측할 수도 있다.In conclusion, the mobility prediction unit 123 calculates' By Bayesian estimation of Equation 1

'Must be calculated. At this time, the mobility prediction unit 123 is'

','

'And'

After calculating '

Calculate Thereafter, the mobility predicting unit 123 may calculate the probability as described above to predict the mobility of the group.

'에 대응되는 값을 베이지안 추정을 통해 그룹의 이동성을 예측하는 방식의 복잡한 계산을 수행하지 않고 얻을 수 있는 결과를 제공한다. 이는 가중치가 가장 큰 위치가 그룹의 다음위치로 결정될 확률이 크기 때문이다.As described above, the method of having the tree structure manager 122 to predict the mobility of the group according to the weight of the tree structure is'

The value corresponding to 'provides a result that can be obtained without performing complicated calculation of the method of predicting group mobility through Bayesian estimation. This is because the position with the largest weight is likely to be determined as the next position in the group.

The database 124 stores and manages a three-level tree structure managed by the tree structure manager 122. At this time, the database 124 includes a three-level tree structure corresponding to the number of group members (ie, the number of elements of the group movement information).

In addition, the database 124 manages a list in which the identification information ID of the sensor 110 corresponds to the position of each sensor 110 installed indoors. Through this, the mobility prediction unit 123 may check the position of the sensing information when the mobility of the group is predicted.

Figure 2 is an exemplary illustration of a tree structure of a single member according to the present invention. Here, assume that FIG. 2 is a tree structure for the mobility of group member A. FIG. At this time, the information in the circle represents the sensing information, that is, identification information (ID) of each sensor 110, the number outside the circle represents the number of detection, that is, the weight of the sensor 110.

As shown in FIG. 2, the mobility prediction apparatus 120 of the present invention stores the mobility of group member A in a three-level tree structure. That is, group member A moves in a pattern such as "aba" (2 times), "abc" (4 times), "abe" (4 times), "ada" (8 times), and "ade" (12 times). It was.

Here, when the mobility prediction apparatus 120 predicts the mobility of the group member A, if the group member A moves in a pattern such as "a (201) -d (202)", the group member A is "a (201) -d. (202) "and then move to" e (203) ". This is because the group member A will move to " e 203 " after " a 201-d 202 " is larger than the weight moving to " a 204 ".

3 is a diagram illustrating an embodiment of a tree structure of a group according to the present invention. 3 assumes a three-level tree structure for the mobility of groups, ie group members A and B. FIG. At this time, the information in the circle represents the group movement information, the number outside the circle represents the number of detection, that is, the weight.

3 shows "(a, c) (a, b) (b, b) (a, b) (a, c) (b, c) (a, b) (b, b) (a, b)" As shown in the drawing, a three-level tree structure is generated from a sequence of group movement information continuously entered into the tree structure manager 122.

The tree structure manager 122 secures three group movement information sequentially provided from the sensing information collecting unit 121 when the three-level tree structure is generated. At this time, the tree structure management unit 122 generates a three-level tree structure by repeating the process of including new group movement information that is continuously performed except for the highest priority group movement information.

As described above, the tree structure management unit 122 gives a basic weight (for example, '1' is given at the time of initial creation) when generating the three-level tree structure and adds additional weight (for example, to the repeated pattern of the tree structure). When repeating, add '1'.

In the sequence of group movement information in FIG. 3, (a, c)-(a, b)-(b, b) [301], (a, b)-(b, b)-(a, b) [302 ], (b, b)-(a, b)-(a, c) [303], (a, b)-(a, c)-(b, c) [304], (a, c)- (b, c)-(a, b) [305], (b, c)-(a, b)-(b, b) [306], (a, b)-(b, b)-(a , b) A three-level tree structure is created as shown in [307].

In addition, the tree structure management unit 122 compares each node of the existing three-level tree structure and the new three-level tree structure, 'when there is no identical node', 'when there is the same node under the root', 'all the same' It manages 3-level tree structure according to the case of node.

In other words, when there is no identical node at all, the tree structure manager 122 stores the new three-level tree structure as well as the existing three-level tree structure in the database 124.

(B, b)-(a, b)-(a, c) [303] and (b, c)-(a, b)-(b, b) [306] of FIG. Since the structure and 'the same node does not exist at all' is stored in the database 124 as a new independent three-level tree structure.

In the case where the same node exists under the root, the tree structure manager 122 updates the tree structure by sharing a part corresponding to the existing 'three level tree structure' previously stored in the database 124.

(A, c)-(b, c)-(a, b) [305] in FIG. 3 is a portion corresponding to (a, c)-(a, b)-(b, b) [301] ( In this case, the tree structure is updated by sharing the root). Similarly, (a, b)-(a, c)-(b, c) [304] in FIG. 3 matches (a, b)-(b, b)-(a, b) [302]. Update the tree structure by sharing the part (here root).

In the case of 'all the same node', the tree structure manager 122 updates the tree structure by additionally assigning a weight of the 'existing three level tree structure' previously stored in the database 124.

(A, b)-(b, b)-(a, b) [307] of FIG. 3 is the same as (a, b)-(b, b)-(a, b) [302]. b)-(b, b)-(a, b) [302] is updated to update the tree structure (i.e. update from weight 1 to 2).

4 is a diagram illustrating an embodiment of mobility prediction of a group according to the present invention. Here, FIG. 4 is a three-level tree structure for mobility of groups, that is, group members A and B, and is assumed to be stored in the database 124. At this time, the information in the circle represents the group movement information, the number outside the circle represents the number of detection, that is, the weight.

The mobility predicting unit 123 collects sensing information by 'group movement information of the previous position' (b, c) [401] and 'group movement information of the current position' (a, b) [402] when predicting mobility of the group. In the case of transmission from the unit 121, this generates a two-level tree structure "(b, c)-(a, b)".

The three-level tree structure shown in FIG. 4 includes three tree structures sharing (b, c)-(a, b), that is, group movement information (a, c) in (b, c)-(a, b). ) [403], (b, b) [404] and (b, c) [405] respectively.

At this time, the mobility predicting unit 123 predicts which group movement information is 'group movement information of the next position' following 'group movement information of the current position' (a, b) [402]. That is, the mobility predicting unit 123 searches for a three-level tree structure having the same two-level tree structure "(b, c)-(a, b)" in the tree structure previously stored in the database 124. That is, in FIG. 4, (b, c)-(a, b)-(a, c), (b, c)-(a, b)-(b, b), (b, c)-(a, b)-(b, c) is found.

Here, the mobility predictor 123 checks (b, c)-(a, b)-(b, b) as the case where the weight is the largest. In this way, the mobility predictor 123 predicts that the group members A and B are currently located at a point corresponding to (a, b) but later move to the position corresponding to (b, b). . This means that when group members A and B exhibit a behavior pattern of "(b, c)-(a, b)", the probability that group members A and B move to the same position is high.

5 is a graph showing the prediction probability of the mobility of the group according to the level. The x-axis of FIG. 5 represents a value obtained by dividing the number of measured information by 100, and the y-axis represents a prediction probability.

As shown in FIG. 5, the tree structure of the present invention shows that the prediction probability of the mobility of the group may vary according to the level of the tree structure when checking based on the measured information. That is, the mobility prediction apparatus 120 may adjust the degree of prediction for the mobility of the group through the level of the tree structure when predicting the mobility of the group.

As shown in FIG. 5, it may be seen that the mobility prediction apparatus 120 shows the highest prediction success rate with respect to the mobility of the group in the case of a three-level tree structure among 1 to 3 levels.

Therefore, those skilled in the art can not only adjust the predicted success rate for the mobility of group members through the level of the tree structure, but also determine the level of the tree structure representing the optimal predicted success rate based on the measured information.

6 is a flowchart illustrating a method for predicting mobility of a group in a wireless sensor network according to the present invention.

As illustrated in FIG. 6, the mobility prediction apparatus 120 generates group movement information by collecting sensing information transmitted from at least one or more sensors 110 (S601). In this case, the mobility prediction apparatus 120 configures elements of the group mobility information according to a predetermined number of group members.

Thereafter, the mobility prediction apparatus 120 generates a tree structure according to a preset level (that is, a three-level tree structure in the present invention) based on the group movement information (S602). That is, the mobility prediction apparatus 120 generates three-level tree structure by securing three group movement information sequentially from a sequence of group movement information transmitted continuously.

In this case, the mobility prediction apparatus 120 checks whether the new three-level tree structure corresponds to an existing three-level tree structure previously stored (S603).

That is, the mobility prediction apparatus 120 determines the new three-level tree structure when the new three-level tree structure corresponds to the existing three-level tree structure (that is, when the new three-level tree structure and the existing three-level tree structure are the same). Weight is additionally given to the existing three-level tree structure without storing (S604).

In addition, when the new three-level tree structure is partially below the existing three-level tree structure root, the mobility prediction device 120 updates the portion corresponding to the root of the existing three-level tree structure to share the same part. S605).

The mobility prediction apparatus 120 may determine the new three-level tree structure when the new three-level tree structure does not correspond to the existing three-level tree structure (that is, when the new three-level tree structure and the existing three-level tree structure are different). Save (S606).

Subsequently, when one group movement information is added (S607), the mobility prediction apparatus 120 may select the highest priority group movement information (that is, the oldest group movement information among the three group movement information) in the three-level tree structure previously generated. The 3 level tree structure is newly regenerated by deleting the remaining two group movement information and then adding the group movement information, and performing any of the steps S604 to S606 following the step S603 (S608).

On the other hand, the mobility prediction apparatus 120 forms a two-level tree structure based on the group movement information of the previous position and the group movement information of the current position at the time of request for the mobility prediction of the group (S609), and the existing tree structure having the highest weight. Search for (S610).

Thereafter, the mobility prediction device 120 predicts mobility of the group through the existing tree structure (S611).

As described above, in the present invention, as the weight of the tree structure continuously changes as the group members move, the tree structure can be managed by reflecting the change in the situation of the group members, and the mobility of the group can be predicted by adapting to the change in the situation. .

On the other hand, the method of the present invention as described above can be written in a computer program. And the code and code segments constituting the program can be easily inferred by a computer programmer in the art. In addition, the written program is stored in a computer-readable recording medium (information storage medium), and read and executed by a computer to implement the method of the present invention. The recording medium may include any type of computer readable recording medium.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited by the drawings.

1 is a configuration diagram of an apparatus for predicting mobility of a group in a wireless sensor network according to the present invention;

Figure 2 illustrates an embodiment of a tree structure of a single member according to the present invention,

3 is an exemplary view illustrating a tree structure of a group according to the present invention;

4 is a diagram illustrating an embodiment of mobility prediction of a group according to the present invention;

5 is a graph showing the prediction probability of the mobility of the group according to the level,

6 is a flowchart illustrating a method for predicting mobility of a group in a wireless sensor network according to the present invention.

* Explanation of symbols for the main parts of the drawings

110: sensor 120: mobility prediction device

121: sensing information collection unit 122: tree structure management unit

123: mobility prediction unit 124: database

Claims (22)

In the mobility prediction device, Sensing information collecting means for generating group movement information by collecting sensing information corresponding to the number of group members and combining them into one; Tree structure management means for securing the group movement information to generate a new tree structure at a predetermined level, and managing the entire tree structure according to the comparison result between the new tree structure and the existing tree structure; And When the mobility of the group is predicted, the tree structure is searched in the entire tree structure by using the group movement information of the previous position and the group movement information of the current position collected by the sensing information collecting means. Mobility prediction means for predicting the mobility of the group according to the weight of Device for predicting mobility of a group in a wireless sensor network comprising a. The method of claim 1, The tree structure management means, When adding group movement information after creating a new tree structure, the tree structure is regenerated by deleting the highest group movement information of the new tree structure and connecting the group movement information which is added following the remaining group movement information. And re-managing the entire tree structure according to the present invention. The method of claim 2, The tree structure management means, According to the contrast between the new tree structure and the existing tree structure, the 'new tree structure corresponds to the existing tree structure', 'the new tree structure partially corresponds to the root of the existing tree structure' and 'the new tree structure' Is not all of the existing tree structure 'class mobility management apparatus of the wireless sensor network, characterized in that classified and managed. The method of claim 3, wherein The tree structure management means, In the case where the new tree structure corresponds to the entire tree structure, the mobility estimation apparatus of the group in the wireless sensor network, characterized in that the new tree structure is stored without additional weight. The method of claim 3, wherein The tree structure management means, In the case where the new tree structure is partially below the root of the existing tree structure, the mobility prediction apparatus of the group in the wireless sensor network, characterized by sharing and managing the corresponding portion below the root of the existing tree structure. The method of claim 3, wherein The tree structure management means, In the case where the new tree structure does not correspond to the existing tree structure, the new tree structure is stored and managed as it is. The method of claim 1, The mobility prediction means, Wireless sensor, characterized in that in the tree structure having the largest weight of the tree structure, the group movement information following the group movement information corresponding to the 'group movement information of the current position' is checked to predict the next position to which the group will move. Device for predicting mobility of groups in a network. The method of claim 1, The preset level is An apparatus for predicting mobility of a group in a wireless sensor network, which is set to '3' in order to reflect an optimal prediction success rate for the mobility of the group. The method of claim 1, The number of group members is A device for predicting mobility of a group in a wireless sensor network, characterized in that a predetermined entrance to a building is detected for at least one member or more. In the mobility prediction method, Generating group movement information by collecting sensing information corresponding to the number of group members and combining the group information into one; A management step of creating a new tree structure at a predetermined level by securing the group movement information, and managing the entire tree structure according to a comparison result between the new tree structure and the existing tree structure; And The tree structure is searched in the entire tree structure using the group movement information of the previous position and the group movement information of the current position, which are collected for the group mobility prediction, and the group mobility according to the weight of the tree structure. Prediction Step to Predict A method for predicting mobility of a group in a wireless sensor network comprising a. The method of claim 10, The management step, When adding group movement information, delete the highest priority group movement information of the new tree structure and reconstruct the tree structure by connecting the added group movement information following the remaining group movement information, and reconstruct the entire tree structure according to the comparison result of the existing tree structure. A method for predicting mobility of a group in a wireless sensor network, characterized by re-management. The method of claim 11, The management step, According to the contrast between the new tree structure and the existing tree structure, the 'new tree structure corresponds to the existing tree structure', 'the new tree structure partially corresponds to the root of the existing tree structure' and 'the new tree structure' Is a case that does not correspond to the existing tree structure is all classified and managed in the wireless sensor network, characterized in that the group mobility prediction method. 13. The method of claim 12, The management step, If the new tree structure corresponds to the entire tree structure, the weight of the existing tree structure is additionally managed without storing the new tree structure. 13. The method of claim 12, The management step, In the case where the new tree structure is partially below the root of the existing tree structure, the part corresponding to the root of the existing tree structure is shared and managed. 13. The method of claim 12, The management step, When the new tree structure does not correspond to the existing tree structure, the new tree structure is stored and managed as it is, characterized in that the mobility mobility group method in the wireless sensor network. The method of claim 10, The prediction step, Wireless sensor, characterized in that in the tree structure having the largest weight of the tree structure, the group movement information following the group movement information corresponding to the 'group movement information of the current position' is checked to predict the next position to which the group will move. How to predict group mobility in a network. The method of claim 10, The preset level is A method for predicting mobility of a group in a wireless sensor network, which is set to '3' in order to reflect an optimal prediction success rate for mobility of a group. The method of claim 10, The number of group members is A method for predicting mobility of a group in a wireless sensor network, characterized in that a predetermined entrance to a building is detected for at least one member or more. delete In the mobility prediction device, Sensing information collecting means for generating group movement information by collecting sensing information corresponding to the number of group members and combining them into one; And When inputting the group movement information, including mobility prediction means for predicting the mobility of the group through calculation through Bayesian estimation, The mobility prediction means, In Bayesian estimation, 'probability distribution for the probability of occurrence of the sequence of the current position' and 'probability distribution for the probability of the next position' and 'probability of the current position and the previous position when the next position is determined In the wireless sensor network, the probability of the group is predicted according to the probability distribution of the probability of the next position when the sequences of the positions occurring up to now are determined using the probability distribution of. Mobility prediction device. delete In the mobility prediction method, Generating group movement information by combining sensing information corresponding to the number of group members and combining them into one; And A group mobility prediction step of predicting mobility of a group through calculation through Bayesian estimation when the group movement information is input, The group mobility prediction step, In Bayesian estimation, 'probability distribution for the probability of occurrence of the sequence of the current position' and 'probability distribution for the probability of the next position' and 'probability of the current position and the previous position when the next position is determined In the wireless sensor network, the probability of the group is predicted according to the probability distribution of the probability of the next position when the sequences of the positions occurring up to now are determined using the probability distribution of. Mobility prediction method.

Images