CN-122020068-A - Space-time trajectory nested map algorithm for cognitive disorder patient and abnormal behavior recognition method and system

CN122020068ACN 122020068 ACN122020068 ACN 122020068ACN-122020068-A

Abstract

The invention provides a space-time track nesting atlas algorithm for a cognitive disorder patient and an abnormal behavior recognition method and system, and relates to the technical field of behavior data analysis based on live tracks, comprising the steps of continuously monitoring track points of an individual to be observed in a preset time period and forming a track sequence according to time sequence; identifying stay segments based on track sequences, generating location nodes, outputting attribute information of the location nodes, giving a location node label to each track point in the track sequences, constructing a periodic track map, combining the periodic track maps of a plurality of past time periods, constructing a multi-period space-time track nested map which comprises all the periodic track maps and nested track maps, constructing an individual reference behavior mode according to all the periodic track maps, constructing an individual current behavior mode according to the nested track maps, and analyzing track deviation. The invention has the advantage that the reliability of individual monitoring of cognition disorder patients can be improved.

Inventors

WAN RUYI
HUANG PING
LIU YU

Assignees

四川互慧软件有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. The space-time trajectory nested map algorithm for the cognitive disorder patient and the abnormal behavior recognition method are characterized by comprising the following steps: Continuously monitoring track points of an individual to be observed in a preset time period, forming a track sequence according to a time sequence, and preprocessing the track sequence; Identifying stay segments based on the track sequence, generating location nodes and outputting attribute information of the location nodes, wherein the attribute information comprises center coordinates, stay time, access frequency and access time distribution of all track points in the location nodes, and endowing each track point in the track sequence with a location node label; Constructing a periodic track graph according to the track sequence and the site node labels, wherein the periodic track graph comprises site nodes, transfer relations among the site nodes and transfer attributes, and the transfer attributes comprise start time, transfer time consumption, stay time after transfer and transfer times of the site nodes; Combining the cycle track graphs of a plurality of past time periods to construct a multi-cycle space-time track nesting graph, wherein the multi-cycle space-time track nesting graph comprises all cycle track graphs and nesting track graphs, and the nesting track graphs comprise all place nodes after adjacent combination, transfer relations among the place nodes and average transfer attributes; constructing an individual reference behavior mode according to all the periodic track diagrams, and constructing an individual current behavior mode according to the nested track diagrams; And carrying out track deviation analysis according to the similarity of the individual reference behavior mode and the individual current behavior mode.
2. The method for identifying the space-time trajectory nesting map algorithm and the abnormal behavior of the cognition impaired patient according to claim 1, wherein the method for preprocessing the trajectory sequence comprises noise reduction, abnormal data elimination and missing data supplementation.
3. The method for recognizing a stay segment based on a track sequence according to claim 1, wherein if distances between a plurality of continuous track points in the track sequence and central coordinates of the track points are smaller than a first distance threshold r, and the duration of the plurality of continuous track points is longer than a time threshold These successive track points are considered as one dwell segment.
4. The method for identifying the space-time trajectory nesting atlas algorithm and abnormal behavior of the patient with cognitive impairment according to claim 3, wherein the method for generating the location node is as follows: After all the stay segments are identified, acquiring the center coordinates of each stay segment; the distance of the central coordinate is smaller than a preset second distance threshold value Merging the stay segments into a location node, and recording the center coordinates, stay time length, access frequency and access time distribution of all track points in each location node; the method for endowing each track point in the track sequence with a spot node label comprises the following steps: ; Wherein, the The locus point being the time point t Is used for the identification of the tag of (c), Representing the node of the location, As the number of the place node, null represents a node not belonging to any place.
5. The method for identifying the space-time trajectory nesting map algorithm and abnormal behaviors of the cognitive disorder patient according to claim 1, wherein the method for constructing the periodic trajectory map is as follows: drawing all the place nodes; forming a time sequence place node label sequence L according to the place node labels of the track points in the track sequence, and removing the place node labels which are continuously repeated; And acquiring a place node transfer relation of the individual activities to be observed based on a place node label sequence L, if the transfer exists between two place nodes, establishing a directed edge between the two place nodes according to the transfer direction, and recording transfer attributes at the directed edge.
6. The method for identifying the space-time trajectory nesting map algorithm and abnormal behaviors of the cognitive impairment patient according to claim 5, wherein the method for constructing the nesting map is as follows: And (3) aligning the position nodes of the periodic track graph of each time period, wherein the method comprises the following steps: acquiring all the place nodes and the corresponding center coordinates thereof in all the periodic track diagrams; The distance between the center coordinates is smaller than a preset second distance threshold value Is combined into the same place node; after alignment is completed, all the merged place nodes are obtained, and the transfer relation and the average transfer attribute between the merged place nodes are comprehensively counted.
7. The method for identifying abnormal behaviors and space-time trajectory nesting atlas algorithm for cognition impaired patients according to claim 1, wherein the method for constructing the individual reference behavior pattern is as follows: calculating site node similarity : ; Wherein, the The number of said place nodes of said periodic trajectory graph representing a kth historical time period in said multi-periodic spatiotemporal trajectory nesting chart, A pair of pairs of similar location nodes in the periodic trajectory graph representing the kth and the (k+1) th historical time periods in the multi-period spatiotemporal trajectory nesting graph, wherein if the distance between the center position of a location node in one kth periodic trajectory graph and the center position of a location node in one (k+1) th historical time period periodic trajectory graph is smaller than a first distance threshold value, the two location nodes are defined as a pair of pairs of similar location nodes, and N is the total number of the periodic trajectory graphs in the multi-period spatiotemporal trajectory nesting graph; Calculating transfer path similarity : ; Wherein min and max are functions that are minimum and maximum, respectively, A set of said location nodes in said cycle trace plot representing a kth historical time period, In the cycle trace map for the kth historical time period Site node to The number of transfers of the site node, A positive number of not more than 0.01; Calculating the similarity of transfer time : ; ; Wherein, the As an intermediate parameter, a parameter which is a function of the parameter, 、 And In the cycle trace map of the kth historical time period respectively Site node to Start time of transfer of the place node, transfer time consumption and stay time after transfer; constructing the individual reference behavior pattern B: 。
8. the method for identifying the abnormal behavior and the space-time trajectory nesting atlas algorithm for the cognition impaired patient according to claim 7, wherein the method for constructing the current behavior pattern of the individual is as follows: Calculating the similarity of the current node based on the cycle track graph and the nested track graph of the current time cycle Similarity of current transfer paths Similarity to the current transition time : ; Wherein, the Representing the total number of location nodes in the nested trajectory graph after the proximity merge, The total number of location nodes in the cycle trajectory graph representing the current time cycle, A pair of pairs of similar location nodes in the periodic trace plot representing the nested trace plot and a current time period, the two location nodes being defined as a pair of similar location node pairs if a distance between a center position of a location node in the periodic trace plot for a current time period and a center position of a location node in the nested trace plot is less than a first distance threshold; ; Wherein, the A set of said place nodes in said period trace plot representing a current time period, In the place nodes after the adjacent combination in the nested trajectory graph Site node to The number of transfers of the site node, In the cycle trace map for the current time cycle Site node to The number of transfers of the site node; ; ; Wherein, the As an intermediate parameter, a parameter which is a function of the parameter, 、 And In the cycle trace map of the current time cycle respectively Site node to The start time of the transfer of the location node, the transfer time consuming and the stay time after the transfer, 、 And Respectively in the place nodes after the adjacent combination in the nested trajectory graph Site node to Start time of transfer of the place node, transfer time consumption and stay time after transfer; constructing the current behavior pattern B' of the individual: 。
9. The method for identifying the abnormal behavior and the space-time trajectory nested graph algorithm for the cognitive disorder patient according to claim 1, wherein the method for analyzing the trajectory deviation according to the similarity between the individual reference behavior pattern and the individual current behavior pattern is as follows: calculating cosine similarity of the individual reference behavior pattern and the individual current behavior pattern; If the cosine similarity is larger than a preset similarity threshold, judging that the behavior of the individual to be observed is not abnormal, and if the cosine similarity is not larger than the preset similarity threshold, judging that the behavior of the individual to be observed is abnormal.
10. The space-time trajectory nested spectrum algorithm and the abnormal behavior recognition system for the cognitive disorder patient are applied to the space-time trajectory nested spectrum algorithm and the abnormal behavior recognition method for the cognitive disorder patient, which are characterized by comprising the following steps: The track acquisition module is used for continuously monitoring track points of an individual to be observed in a preset time period, forming a track sequence according to a time sequence and preprocessing the track sequence; the track identification module is used for identifying stay segments based on the track sequence, generating the site nodes and outputting attribute information of the site nodes, wherein the attribute information comprises center coordinates, stay time, access frequency and access time distribution of all track points in the site nodes, and a site node label is given to each track point in the track sequence; The periodic track diagram construction module is used for constructing a periodic track diagram according to the track sequence and the site node labels, wherein the periodic track diagram comprises site nodes, transfer relations among the site nodes and transfer attributes, and the transfer attributes comprise start time, transfer time consumption, stay time after transfer and transfer times of the site nodes; The multi-period space-time track nesting map construction module is used for merging the period track maps of a plurality of past time periods to construct a multi-period space-time track nesting map, wherein the multi-period space-time track nesting map comprises all period track maps and nesting track maps, and the nesting track maps comprise all place nodes after adjacent merging, transfer relations among the place nodes and average transfer attributes; the behavior pattern construction module is used for constructing an individual reference behavior pattern according to all the periodic track diagrams and constructing an individual current behavior pattern according to the nested track diagrams; And the behavior analysis module is used for carrying out track deviation analysis according to the similarity between the individual reference behavior mode and the individual current behavior mode.

Description

Space-time trajectory nested map algorithm for cognitive disorder patient and abnormal behavior recognition method and system Technical Field The invention relates to the technical field of behavior data analysis based on a live track, in particular to a space-time track nested graph algorithm for a cognitive disorder patient and an abnormal behavior recognition method and system. Background In the disease development process, cognition disorder patients often show behavior characteristics such as wandering, lost, repeated round trip, abnormal time going out, gradually reduced activity range and the like, and the behaviors not only increase the risk of wandering, but also bring great challenges to personal safety and care management of the patients. In the traditional care mode, the behavior of a patient is mainly analyzed in a accompany, inquiry and data tracing mode, the method has hysteresis, abnormal behavior is difficult to discover in time, quantitative evaluation and trend analysis are difficult to conduct, and early risk signals are easy to miss. In the prior art, a kind of anti-lost scheme based on a positioning technology (such as GPS or RFID) combined with an electronic fence is proposed, and the basic principle is that a target individual is monitored in a preset space range, and an alarm is triggered when the individual exceeds a limited area. However, the method essentially belongs to a space constraint strategy based on rules, and has the defects that the space position change is focused, the modeling capability of time-space combination dimension is lacking, different cognition disorder patients have obvious differences in the aspects of the moving range, the travel path, the work and rest rules and the like, misinformation or omission is easily caused by adopting a unified threshold value or a fixed rule, the fine care requirement is difficult to meet, and the structural change in the track cannot be effectively depicted because the conventional method is generally judged based on single out-of-range. Therefore, the track monitoring of the cognitive impairment patient needs to be optimized, the problems of the prior art are overcome, a track monitoring method with more adaptability and space-time feature extraction capability is realized, and the reliability of individual monitoring of the cognitive impairment patient is further improved. Disclosure of Invention The invention aims to provide a space-time track nesting atlas algorithm for a cognitive disorder patient, an abnormal behavior recognition method and a system, which can realize a track monitoring method with space-time feature extraction capability and higher adaptability. The invention is realized by the following technical scheme: a space-time trajectory nested graph algorithm and an abnormal behavior recognition method for cognition disorder patients comprise the following steps: Continuously monitoring track points of an individual to be observed in a preset time period, forming a track sequence according to a time sequence, and preprocessing the track sequence; Identifying stay segments based on the track sequence, generating location nodes and outputting attribute information of the location nodes, wherein the attribute information comprises center coordinates, stay time, access frequency and access time distribution of all track points in the location nodes, and endowing each track point in the track sequence with a location node label; Constructing a periodic track graph according to the track sequence and the site node labels, wherein the periodic track graph comprises site nodes, transfer relations among the site nodes and transfer attributes, and the transfer attributes comprise start time, transfer time consumption, stay time after transfer and transfer times of the site nodes; Combining the cycle track graphs of a plurality of past time periods to construct a multi-cycle space-time track nesting graph, wherein the multi-cycle space-time track nesting graph comprises all cycle track graphs and nesting track graphs, and the nesting track graphs comprise all place nodes after adjacent combination, transfer relations among the place nodes and average transfer attributes; constructing an individual reference behavior mode according to all the periodic track diagrams, and constructing an individual current behavior mode according to the nested track diagrams; And carrying out track deviation analysis according to the similarity of the individual reference behavior mode and the individual current behavior mode. Preferably, the method for preprocessing the track sequence comprises noise reduction, abnormal data elimination and missing data supplementation. Preferably, the method for identifying the stay segment based on the track sequence is that if the distances between the continuous track points and the central coordinates of the track points in the track sequence are smaller than a first distance threshold r, and the dura