Distributed Behavior Recognition Method Based on Wireless Sensor Network

Abstract

The invention discloses a distributed behavior recognition method based on a wireless sensor network, which includes the following steps: Step 1, collecting sensor sampling data, and passing the sensor sampling data to the Centralized control system; step 2, the centralized control system trains and mines the sampled data after the calculation, and excavates the frequent behavior trajectory pattern, thereby generating frequent behavior trajectory sets and frequent behavior state sets respectively; step 3, to the generated frequent behavior trajectory The set is further mined to obtain the relevant frequent behavior recognition knowledge set that can be distributed and stored on each sensor; step 4, the frequent behavior recognition knowledge set is stored in each sensor, and when the user performs a behavior action, the multicast report The completed recognition information in this article and the knowledge set stored by the sensor complete the behavior calculation process, identify frequent behaviors, and identify user behaviors.

Description

technical field [0001] The invention relates to the field of intelligent control, in particular to a distributed behavior recognition method based on a wireless sensor network. Background technique [0002] There are many deficiencies in the research and implementation of behavior recognition in traditional intelligent environments, mainly including the following two points: [0003] 1. Use video monitoring data and RFID data carrying tags to identify residents or pedestrians. [0004] The technology based on video surveillance collects data through video equipment, and uses image processing technology to analyze the data to identify user activities. Because in order to achieve more detailed activity recognition, it is necessary to arrange a large number of dense information collection equipment, and due to the high cost of video surveillance equipment, it is difficult to meet this requirement due to the defects of small coverage areas, and problems such as house structure ...

AI Technical Summary

Problems solved by technology

Because in order to achieve more detailed activity recognition, it is necessary to arrange a large number of dense information collection equipment, and due to the high cost of video surveillance equipment, it is difficult to meet this requirement due to the defects of small coverage areas, and problems such as house structure and indoor light in real scenes It will affect the effectiveness of the recognition results. In addition, the research method using visual monitoring equipment is an intrusive method. Directly taking pictures of the user's life leads to insufficient privacy protection for the user. In the home environment, people do not accept video for a long time. monitoring, so user acceptance of such methods is not high

In addition, actual participants are not willing to carry special gloves that can trigger sensors or other special devices that carry tags, so the RFID sensor method based on tags is not conducive to large-scale promotion

In addition, the application of intelligent environment in life assistance requires more accurate activity recognition. However, the traditional intelligent environment based on tags and video surveillance has relatively rough recognition of activities and cannot meet the required requirements. These methods are difficult to implement. Applying scale to real-world environments

[0005] 2. The current activity detection algorithms based on wireless sensor networks are mostly centralized algorithms, using a centralized data processing method, which requires each sensor to transmit the detected data back to the central node (usually a computer). The central node performs analysis and reasoning, but the real-time performance of this centralized algorithm is not strong, it cannot solve the detection errors that may be caused by network delays, and it does not utilize the computing and storage capabilities of the sensor nodes themselves.

Images