Unlock instant, AI-driven research and patent intelligence for your innovation.

CSI-based position-independent human body activity identification method

A recognition method and technology of human activity, applied in the recognition of patterns in signals, neural learning methods, character and pattern recognition, etc., to achieve the effect of improving performance, reducing samples, and enhancing the ability of position generalization

Pending Publication Date: 2022-06-28
HEFEI UNIV OF TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these two methods still require a small number of samples of the target location, which has certain limitations.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • CSI-based position-independent human body activity identification method
  • CSI-based position-independent human body activity identification method
  • CSI-based position-independent human body activity identification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0060] In this embodiment, figure 1 is the structure diagram of the whole system, figure 2 is the system model diagram. like figure 1 and figure 2 As shown, a position-independent action recognition method based on multi-task learning hierarchical graph neural network is carried out as follows:

[0061] Step 1. Select a rectangular area of ​​the indoor space; use the router outside the rectangular area as the sending device of the WIFI signal, denoted as AP, and use the network card as the receiving device, denoted as RP;

[0062] In this embodiment, the AP adopts a TL-WDR6500 router, the RP adopts an Intel 5300 network card, and the AP and RP have 2 transmitting antennas and 3 receiving antennas respectively;

[0063] Step 2. Divide the rectangular area into m blocks evenly, and set the position of the center point of each block as a calibration point, so as to obtain m calibration points in total; randomly select n points at the positions of non-calibration points as t...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a position-independent human body activity identification method based on CSI (Channel State Information), which comprises the following steps that: 1, action samples of a plurality of positions are acquired indoors by using CSI signals, and each position comprises a plurality of action types; 2, de-noising the sample by using wavelet transform, and extracting an amplitude value as sample data; 3, representing the sample as a hidden representation, and constructing sample structure information, class structure information and position structure information for the hidden representation by using a hierarchical graph neural network for data augmentation of the hidden representation; 4, representing the classification of the action at each position as a multi-class classification task, selecting data of a plurality of positions for training by utilizing multi-task learning, and training a classification model; and 5, finally, performing test classification on data at non-training positions by using the trained model. According to the method, action recognition of any indoor position can be realized with relatively high precision by utilizing WiFi-CSI under the condition of limited positions and samples.

Description

technical field [0001] The invention belongs to the technical field of wireless communication, in particular to a position-independent action recognition technology based on a multi-task graph neural network. Background technique [0002] In recent years, human motion recognition has played an important role in medical, entertainment, smart home and other fields. Human motion recognition systems employ different technologies, such as those based on wearable sensors, radar, computer vision, etc. Since the actions of the human body can be captured by Wifi signals, Wifi-based action recognition has attracted people's attention. Compared with other human motion recognition technologies, WiFi-based technologies have the advantages of low cost, no need for wearable devices, and no privacy concerns. [0003] WiFi signals include Received Signal Strength Information (RSSI) and Channel State Information (CSI). RSSI is widely used in WiFi-based human motion recognition, it is the a...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06N3/045G06F2218/02G06F2218/12G06F18/24G06F18/214Y02D30/70
Inventor 张勇程安东于光伟殷雨晴
Owner HEFEI UNIV OF TECH