Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Stranger action recognition method based on channel state information

A channel state information and action recognition technology, applied in the field of stranger action recognition, can solve problems such as affecting the recognition accuracy.

Pending Publication Date: 2021-02-02
HARBIN ENG UNIV
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Since different users will not be exactly the same even if they do the same action, and this inconsistency is more obvious in the channel state information that is easily affected by the environment, so the difference between users directly affects the accuracy of recognition

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
  • Stranger action recognition method based on channel state information
  • Stranger action recognition method based on channel state information

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0018] combine figure 1 , a method for stranger action recognition based on channel state information of the present invention includes four modules such as CSI data collection, signal preprocessing, feature extraction, and behavior recognition, and the specific steps include:

[0019] 1) Place the transmitter and receiver anywhere in the room, and the user stands between the transmitter and receiver to perform actions to obtain channel status information;

[0020] 2) Use the LOF algorithm to filter out abnormal values ​​on the collected original information;

[0021] 3) After preprocessing, the data is brought into the convolutional neural network to extract features and obtain feature vectors;

[0022] 4) Send the feature vector into the SVM for secondary training to obtain the CNN-SVM combination model. The above is the online training stage;

[0023] 5) In the offline recognition stage, after the data to be tested is processed in the first two steps, it is brought into ...

Embodiment 2

[0025] combine figure 1 , a stranger action recognition method based on channel state information of the present invention, specifically:

[0026] 1. First collect the original CSI data of the wireless signal from the receiver, and then perform data preprocessing, feature extraction, and action recognition in sequence.

[0027] 2. In the data collection stage, place the transmitter and receiver at any position in the room, but it is best to separate them by a certain distance. The user stands between the transmitter and the receiver to perform actions. The transmitter uses a The root antenna is a TP-Link802.11n wireless router, and the receiver is a levono laptop equipped with an Intel 5300 network card and three external antennas. Set the sampling frequency to 1000Hz, and use the receiver to obtain channel status information.

[0028] 3. First, preprocess the received raw data and use the LOF algorithm. If the obtained local anomaly factor LOF value is greater than 1, it wil...

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 stranger action recognition method based on channel state information. The stranger action recognition method comprises the following steps: S1, acquiring the channel state information; S2, preprocessing the channel state information obtained in the step S1; S3, inputting the preprocessed data into a convolutional neural network to extract features to obtain feature vectors; S4, sending the feature vectors into an SVM for secondary training to acquire a CNN-SVM combination model; and S5, in an offline identification stage, preprocessing the acquired to-be-identified data, substituting the preprocessed to-be-identified data into the CNN model to obtain a feature vector, and substituting the obtained feature vector into the trained model to perform action identification. The SVM is used for replacing softmax in the CNN, the training duration of the CNN-SVM combination model is relatively shorter, prediction speed is relatively higher, convergence speed is high,and recognition accuracy is high. Action recognition of more users can be realized only through training of a few users.

Description

technical field [0001] The invention relates to a stranger's action recognition method, in particular to a stranger's action recognition method based on channel state information and not trained in a system. Background technique [0002] Action recognition methods can be divided into sensor-based recognition, infrared-based recognition, vision-based recognition, sound wave-based recognition, and channel state information-based recognition. Among them, the recognition based on channel state information does not require users to wear additional equipment during data collection and does not violate user privacy, which is the most convenient method in today's action recognition methods. The identification based on channel state information benefits from the development of WiFi technology. In the beginning, WiFi was only invented to connect POS machines, and gradually entered the daily life of the public as a wireless network. Due to its characteristics of frequency bands that ...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06N3/04
CPCG06V40/20G06N3/045G06F18/2411G06F18/214
Inventor 吕继光杨武苘大鹏王巍玄世昌丁宁宁
Owner HARBIN ENG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products