Gesture behavior recognition method not affected by individual wearing position and wearing style

Abstract

The invention provides a posture and behavior recognition method that is not affected by an individual's wearing position and manner, and belongs to the field of pattern recognition. By collecting the original acceleration and gyroscope data, then performing window length interception, and then performing feature extraction, the extracted features are mainly the overall characteristics of the three-axis acceleration data and the overall characteristics of the three-axis gyroscope data, and then the features are normalized. , and finally use a Gaussian process classifier for classification processing. Through the method of the present invention, when the sensor device is placed in any pocket of tight-fitting trousers in any direction, the three actions of walking, running, and standing can be accurately identified, and the system has low computational complexity and high applicability. widely.

Description

technical field [0001] The invention belongs to the field of pattern recognition and sensor technology, and in particular relates to a human body behavior recognition method that is independent of the position and direction of the trouser pocket where the sensor is placed. Background technique [0002] With the rapid development of sensor technology, the continuous improvement of sensor computing power and the continuous reduction of cost and size make it more and more a part of daily life to use these sensor devices to interact with people to improve the quality of life. In this field, sensor-based human behavior analysis and recognition has become a hot research topic in recent years because of its important role in medical care, security emergency, military and sports competitions. [0003] Now sensor-based human behavior recognition can generally be divided into two implementation methods, one is to use external sensors, and the other is to use wearable sensors. Externa...

AI Technical Summary

Problems solved by technology

[0004] However, the realization of human behavior recognition based on external sensors has relatively high requirements for the environment, and the sensor equipment needs to be fixed at a specific position in the environment. In practice, the human movement environment is constantly changing, and this method has great potential limitation

For wearable sensors, the current application generally requires the sensor to be fixed at a fixed position on the human body, so that the user's comfort level will be greatly reduced when using it, and it is necessary to ensure that the sensor is fixed on the user's body for a long time. Some parts are also difficult to achieve

For the realization of sensors in mobile phones, since people use mobile phones mainly for communication, entertainment and other applications that require handheld applications, it is unrealistic to continuously fix the mobile phone at a certain position on the user's body.

Images