Intelligent Wheelchair Control Based On Sitting Posture And The Fusion Of Multi Sensors

Intelligent wheelchair is part of service robot family, the importance of this study is to help mobility-impaired elderly or disabled better integrate into society and improve their quality of life. With robot technology, embedded technology, artificial intelligence and other fields of development and study of intelligent wheelchair, absorbing a variety of technologies, such as multi-sensor information fusion, pattern recognition, and so on. Its performance continues to increase, and features a more diverse and powerful. This paper analyzes the present situation and the development of related technologies for intelligent wheelchair, based on users is based on research posture of intelligent wheelchair control, presented using support vector machine to treat pressure sensor data position detection method and study on multi-sensor information fusion, final design of human-machine interface based on Android platform.Firstly, from a functional perspective, modular hardware structure analysis of wheelchair designed obstacle range, position detection module. Presents posture recognition method based on Gaussian kernel support vector machines. First Tekscan distributed diaphragm pressure sensors on the wheelchair seat and backrest pressure distribution detection, pressure-sensitive distribution and statistical methods, and provides scientific basis for installing low cost pressure sensors, pressure sensor data is more accurate; after extracting characteristics of sitting, learning with Gaussian radial basis function based SVM algorithm pressure data can identify the users sitting_posture.Study on the BP neural network multisensory information fusion used bearing partition algorithm for ultrasonic sensor data pretreatment and fuzzy distance information. L-m optimization algorithm for neural network training is used to improve the convergence speed. Training of neural networks can identify obstruction category.Posture recognition based on multi-sensor information environment categories based on three-dimension fuzzy controller is designed, according to the use by wheelchair users sit and autonomous obstacle avoidance obstacle Environment category; the last was developed based on the Android platform wheelchair operation makes human-computer interaction interface simple, practical, friendly, and convenient for functional integration modules and extensions.The above research, improves safety, intelligent, friendly and intelligent wheelchairs. Especially due to physical discomfort to users who maintain normal posture for a long time, the intelligent wheelchair obstacle, turns according to the user sat output users safe, comfortable speed and angle. This study also provide posture detection and multi-sensor information fusion with new ideas and methods.