Loading...
Please wait, while we are loading the content...
Similar Documents
NEURAL NETWORK WITH VARIABLE TYPE CONNECTION WEIGHTS FOR AUTONOMOUS OBSTACLE AVOIDANCE ON A PROTOTYPE OF SIX-WHEEL TYPE INTELLIGENT WHEELCHAIR (2006)
| Content Provider | CiteSeerX |
|---|---|
| Author | Yasuda, Toshihiko Nakamura, Kazushi Kawahara, Akihiro Tanaka, Katsuyuki |
| Abstract | Abstract. In this paper, an assist method for human operation of electric-powered wheelchairs is studied. The purpose of this research is to make powered wheelchairs intelligent and to realize a mobility aid for people, who find it difficult or impossible to drive a conventional wheelchair. On a prototype of our group, a neural network produces an obstacle avoidance function. In this research, by the approach that connection weights of the neural network change according to the condition of obstacles in the vicinity of the wheelchair and the running state of the wheelchair, we improve the obstacle avoidance function. First, neural networks evolve by using digital computer simulator. Secondly, experiments, using a prototype with six wheels implemented neural networks whose connection weights are determined by numerical studies, demonstrate that the neural network with variable connection weights exhibits the excellent level of ability of obstacle avoidance. |
| File Format | |
| Publisher Date | 2006-01-01 |
| Access Restriction | Open |
| Subject Keyword | Six-wheel Type Intelligent Wheelchair Obstacle Avoidance Function Connection Weight Neural Network Change Electric-powered Wheelchair Variable Connection Weight Conventional Wheelchair Mobility Aid Powered Wheelchair Intelligent Numerical Study Excellent Level Assist Method Obstacle Avoidance Human Operation Digital Computer Simulator |
| Content Type | Text |
| Resource Type | Article |