As the world's aging trend is getting faster and faster, the quality of life of the elderly gradually into the public view. As an auxiliary tool of intelligent wheelchair has also been a wide range of attention and research at home and abroad. Intelligent wheelchair as a wheeled mobile robot, a combination of a number of areas of research technology, such as navigation obstacle avoidance technology, human-computer interaction technology. As an auxiliary tool for the elderly and the disabled, the smart wheelchair must have a more secure and reliable control system, as well as a more convenient and versatile mode of operation.
In this paper, a variety of methods to achieve intelligent wheelchair motion control system, navigation obstacle avoidance system and human-computer interaction interface control system. In this paper, we use an algorithm based on 3-spline trajectory planning to avoid obstacles. The trajectories planned according to the three splines have the characteristics of 3-order geometric continuity, which ensures smooth and smooth wheelchair movement. In order to realize the visual tracking and navigation function of intelligent wheelchair, this paper puts forward the algorithm of visual tracking and navigation avoidance based on Kinect camera, and has achieved some effect. In order to improve the diversity of intelligent wheelchair operation, this paper designs two kinds of control methods: brain movement control based on motion imagination and control mode based on speech recognition. In the EEG control interface based on motion imagination, this paper uses the CSP algorithm to extract the characteristics of EEG signals, and classifies the features with support vector machine (SVM) to identify human EEG signals , In response to people's "thinking" to control the purpose of the wheelchair. In the voice recognition control interface,
This paper presents a combination of speech detection technology and Google Speech speech recognition engine approach. The voice detection technique realizes the start and stop of the person's speech and uploads the valid voice data to the recognition engine for identification. The recognition result obtained is compared with the preset control instruction, and finally the purpose of controlling the wheelchair is achieved. Based on the above functional requirements, this paper has modified an electric wheelchair, designed a smart wheelchair prototype, and in a number of intelligent wheelchair verification experiments, including wheelchair obstacle avoidance experiment, visual tracking navigation experiment, EEG control experiment, voice Control experiments and so on. The experimental results show that the proposed algorithm and the proposed method have certain effect. The intelligent wheelchair designed in this paper has certain feasibility and practicability.