Intelligent wheelchair is designed to aid people in their daily life, especially for the disabled. Its work environment is mainly indoors. Intelligent wheelchair is expected to be able to transport its user or deliver items automatically. For example it can take the rider from one position to another required place by itself, or it is capable of picking up and delivering something as commanded by the user. Intelligent wheelchair is the application of intelligent mobile robots. They share similar research areas, including self-localisation, obstacle detection and avoidance, path planning. Now there are more and more researchers working on intelligent wheelchair and thus more and more new strategies have been proposed. In this thesis, some of the key issues of an intelligent wheelchair are investigated and some new algorithms are proposed. The contributions of this thesis include: (1) a novel Landmark Tree based robot self-localisation algorithm. This is a framework used to map the wheelchair's work environment. In this framework a robot can locate itself in a multi-scale as required by its user. In addition, based on Landmark Tree, finding the nearest path between two positions is mapped to linking two nodes in a tree, which makes the problem of path planning easy to be solved. (2) a vanishing point based obstacle avoidance strategy. Taking perspective effect into consideration, a robot only pays its attention to the real road when detecting and finding its path, which, as we know, reduces the influence of surrounding and irrelevant environment. Therefore this algorithm is more robust. Experiments are described to justify the algorithms.
修改评论