非全自主控制与避障研究模式识别与智能系统专业论文.docxVIP

ABSTRACTABSTRACT ABSTRACT ABSTRACT With the aggravation of the senior citizen and the increasing handicapped persons，great attention is paid to the problem that how to improve these people’S mobility who have difficulty on walking and make them to be in harmony with society．Applying robot technology to electric wheelchair,intelligent wheelchair has just been raised recently．One of the research topics focus on avoiding obstacle automatically which is an integral part of the whole． On the basis of summarizing the development of the intelligent wheelchair,this thesis builds a hybrid architecture including reactive behavior control layer, deliberative behavior control layer,and supervision and management behavior control layer according to the general structure of Saphira intelligent control system design． For driving the wheelchair obtaining environmental information in unknown circumstances and realizing obstacle—avoidance，hardware circuit and software of multi-channel ultrasonic distance acquisition module is designed with DSP TMS320LF2407A in the thesis．And the least square method on the relative error is adopted to improve the accuracies of distance data collected． In the reactive layer,the thesis realizes designed T-S fuzzy neural network to avoid obstacles in unknown circumstances，and applies fuzzy control method to realize goal-·trending behavior．A decision-·making method of arbitration mechanism is proposed to cope、析tll decision problem in the course of obstacle avoidance．The validity of the algorithm is proved by the simulation results in Matlab environment． For the case that there are dynamic obstacles in the environment，the thesis adopts two different methods to predict obstacle’S reference position depending on their trajectory path．When the dynamic obstacle is in a similar linear movement，a linear regression model is used to predict the obstacle’S position based on Ordinary Least Squares Method．A Radial Basis Function Neural Network is used to build the pr