The Research Of Intelligent Wheelchair Interaction And Motion Control Based On The Center Of Gravity Model

With the development of society, the wheelchair related sports for the disabled toparticipate is gradually popular. In the conventional wheelchair ball games, thedisabled athletes must rely on upper body strength to control the direction and speedof wheelchair while sports, and serve the ball at the same time. Therefore, it hashigher requirements for sports wheelchair in safety, controllability and efficiency.This research focused on human-computer interaction of the intelligent wheelchairsystem. The wheelchair control system based on gravity center of human body wasdesigned. This system can determine the user’s desired moving direction and speed byanalyzing the variations of the two-dimensional projection coordinates of human’scenter of gravity which was produced by body tilting on chair surface Therefore, thewheelchair motion can be controlled exactly and efficiently. The obtainedachievements or improvements in this dissertation are as follows.1) Direction control method based on improved fuzzy Kohonen clusteringalgorithmIn order to analyze the variations of the two-dimensional projection coordinates ofhuman’s gravity center on chair surface accurately, and determine the user’s movingintention, the fuzzy Kohonen clustering network was presented to cluster the gravitycenter data. Considering the influence of the abnormal data generated fromdestabilization on the clustering center, to eliminate misjudging the user’s intentionduring the motion control, a novel fuzzy Kohonen clustering network has beenimproved. Firstly, the lateral inhibition membership function was added into thissystem. Moreover, the classification accuracy can be achieved. Secondly, the learningrate of the network can be adjusted dynamically. The convergence rate was adapted inreal-time according to the membership degree of the data on the clustering nodes. Asresults, the system computational complexity has been optimized and the algorithmefficiency has been improved.2) Complex velocity control method based on gravity modelAccording to the interface of intelligent wheelchair based on gravity model, anexpert system combined with fuzzy PID controller was designed. First of all, theexpert knowledge and control rule set were established by the user’s intention andexperience. Thus the wheelchair speed was determined according to the occupant’stilting degree. Furthermore, the output of the expert system and current velocity ofwheelchair were used as the inputs of fuzzy PID controller. The fuzzy rules wereadopted to adjust PID parameters to trimmer the speed precisely. Consequently, thewheelchair velocity control system showed better dynamic performance and lowerovershoot. 3) Development and implementation of the interface software based on GUIFor convenience of controlling intelligent wheelchair in real time, and analyzingthe performance of the controller, the interface software with GUI component isdeveloped. Moving direction and speed control of the intelligent wheelchair based onGUI can be realized. Moreover, the algorithm parameter adjustment and systemperformance analysis modules were designed for the users. The software can displaythe data of pressure sensors, the state of controller operation, the human body centerof gravity, as well as, the velocity and trajectory of wheelchair. As results, the stabilityof wheelchair controller and the safety of users have been achieved.The wheelchair moving direction control algorithm and speed control algorithmwere conducted on the intelligent wheelchair hardware platform in terms of validityand accuracy. Firstly, according to the direction controller, the influence on clusteringresults caused by isolated points can be eliminated effectively by the improved FuzzyKohonen clustering algorithm. Furthermore, it can reduce the error between desireddirection and actual direction, and improve the control accurately. Secondly, thecomposite speed controller ensures that the system has improved the ability of thewheelchair in terms of its real-time performance, stability and precision.