Obstacle Avoidacne Strategy Research Based On Fuzzy Neural Network For Omni Intelligent Wheelchair And Application In Tennis Doubles

Many countries have focused on intelligent wheelchairs with the purpose ofproviding superior transportation facility for the disabled to help them improve theirfreedom of action and re-integrate into society. The disabled desired to improve thequality of life, especially to participate in competitive sports, so the wheelchair shouldhave more intelligence functions to ensure that the disabled were safe in sports. Thisdissertation presented an obstacle avoidance strategy based on strategy conversionmechanism. The obstacle avoidance strategy combined obstacle avoidance algorithmbased on improved fuzzy neural network with obstacle avoidance algorithm based onstate control variables, and realized obstacle avoidance of omni-directionalwheelchairs. The main research works are listed as follows:1) Obstacle avoidance algorithm of wheelchair based on improved fuzzy neuralnetworkTo avoid selecting the parameters of the fuzzy logic algorithm according to theexperience of the designer in traditional method, an obstacle avoidance algorithmbased on improved fuzzy neural network was proposed, which can learn theparameters itself. The algorithm combined fuzzy logic with neural network accordingto their advantages. It trained fuzzy logic parameters by neural network. Consequently,fuzzy logic parameters were optimized and the effect of obstacle avoidance wasimproved.2) Omni-directional obstacle avoidance algorithm of wheelchair based on statecontrol variablesIn order to solve confusion of user’s desire direction and the direction producedby obstacle avoidance algorithm, a method based on state control variables waspresented. The algorithm recorded omni-wheelchair state of motion to solve theconfusion by state control variables. As a result, obstacle direction was close to theoriginal direction of the user. This method can meet user’s requirement for obstacleavoidance.3) Obstacle avoidance strategy of wheelchair tennis doubles based on strategyconversion mechanismIn order to avoid colliding between the wheelchair tennis doubles and increasingthe tacit understanding, an obstacle avoidance strategy based on conversionmechanism was presented. The strategy combined the algorithm based on improvedfuzzy neural network with the algorithm based on state control variables. Differentweights were assigned to avoid obstacles in wheelchair tennis doubles. Whenobstacles were in the field of view, the algorithm based on improved fuzzy neuralnetwork would play a leading role. On the contrary, the algorithm based on state control variables was primary obstacle avoidance algorithm. As a result, the strategycan improve the effect of obstacle avoidance.The experimental results showed that the proposed method can improve theobstacle avoidance effect of intelligent wheelchair and have practical applicationvalue.