基于模糊神经网络在智能轮椅避障中的应用

Based on fuzzy neural network for intelligent wheelchair avoiding obstacle

针对机器人在未知环境中的避障问题,提出了一种多传感器信息融合的避障方法。利用多传感器(声纳、摄像头)来采集外部环境信息,使得智能轮椅在移动过程中可以得到更充分的外部环境信息;使用基于Takagi-Sugeno(T-S)模型的模糊神经网络来对环境信息进行融合;通过融合的结果来控制轮椅的避障行为。通过模拟实验验证和分析,表明了该方法在解决轮椅避障问题方面有很好的效果,同时优化了轮椅避障的路径,提高了智能轮椅使用的安全性和方便性。

In the environment of unknown for robot obstacleavoidance problem, A multisensor information fusion of obstacle a voidance method. This method is to use multiple sensors （sonar, camera） to collect the external environment information, make intelligent wheelchair in mobile process can get more sufficient externaI environment information Then the fuzzy neural network based on the TakagiSugeno （TS） mode is used to fuse the information of the environment At last, we use the result of the in formation fusion to control the wheelchair＇s moving. Through the simulation experiment and analysis, show that the proposed method in solving wheelchair obstacle avoidance has very good effect, while optimizing the wheelchair obstacle avoidance path, improved the intelligence of the safety and convenience to use a wheelchair.