基于自适应模糊逻辑和神经网络的双足机器人控制研究

Study on Biped Robot Control Based on Adaptive Fuzzy Logic and Neural Network

在双足机器人行走控制中 ,为了改善系统的行走性能 ,提出了一种基于RBF神经网络前馈控制的力矩补偿控制方法。该方法将自适应模糊控制和神经网络逆模控制有效地结合起来 ,利用神经网络来逼近系统的逆动力学模型 ,提高系统了的控制性能 。

In study on biped robot control, in order to improve the locomotion, a novel torque compensation control scheme based on the combination of adaptive fuzzy logic and neural network is proposed in this paper. This method combines the advantage of adaptive fuzzy logic and neural network. Although the accurate structure and parameters of biped robot are not known, the inverse dynamic model can be approximated. The resulting control scheme has improved performance without increase of complexity. The biped robot has a good walking property.