基于自适应模糊神经网络的下肢关节运动意图估计

Motion intention estimation of lower limb joint angles based on adaptive fuzzy neural network

利用神经网络和模糊推理技术建立了腿部关节角度预估模型。模型输入为处理后的人体下肢股直肌(VR)、股外侧肌(VL)和长伸肌(EP)的表面肌电信号,输出为髋、膝、踝三个关节角度预测值。在算法迭代更新过程中,采用混合策略和误差反向传播算法实现结构参数的自适应调整与优化。通过了数值仿真实验。

With neural networks and fuzzy inference techniques,an estimation model for lower limb joint angles is built in which the surface electromyography signals(sEMG)of human vastus rectus muscle(VR),vastus lateralis muscle(VL)and extensor pollicis longus(EP)are inputs,while the outputs are hip,knee and ankle joint angles estimations.In the iterative update process of the algorithm,both the hybrid strategy and error back propagation algorithm are adopted to adjust the parameters.Simulation is carried out.