RBF神经网络在柔性触觉传感器解耦中的应用

Application of RBF neural network in decoupling research of flexible tactile sensor

针对柔性触觉传感器模型高度非线性、解耦难度大等问题,提出一种有效的方法来模拟柔性触觉传感器在实际应用中含噪声的情形。首先在理想条件下的传感器模型上添加不同幅度的高斯白噪声并建立其数学模型,之后通过K-均值和递归最小二乘法优化RBF神经网络,并利用优化后的RBF神经网络算法逼近受噪声干扰的传感器阻值与形变之间的高维非线性映射关系,最后基于不同的展开幅度通过行列阻值解耦出传感器三维形变信息,获得了较好的解耦精度。解耦结果表明,RBF神经网络算法具有较强的鲁棒性和抗噪声能力,能够很好地逼近含噪声的传感器高维非线性数据之间的映射关系。

In view of the high nonlinearity and difficulty of decoupling process of the flexible tactile sensor, an effective method is proposed to simulate the situation of a flexible tactile sensor interfered by noises in practical application. Firstly, different white Gaussian noises are added into the ideal tactile sensor model, and its mathematical model is established. Then, the K-means and recursive least squares methods are used to optimize the Radial Basis Function Neural Network （RBFNN）. The optimized RBF neural network algorithm is used to approximate the high dimensional nonlinear mapping relationship between the sensor resistance interfered by noise and deformation. The three-dimensional deformation of the sensor is decoupled by the row-column resistance based on different spreads. The decoupling results show that the RBFNN with strong robustness and anti-noise ability has good performance in approximate the highly nonlinear relationship between the sensor variables interfered by noises.