Inverse kinematics Solution based on IPSO-RBF Neural network

Solving inverse kinematics of concrete pump truck boom system is inefficient and sometimes may have the singular solution. To handle this, this paper proposed an improved particle swarm optimization-radial basis function neural network(IPSO-RBFNN) algorithm. First, we simplified the boom system of the concrete pump truck as a four-degree-of-freedom serial robot and established the forward kinematics model of the robot. Then, we used the IPSO algorithm to generate training and testing samples and search the optimal parameters of the RBF network by iteration. After that, we used the optimized RBFNN to solve inverse kinematics. Finally, the simulation shows that IPSO-RBFNN can quickly solve the inverse kinematics of the points on a straight pouring path. The experimental results show that the error of IPSO-RBFNN is reduced by 68% compared with the traditional RBFNN and 35% compared with the PSO-RBFNN.