Hybrid computational intelligence based kinematic analysis for parallel machine tool

Compared with serial Machine Center (MC), the Parallel Machine Tools (PMTs) have the advantages of high rigidity, low cost, high accuracy, high velocity, high acceleration and especially suitable for free surface cutting. The main kinematic part in PMTs is Parallel Manipulator (PM). PM includes a fixed base, a platform and several kinematic chains connected with fixed base and platform. The parallel structure characteristics of PM lead to non-linear between actuators and platform, and results in difficulty and complexity in PMT design and PMT control algorithm design. Now the popular method for kinematic analysis includes two steps: non-linear modeling and non-linear equations solving, some of which are difficult and inefficient. A hybrid computational intelligent method is proposed for kinematic analysis of PMTs. The approach includes three main steps: 1) artificial neural network (ANN) models and parameters selection; 2) artificial neural network learning and training; 3) kinematic analysis. The approach overcomes the shortages of ANN. The main processes and key techniques are illuminated for kinematic analysis of PMTs with several analysis samples finally.