Control strategies for robotic manipulators

The dynamic end-point positioning problem for robotic manipulators is considered. A review of the current approaches to the problem is presented with a critical examination of each technique. A close inspection of the structure of a typical manipulator model is performed to investigate the suitability of the various control approaches. In particular, the PUMA 5 60 Industrial Manipulator is considered. The advantages and disadvantages of various techniques, including linear and non-linear, fixed and adaptive methods are given and a current approach, based on non-linear self-tuning theory is presented. An attempt is made to give some appreciation of the relative computational complexity of various algorithms and a number of possible hardware architectures for control are discussed. Finally, conclusions regarding the choice of dynamic control algorithms for manipulators are drawn