In this chapter, the physically constrained RMP scheme is applied to the PA10 robot manipulator by using two QP solvers, i.e., the LVI-PDNN and DNN. Different desired motion paths of the PA10 end-effector (i.e., circular and straight-line paths) are tested for illustrative and comparative purposes.
- 2.Wang J, Zhang Y (2004) Recurrent neural networks for real-time computation of inverse kinematics of redundant manipulators. In: Sincak P, Vascak J, Hirota K (eds) Machine intelligence: quo vadis? World Scientific, Singapore, pp 299–319 Google Scholar