Neural network based optimal position/force control for constrained robot manipulators

In this paper the application of quadratic optimization and sliding mode approach is considered for hybrid position and force control of a robot manipulator. The dynamic model of the manipulator is transformed to a state-space model to contain two sets of state variables, where one describes the constrained motion and the other describes the unconstrained motion. The optimal feedback control law is derived solving matrix differential Riccati equation, which is obtained using Hamilton Jacobi Bellman optimization. The dynamic model uncertainties are compensated with a feedforward neural network. The FFNN requires no preliminary off-line training and is trained with on-line weight tuning algorithms that guarantee small errors and bounded control signals. The application of the derived control law is demonstrated through simulation with a two-arm robot manipulator to track a circular constrained surface while applying the desired force on the surface.