Optimal Decision Control for Robot with Series Compliance and Bounded Inputs

Position control of a manipulator involves the practical problem of solving the correct inputs to apply to the actuators order to follow a specified trajectory. This paper presents a robotic control scheme that accounts for a realistic model of the arm. This model includes the non-linearities inherent to the arm, the interaction between the drive motor and the arm coupled by harmonic drives and the possible saturation of the actuators. In the control scheme presented, the system is decomposed into two subsystems: the robot subsystem and the drive subsystem. A nonlinear control procedure is used to decouple the outputs of the arm subsystem. The decoupling control is designed so that the output of the robot follows the desired trajectory. The desired motor positions and velocities are obtained through the solution of the interaction equation, between the two subsystems. In order to control the actuators a pointwise optimal control is used. this control law minimizes, at each time, the error between the desired state velocity vector and the actual state vector of the motor subsystem.