Abstract
Low cost "smart" servomotors widely used in robotics research present several control issues due to non-linearity at high torque and low performance controllers. Using publicly available physical simulation packages with PGS and Danztig solvers we can create very accurate faster than real-time models of a smart servomotor for use in control optimization. Simulation results are within 1.0 degree of the experimental results without external loading, and demonstrate identical non-linearity when the motor is driven above the torque limit.
