A distributed model predictive control (D-MPC) for modular robots in chain configuration

Abstract

This paper shows the design of a distributed predictive control, applied to a robotic chain of six modules. The dynamic model is obtained for a single module which is validated through the SimMechanics blockset of MATLAB. With the obtained model, a controller is designed using different predictive control strategies to choose the one that shows a better performance in terms of overshoot and settling time. Finally, the predictive control is distributed up to six modules, and its performance is evaluated by varying typical parameters of the predictive control such as the horizons of prediction, the control horizon, and the weights among others.