Comparison of Kinematic and Dynamic Model Based Linear Model Predictive Control of Non-Holonomic Robot for Trajectory Tracking: Critical Trade-offs Addressed

Kaustav Mondal, Armando A. Rodriguez, Sai Sravan Manne, Nirangkush Das, and Brent Wallace


Non-holonomic System, Model Predictive Control, Convex Optimization, Trajectory Tracking Control, Quadratic Programming, Mobile Robot


This paper presents a hierarchical inner-outer control structure with model predictive controller (MPC) in the outer-loop and a PI controller in the inner loop, to perform trajectory tracking on non-holonomic robots. Two different MPC formulations are considered: (1) kinematic model based MPC which assumes that inner-loop has infinite bandwidth (2) dynamic model based MPC which takes into consideration the bandwidth limitations imposed by inner-loop. In order to emphasize the importance of dynamic model based MPC over kinematic model based MPC, critical tradeoffs involving tracking errors vs inner-loop bandwidth, for varying reference velocities, are studied. The novelty of this paper lies in the systematic approach taken to answer: (1) when is a kinematic model based MPC sufficient, (2) when is a dynamic model based MPC necessary, to obtain good trajectory tracking properties. Both, simulation and hardware results are taken into consideration.

Important Links:

Go Back