In this paper, a reference tracking problem for wheeled differential-drive robots subject to obstacle avoidance requirements is considered. By resorting to feedbacklinearization techniques, the differential-drive nonlinear dynamics is first recast as two simple integrators under statedependent input constraints; then, an ad-hoc switching receding horizon controller is derived with the aim to efficiently handle time-varying constraints and non-convex obstacle-free regions. Effectiveness and benefits of such an approach are finally discussed by means of experimental results obtained using an E-puck2 differential-drive robot. 

C. Tiriolo, G. Franze', W. Lucia, IEEE Conference on Control Technology and Applications (CCTA), DOI: 10.1109/CCTA41146.2020.9206301, 2020