Towards trajectory planning and collision avoidance for the voyager unmanned ground vehicle in dynamic environments

Department of Electrical and Electronic Engineering, Stellenbosch University, South Africa

^* Corresponding author: 23676787@sun.ac.za

Abstract

This paper presents a Model Predictive Contouring Control (MPCC) framework for trajectory planning on the CSIR Voyager differential-drive UGV, scoped to structured racetrack environments. The system unifies path following and short-horizon planning using a linear-spline track representation and curvilinear progress dynamics, targeting high progress with low contouring and lag error. Dynamic obstacles are assumed to have known, broadcast trajectories; collision avoidance is formulated either by penalizing vehicle-to-obstacle distance in the objective function or by enforcing minimum-distance inequality constraints over the prediction horizon. The framework is validated in MATLAB simulations. Real-time feasibility depends strongly on horizon selection: smaller prediction/control horizons achieve >10 Hz closed-loop rates, whereas longer horizons that improve lookahead can exceed the 10 Hz budget. The work establishes a baseline in simualtion for integrating dynamic-obstacle avoidance under known trajectories onto the Voyager platform; future efforts will progress toward onboard perception-driven obstacle prediction and constraints before deployment on the physical Voyager platform.