<![CDATA[In multi-robot formation control systems, trajectory planning plays a crucial role as it determines the ability of a formation to follow desired paths smoothly and accurately. This study presents a framework for Mecanum-wheeled robot teams, combining quintic-polynomial waypoint generation, a virtual-structure model, an artificial potential field for tracking and obstacle avoidance. The polynomial trajectories generate smooth references that satisfy position, velocity, and acceleration constraints at selected waypoints. Mecanum-wheel omnidirectional motion improves maneuverability in complex environments. The virtual rigid-body model assigns each robot a fixed formation offset to support coordinated movement. Artificial potential field methods attract robots toward their formation targets while repelling them from nearby robots and obstacles. The simulation results show that the generated waypoint trajectories are continuous and smooth. The robots successfully tracked their assigned formation positions while avoiding nearby obstacles. The mean Euclidean tracking errors between the desired and actual positions for Robots 1–3 were 0.268, 0.153, and 0.118 m, respectively. Additionally, three fixed circular obstacles with a diameter of 0.2 m were positioned at (3, 2), (5, 3.5), and (7, 2.5) m. During the simulation, the minimum distances from each robot's reference point to the center of its nearest obstacle were 0.642, 0.811, and 0.822 m, respectively. The inverse-kinematic transformation generated individual wheel angular velocities, which will be utilized for physical experiments in the next phase. The simulation results demonstrate the integrated performance of the trajectory generation, virtual-structure model, and artificial potential field, validating the approach prior to physical implementation.]]>
Copyrights © 2026
