Cooperative control and trajectory planning of the multiple mobile vehicles via distributed model predictive control to implement the following tasks including formation control, inter-vehicle obstacle avoidance and environment obstacle avoidance. Stability, feasibility and optimality must be guaranteed. For real test, these codes will be deployed in the three Raspberry PI 3. A cameral and AprilTags visual localization system are used for localisation. Codes and demos are not complete and are debugged and kept updating.
Centralized structures usually leads to large optimization problem, which is time-consuming. The second graph below shows the case in which three vehicles are moving in the formation of a triangle.
Decentralized structures contributes to faster solving of the optimization problem. The trajectories of MPC are shown with the horizon 10
The three vehicles form a group as a triangle and break the form due to the emergency.
Three raspberry pi 4 are used for running the main programs while a Realsense D435 RGBD cameral and AprilTag system are used for localization.
The following animation shows the MPC implementation for obstacle avoidance, which is the important fundation for DMPC of the multiple vehicles.
The following animation shows the DMPC implementation for two vehicles to avoid inter-vehicles collision and approach their destinations.
The following animations show the DMPC implemetation for three vehicles to avoid inter-vehicles collision and approach their destinations.