VR-supported method for studying the MPC algorithm in controlling snake robot motion

The snake robot is suited for difficult terrain and limited spaces for searches, rescue, exploration, and medical procedures. However, such underactuation makes their control and locomotion quite challenging. Current methods mostly lack real-time adaptability, optimisation under constraints, and effective visualization tools. This work contributes toward filling some of the gaps in the control and locomotion of under-actuated snake robots through the use of model predictive control and virtual reality. Five control strategies were implemented and tested using simulations in MATLAB and Unity with a three-degree-of-freedom snake robot model. Simulations included ‘ghost robots’ to visualize predicted and actual trajectories under different conditions. Proposed techniques improve the navigation, and more complex algorithms provide better accuracy and robustness. VR is proving to be an important tool in optimizing robot performance. Integrating advanced control algorithms together with VR increases the effectiveness of robotic systems development. Future work will be concerned with cluttered environments and more complex interaction models.