Vision-Based Cooperative Transport for Two Mobile Robots in Communication-Free Mapless Environments with Guaranteed Payload Safety

This letter addresses the vision-based cooperative transportation problem for two nonholonomic mobile robots in a leader-follower configuration, operating in obstacle-filled, mapless environments without inter-robot communication. Unlike most existing methods, the proposed approach considers key practical constraints, including the absence of maps, global localization systems, communication devices, and bounded robot velocities, as well as obstacle avoidance and payload protection. To address these challenges, a visual servo system on the follower continuously obtains relative pose measurements of the leader using a limited field-of-view (FOV) camera. A bounded navigation strategy guides the leader to the destination without maps, while the follower maintains a safe distance to protect the payload from stretching or ground collision. A bounded multi-objective control method enables the follower to maintain this distance while coordinating obstacle avoidance. To ensure safety, a recovery mechanism is activated when obstacles are encountered due to parameter issues or accidents. Throughout the process, no inter-robot communication is required. Finally, simulations and experiments validate the effectiveness of the proposed approach.