This study presents a decentralized initial configuration algorithm for cooperative object transportation using swarm robots. The proposed algorithm enables each robot to determine whether to participate in transport based on locally sensed load and the light color signals emitted by nearby robots. This approach avoids the need for centralized coordination or complex communication infrastructure, improving system scalability and robustness. To evaluate the effectiveness of the algorithm, simulations were conducted using ROS2 and Gazebo with modified TurtleBot3 Waffle Pi models. Two types of transport objects were tested: one with a uniformly distributed mass and another with an off-center center of mass. Results showed that in most cases, the robots autonomously formed a stable configuration, with a 90% success rate for the uniform object and 70% for the unbalanced one. The number of participating robots dynamically adjusted according to the local load distribution. These findings demonstrate that swarm robots can achieve adaptive and stable transport through local interaction.
Tanibuchi et al. (Wed,) studied this question.