Area Division Using Affinity Propagation for Multi-Robot Coverage Path Planning

In the wake of advancing technology, autonomous vehicles and robotic systems have burgeoned in popularity across a spectrum of applications ranging from mapping and agriculture to reconnaissance missions. These practical implementations have brought to light an array of scientific challenges, a crucial one among them being Coverage Path Planning (CPP). CPP, the strategic planning of a path that ensures comprehensive coverage of a defined area, while being widely examined in the context of a single-robot system, has found its complexity magnified in the multi-robot scenario. A prime hurdle in multi-robot CPP is the division and allocation of the operation area among the robots. Traditional methods, largely reliant on the number of robots and their initial positions to segment the space, often culminate in suboptimal area division. This deficiency can occasionally render the problem unsolvable due to the sensitivity of most area division algorithms to the robots' starting points. Addressing this predicament, our research introduced an innovative methodology that employs Affinity Propagation (AP) for area allocation in multi-robot CPP. In our approach, the area is partitioned into 'n' clusters through AP, with each cluster subsequently assigned to a robot. Although the model operates under the assumption of an unlimited robot count, it offers flexibility during execution, allowing the user to modify the AP algorithm's similarity function factor to regulate the number of generated clusters. Serving as a significant progression in multi-robot CPP, the proposed model provides an innovative approach to area division and path optimization, thereby setting a strong foundation for future exploration and practical enhancements in this field.