Online Coverage Path Planning in LIDAR Perception Constrained Environments

Coverage Path Planning (CPP) is one of the vital factors in several fields where robots are being engaged. Online CPP is required for robots deployed in unknown environments. A proper perception system is crucial for a robot to perform autonomous navigation without colliding with obstacles. LIDAR is one of the prominent technologies utilized in recent path-planning research. However, having a limited range of perception and the lack of responsiveness against certain objects leads to issues when engaging CPP, which relies only on LIDAR. A secondary perception method can be integrated to overcome this limitation. This paper proposes an online CPP approach that occupies a LIDAR and a vision system that enhances perception range and obstacle recognition. The vision systems identify the specific set of obstacles: wires, carpet, floor socket outlets, and glass, which LIDAR can not perceive because of range and non-reflectiveness issues. The experimental results confirm the effectiveness of the proposed CPP to perform online coverage in an unknown environment while coping with obstacles that a LIDAR cannot perceive.