Hierarchical Topology Map with Explicit Corridor for global path planning of mobile robots

One way to construct a global path for a mobile robot is to extract the equidistant points from obstacles in an occupancy grid map (OGM) as its backbone of thin lines (skeleton) and find an appropriate path from them. Such skeleton-based methods are advantageous as they provide complete path planning, i.e., a feasible path is guaranteed to be found if it exists and often allow safe distances from obstacles with light computation. However, a skeleton path needs to be properly refined to avoid excessive bypasses, which inherently exist, and get an optimal path. This paper introduces hierarchical topology map with explicit corridor (HTM-EC) as a new skeleton-based strategy for global path planning with path optimization. HTM-EC is a skeleton-based hierarchical graph structure with a topology graph and the corresponding skeleton of a map, which can be an alternative map expression of an OGM. A skeleton path can be quickly obtained from the topology graph of HTM-EC as needed with the light computation facilitated by the significantly reduced size of searching space. The novelty of the paper lies on, first, the formulation of HTM-EC structure, and second, the refinement method of the given skeleton path with cost optimization by using corridor discretization and dynamic programming. For cost optimization, we incorporate the allowable speed that limits the navigation speed in narrow areas for safe operation, which enables a time-optimal path with the cost of traveling time. Simulations and experiments were conducted with real OGMs and a mobile robot to validate the proposed global planning method.