Avoidance Driving Method Through Costmap Reflection Using Path Information Between Autonomous Driving Robots

There is a need for traffic management technology to prevent driving deadlocks between robots performing irregular missions in unstructured spaces. We propose a method to safely avoid neighboring robots by sharing driving information through inter-robot communication. The distribution of surrounding robots can be known even before it is detected by sensors such as lidar, and gives a cost to their expected path. It was implemented and tested in the multi-robot-layer costmap, and shows that it reduces the deadlock problem.