Extending Skills of Autonomous Off-Road Robots on the Example of Behavior-Based Edge Compaction in a Road Construction Scenario

Autonomous construction robots operate in challenging environments and fulfill versatile tasks within frequently changing surroundings. To increase quality and productivity, the functionality of autonomous agents has to be continuously extended to achieve robust navigation and task-related skills. Unfortunately, the complexity of autonomous systems increases strongly along with their increasing number of application areas, while the overall maintainability decreases. In the area of road construction, tandem rollers are vehicles with a high potential for automation. This paper proposes an extensible and scalable framework for autonomous off-road robots on the example of road compaction tasks. The solution fosters reuse of components, enables abstraction, and clear interface definitions. New skills are added to the software framework and their performance is demonstrated within a simulated road construction scenario.