A real-time safe path planning system for cooperative robots

Abstract

Background: A cooperative robot is a robot requested to co-work with humans efficiently and safely in an environment with flexible arrangements. Safe path planning is a crucial issue which must be resolved during human-robot cooperation. In this paper, we present a safe path planning system that could plan the manipulation path in real-time based on the environmental changes and guarantees safety when the robot interacts with the environment and humans. Methods: In this system, we first build a real-time obstacle Octomap from the environment RGB-D (red green blue-depth) images, which can effectively differentiate the robot from other obstacles in the environment and eliminate the robots influence during the map building. And then, we adopt the rapidly exploring random trees-Connect method to plan the safe path in the Octomap. When the planning path is obstructed by the dynamic objects, the system will re-plan the new safe path based on the changed Octomap. Results: The experimental results show that our system can effectively avoid obstacles in a dynamic environment and safely reach the manipulation destination. Conclusions: We propose a real-time safe path planning system for cooperative robots, which can guarantee the safety of manipulation.