Human-Force-Based Imitation Learning Enabling Mobile Robotic Carts to Avoid Obstacles

Recently, the demand for transport-support systems that provide physical assistance to humans in carrying out heavy tasks, such as transportation, has surged, and mobile robots with high mobility performance have been studied. However, autonomous movement control and obstacles on the path from the current position to the target position hamper the adaptability of these robots in the real world. In this paper, we propose a machine learning based autonomous control method for a mobile robotic cart to avoid obstacles while moving to the target position. We train a neural network using distance data to the obstacle and human operation information. The proposed method is validated via experiments on obstacle avoidance