Autonomous Navigation for Human-Following Robots Based on Optimized Position Tracking

In the field of human-robot interaction, achieving flexibility in following and companionship with humans in real-world environments presents numerous challenges. Traditional methods typically restrict robots to fixed positions relative to humans, such as tracking from behind, in front, or side-by-side, limiting their adaptability in various workspaces. This study introduces a novel navigation framework designed to enhance the flexibility of mobile robots in adjusting their tracking positions to accommodate workspace constraints. Initially, a virtual workspace is established, including the tracking point, combined with the human-robot interaction space to define a non-collision tracking space. An optimization function determines the optimal tracking position based on the relative distances among the tracking point, obstacles, the robot, and the human. Once the tracking point is identified, MPPI is used to help the robot follow the human through the optimal tracking point and avoid obstacles during movement. The method is implemented on ROS2 and validated in simulation environments using Gazebo and real-world experiments. Results demonstrate the efficacy of the method in enabling robots to follow and accompany humans with increased adaptability in limited and varying workspace environments. This flexible tracking capability allows the robot to adjust its path dynamically, maintaining optimal proximity to the human and enhancing interaction quality and operational efficiency in diverse scenarios.