Collaborative Human-Robot Object Transportation Using a Deformable Sheet
Weijian Zhang, Charlie Street, Masoumeh Mansouri
AI summary
Problem
Collaborative object transportation in cluttered environments is hindered by humans' difficulty predicting robot behavior during obstacle crossings and the rigidity of existing formation methods. Current approaches often assume obstacle-free spaces or holonomic robots, limiting real-world deployment.
Approach
The authors propose a real-time planning system that defaults to human-led navigation but switches to robot-led mode when obstacles require crossing. During robot-led mode, it computes and projects a safe feasible region to guide the human while robots plan collision-free trajectories using homotopy-aware optimization.
Key results
- Multi-modal real-time formation planning framework with dynamic leadership switching
- Safe corridor construction and homotopy-aware trajectory optimization for obstacle avoidance
- Real-time feasible region computation and visual projection to guide human motion
- Successful validation in simulation and real-world hardware experiments
Why it matters
Enables safer, more flexible collaborative transportation in cluttered environments, advancing practical human-robot interaction for industrial and service automation.
Abstract
In this paper, we tackle real-time formation trajec- tory planning for collaborative object transportation in complex environments using a team of nonholonomic robots and a hu- man. The object is transported in a deformable sheet, and robots should follow the human’s lead while autonomously avoiding obstacles. By including a human in the formation, we leverage their adaptability and decision-making to improve transporta- tion. However, it can be difficult for a human to predict how autonomous robots will behave in complex situations, such as when the formation must cross an obstacle, i.e. where the object is transported above it. This could cause human decisions that compromise safety. To overcome these challenges, we introduce a multi-modal formation planning framework. By default the human leads the formation, and the robots plan to remain in the same homotopy class as the human to avoid collisions. If obstacle crossing is necessary the robots take the lead of the formation, where human motion is constrained to a feasible region projected visually in front of them. We demonstrate the efficacy of our framework in simulation and on hardware.