Nonlinear Optimization-Based Real-Time Trajectory Planning for Cooperative Transport of Heterogeneous Mobile Manipulators

Goal :

1. Collaborative Path Planning Optimization
   • Integrate real-time path generation for the leader and nonlinear optimization-based path planning for the follower within a leader-follower structure. The proposed framework aims to generate optimal paths that simultaneously consider inter-robot distance maintenance, geometric consistency, and obstacle avoidance.
2. Nonlinear Optimization for Follower Path Generation
   • Design an algorithm to generate the optimal path for the follower robot in real-time based on the leader’s trajectory. This approach optimizes distance constraints, geometric consistency, and obstacle avoidance through a multi-objective cost function, addressing the trade-offs between these conflicting objectives.

Summary This study presents a Nonlinear Optimization-Based Real-Time Trajectory Planning Framework for cooperative transport of heterogeneous mobile manipulators. The framework addresses the challenges of collaborative control in environments with varying robot structures and kinematics. The key components of this research are as follows:

1. Leader-Follower Architecture
   • The system adopts a leader-follower structure where the leader (MOBY) plans its trajectory using the Timed-Elastic Band (TEB) approach, while the follower (MOCA) generates its path based on the leader’s trajectory and local obstacle information collected via its LiDAR sensors.
   • The follower uses nonlinear optimization to ensure safe distance maintenance, reference path tracking, and obstacle avoidance, enabling stable transport in complex environments.
2. Nonlinear Path Optimization for the Follower
   • The follower’s path is generated by laterally offsetting the leader’s trajectory along a smoothed normal direction and then refined through nonlinear optimization.
   • The optimization process includes objective functions for distance maintenance, reference path tracking, and obstacle avoidance, with carefully tuned weights to balance these conflicting goals.
3. Whole-Body Control using GHC
   • The follower robot is controlled using a Generalized Hierarchical Control (GHC) structure, integrating impedance control for the manipulator and admittance control for the mobile base.
   • This approach allows the follower to respond dynamically to environmental changes while maintaining stable cooperative transport.
4. Experimental Validation
   • The proposed framework was validated through experiments involving two heterogeneous mobile manipulators (MOBY and MOCA) in a cluttered environment. The results demonstrated successful trajectory tracking with an average tracking error of 0.037m and a standard deviation of 0.039m, while maintaining a safe inter-robot distance (1.05–1.65m) during 96.09% of the transport time.