Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Sharing the Load: Distributed Model-Predictive Control for Precise Multi-Rover Cargo Transport

Alexander Krawciw, Sven Lilge, Luka Antonyshyn, Timothy D. Barfoot

TL;DR

Distributed MPC for cooperative cargo transport enables precise inter-robot spacing and path tracking in GNSS-denied environments by leveraging lidar Teach-and-Repeat localization and a shared map. The method compares distributed MPC, centralized MPC, and reactive distance-based controllers, showing distributed MPC matches centralized performance with added scalability; direct leader-to-follower distance measurements yield better short-range tracking but degrade with long-range offsets. Experiments on two- and three-robot convoys across varied terrains demonstrate centimeter-level spacing and robust real-time operation, with performance degrades modestly on rough terrain. The approach offers flexible convoy configurations and scalable deployment for practical autonomous cargo operations.

Abstract

For autonomous cargo transportation, teams of mobile robots can provide more operational flexibility than a single large robot. In these scenarios, precision in both inter-vehicle distance and path tracking is key. With this motivation, we develop a distributed model-predictive controller (MPC) for multi-vehicle cargo operations that builds on the precise path-tracking of lidar teach and repeat. To carry cargo, a following vehicle must maintain a Euclidean distance offset from a lead vehicle regardless of the path curvature. Our approach uses a shared map to localize the robots relative to each other without GNSS or direct observations. We compare our approach to a centralized MPC and a baseline approach that directly measures the inter-vehicle distance. The distributed MPC shows equivalent nominal performance to the more complex centralized MPC. Using a direct measurement of the relative distance between the leader and follower shows improved tracking performance in close-range scenarios but struggles with long-range offsets. The operational flexibility provided by distributing the computation makes it well suited for real deployments. We evaluate four types of convoyed path trackers with over 10 km of driving in a coupled convoy. With convoys of two and three rovers, the proposed distributed MPC method works in real-time to allow map-based convoying to maintain maximum spacing within 20 cm of the target in various conditions.

Sharing the Load: Distributed Model-Predictive Control for Precise Multi-Rover Cargo Transport

TL;DR

Distributed MPC for cooperative cargo transport enables precise inter-robot spacing and path tracking in GNSS-denied environments by leveraging lidar Teach-and-Repeat localization and a shared map. The method compares distributed MPC, centralized MPC, and reactive distance-based controllers, showing distributed MPC matches centralized performance with added scalability; direct leader-to-follower distance measurements yield better short-range tracking but degrade with long-range offsets. Experiments on two- and three-robot convoys across varied terrains demonstrate centimeter-level spacing and robust real-time operation, with performance degrades modestly on rough terrain. The approach offers flexible convoy configurations and scalable deployment for practical autonomous cargo operations.

Abstract

For autonomous cargo transportation, teams of mobile robots can provide more operational flexibility than a single large robot. In these scenarios, precision in both inter-vehicle distance and path tracking is key. With this motivation, we develop a distributed model-predictive controller (MPC) for multi-vehicle cargo operations that builds on the precise path-tracking of lidar teach and repeat. To carry cargo, a following vehicle must maintain a Euclidean distance offset from a lead vehicle regardless of the path curvature. Our approach uses a shared map to localize the robots relative to each other without GNSS or direct observations. We compare our approach to a centralized MPC and a baseline approach that directly measures the inter-vehicle distance. The distributed MPC shows equivalent nominal performance to the more complex centralized MPC. Using a direct measurement of the relative distance between the leader and follower shows improved tracking performance in close-range scenarios but struggles with long-range offsets. The operational flexibility provided by distributing the computation makes it well suited for real deployments. We evaluate four types of convoyed path trackers with over 10 km of driving in a coupled convoy. With convoys of two and three rovers, the proposed distributed MPC method works in real-time to allow map-based convoying to maintain maximum spacing within 20 cm of the target in various conditions.
Paper Structure (19 sections, 8 equations, 4 figures, 3 tables)

This paper contains 19 sections, 8 equations, 4 figures, 3 tables.

Table of Contents

  1. INTRODUCTION
  2. Related Work
  3. Methods
  4. System Overview
  5. Localization
  6. Reference Pose Planning
  7. Leader Reference Poses
  8. Follower Reference Poses
  9. Distance-Aware Pose Selection
  10. PI Feedback Controller Pose Selection
  11. Model-Predictive Control Architectures
  12. Distributed MPC
  13. Centralized MPC
  14. Independent Lateral and Longitudinal Control
  15. Scaling to $N$ Robots
  16. ...and 4 more sections

Figures (4)

  • Figure 1: A coupled robot convoy navigating the Woody Loop while cooperatively carrying cargo. A common set of tracks is clearly visible showing the two robots following each other within a few centimetres of error laterally and longitudinally.
  • Figure 2: The four controllers share similarities but are divided by how the inter-robot distance is measured and what variables are free for the MPC to optimize. All methods allow the steering angle to be controlled (steering wheel icon) but only some allow the linear velocity (gauge icon) to vary.
  • Figure 3: Startup separation as the robots start to move. The MPC controllers maintain close to the target of 1.5 m from the initial motion. The lighter colours show the velocity of the leader as it accelerates from rest.
  • Figure 4: A three-robot convoy repeating in unison using the Distributed MPC approach presented here. Extending the system to $N$ robots is simple with the distributed MPC.