Loosely Coupled Payload Transport System with Robot Replacement

Pulkit Verma, Rahul Tallamraju, Abhay Rawat, Subhasis Chand, Kamalakar Karlapalem

Abstract

In this work, we present an algorithm for robot replacement to increase the operational time of a multi-robot payload transport system. Our system comprises a group of nonholonomic wheeled mobile robots traversing on a known trajectory. We design a multi-robot system with loosely coupled robots that ensures the system lasts much longer than the battery life of an individual robot. A system-level optimization is presented to decide on the operational state (charging or discharging) of each robot in the system. The charging state implies that the robot is not in a formation and is kept on charge, whereas the discharging state implies that the robot is a part of the formation. Robot battery recharge hubs are present along the trajectory. Robots in the formation can be replaced at these hub locations with charged robots using a replacement mechanism. We showcase the efficacy of the proposed scheduling framework through simulations and experiments with real robots.

Introduction: The Battery Bottleneck

In cooperative robotics, tasks like payload transportation often require multiple robots to work in a tight formation. A significant hurdle in these systems is the "weakest link" problem: if a single robot’s battery dies, the entire formation stalls, and the mission fails. While existing research focuses on making robots more energy-efficient or finding optimal paths to chargers, few address how to maintain a continuous task while robots recharge.

A group of robots in formation carrying a payload.

The challenge is twofold. First, there is a scheduling problem: deciding which robot to swap and when, considering the future state of the fleet and the locations of recharge hubs. Second, there is a stability problem: replacing a robot mid-task without dropping or destabilizing the payload. This paper proposes a clever solution by treating a robot formation as a dynamic "tag-team" that swaps members at recharge hubs, extending the operational life of a transport system far beyond the battery capacity of any single unit.

Methodology

The researchers developed a two-part solution involving high-level optimization and low-level mechanical coordination.

Moving Horizon Optimization

The system uses a Quadratic Program (QP) to manage the fleet. Instead of just looking at the current state, it uses a "moving horizon" (looking two hubs ahead) to decide the optimal state (active or charging) for every robot.

  • Objective: Minimize total battery discharge while maximizing "robot retention" (minimizing unnecessary swaps that cause idle time).
  • Constraints: The solver ensures a fixed number of robots stay in formation, no robot's battery drops below a critical threshold, and swaps only occur if a charged robot is actually available at the current hub.

One key insight from this optimization is proactive swapping. The system often chooses to replace a robot even when it has 80-90% battery remaining. This ensures that robots at hubs are always being topped off, preventing a situation where multiple robots hit "empty" simultaneously.

The Replacement Mechanism

To physically swap robots without stopping the mission for long or risking the payload, the authors designed a "loosely coupled" system. Each robot is equipped with a linear actuator (piston) to lift the payload. By not mechanically "locking" the robots to the payload, the system gains the flexibility to swap components on the fly.

The 8-step sequence of a robot replacement within the formation.

As shown in the sequence above, when a swap is needed:

  1. The formation stops at a hub when a low battery state is detected.
  2. A dedicated "Support Robot" joins the formation to take the weight.
  3. The low-battery robot lowers its piston and leaves the formation.
  4. A fresh, charged robot enters the formation and raises its piston to support the payload.
  5. The support robot exits, and the formation continues its mission.

Experimental Results

The authors validated their approach using both simulations and physical hardware. They performed discharge tests on Li-Ion batteries under various loads (6kg to 18kg) to create empirical battery models. Evaluations were then conducted using a fleet of 12 robots in a Gazebo simulation and a physical experiment using 8 custom-built differential drive robots.

In real-world tests with a 6kg payload, the system's operational time increased from 25 minutes (the life of a single battery) to nearly 1 hour—a 140% increase limited only by the number of robots available at the hubs. In simulations, the optimization approach significantly outperformed "naive" baselines that only swap when a battery is nearly dead, maintaining higher average battery levels across the fleet and covering more distance.

Conclusion and Future Work

This work shifts the focus from individual robot efficiency to system-level resource management. By decoupling the task from individual battery limits, the system can operate indefinitely as long as charged robots are available.

Future directions include exploring "mobile" recharge hubs (tanker robots) to remove the constraint of stationary stations. Additionally, navigating uneven terrain would require more complex dynamics modeling beyond the flat, indoor environments tested here. Finally, moving the central optimizer to a fully distributed framework would enhance the system's robustness against communication failures.

Summary updated

Abstract

In this work, we present an algorithm for robot replacement to increase the operational time of a multi-robot payload transport system. Our system comprises a group of nonholonomic wheeled mobile robots traversing on a known trajectory. We design a multi-robot system with loosely coupled robots that ensures the system lasts much longer than the battery life of an individual robot. A system level optimization is presented, to decide on the operational state (charging or discharging) of each robot in the system. The charging state implies that the robot is not in a formation and is kept on charge whereas the discharging state implies that the robot is a part of the formation. Robot battery recharge hubs are present along the trajectory. Robots in the formation can be replaced at these hub locations with charged robots using a replacement mechanism. We showcase the efficacy of the proposed scheduling framework through simulations and experiments with real robots.
BibTeX
@misc{verma2019loosely,
  author = {Pulkit Verma and Rahul Tallamraju and Abhay Rawat and Subhasis Chand and Kamalakar Karlapalem},
  title = {{Loosely Coupled Payload Transport System with Robot Replacement}},
  year = {2019},
  eprint = {1904.03049},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1904.03049},
}
Figure