[Paper Review] Particle Computation: Designing Worlds to Control Robot Swarms with only Global Signals

Aaron T. Becker, Erik D. Demaine+2 more|arXiv (Cornell University)|Feb 16, 2014

Modular Robots and Swarm Intelligence17 references4 citations

TL;DR

This paper introduces a framework for controlling large swarms of simple robots using only global force-field signals, by designing obstacle configurations that enable complex computation and reconfiguration. It proves the motion-planning problem is PSPACE-complete and demonstrates practical implementations of matrix permutations, absolute encoders, and dual-rail logic gates using a reconfigurable hardware platform with gravity-driven particle movement.

ABSTRACT

Micro- and nanorobots are often controlled by global input signals, such as an electromagnetic or gravitational field. These fields move each robot maximally until it hits a stationary obstacle or another stationary robot. This paper investigates 2D motion-planning complexity for large swarms of simple mobile robots (such as bacteria, sensors, or smart building material). In previous work we proved it is NP-hard to decide whether a given initial configuration can be transformed into a desired target configuration; in this paper we prove a stronger result: the problem of finding an optimal control sequence is PSPACE-complete. On the positive side, we show we can build useful systems by designing obstacles. We present a reconfigurable hardware platform and demonstrate how to form arbitrary permutations and build a compact absolute encoder. We then take the same platform and use dual-rail logic to build a universal logic gate that concurrently evaluates AND, NAND, NOR and OR operations. Using many of these gates and appropriate interconnects we can evaluate any logical expression.

Motivation & Objective

• To address the challenge of controlling large swarms of micro- or nanorobots using only global input signals, such as electromagnetic or gravitational fields.
• To investigate the computational complexity of motion planning for such swarms under uniform control inputs.
• To design artificial environments with obstacles that enable complex tasks like matrix permutations and Boolean logic computation.
• To demonstrate practical implementations of these systems using a reconfigurable hardware platform with gravity-driven particle movement.

Proposed method

• Uses a 2D grid model where all robots receive the same global command (e.g., move up, right, down, left) and move until blocked by obstacles or other robots.
• Designs obstacle configurations that act as one-way valves, latches, and logic gates to control robot flow and enable computation.
• Employs dual-rail logic, where each logical value is represented by two particles, enabling concurrent evaluation of AND, NAND, OR, and NOR operations.
• Constructs matrix permutation systems using four global commands to rearrange any number of robots into arbitrary target configurations.
• Develops reversible logic circuits using conservative logic principles, ensuring no particles are created or destroyed.
• Validates the approach with a physical hardware prototype using steel and nylon bearings on a tiltable MDF grid, controlled via gravity.

Experimental results

Research questions

• RQ1 What is the computational complexity of motion planning for robot swarms controlled by global signals?
• RQ2 Can artificial obstacle configurations enable arbitrary reconfiguration of robot swarms using only global commands?
• RQ3 Can such obstacle designs support universal computation, such as Boolean logic operations?
• RQ4 How can reliable, error-free computation be achieved in a system with limited individual control and potential for particle misalignment?
• RQ5 What are the practical limitations and design constraints for implementing such systems in real-world micro-robotic environments?

Key findings

• The problem of determining whether a given initial robot configuration can be transformed into a desired target configuration using global commands is proven to be PSPACE-complete.
• Matrix permutations of any size can be achieved in just four global commands, enabling efficient reconfiguration of robot swarms.
• A physical hardware platform using gravity-driven steel and nylon bearings successfully demonstrates matrix permutations and dual-rail logic gates.
• Dual-rail logic gates were implemented to concurrently compute AND, NAND, OR, and NOR operations using the same command sequence.
• The prototype demonstrated reliable operation in enclosed maze-like logic circuits, though long open paths in permutation systems were prone to errors from particles deviating from path.
• The system enables the construction of an absolute encoder and a universal logic gate, demonstrating that engineered environments can perform complex computation with minimal control input.