Abstract:Many new methodologies for the control of large-scale multi-agent systems are based on macroscopic representations of the emerging system dynamics, in the form of continuum approximations of large ensembles. These techniques, that are developed in the limit case of an infinite number of agents, are usually validated only through numerical simulations. In this paper, we introduce a mixed reality set-up for testing swarm robotics techniques, focusing on the macroscopic collective motion of robotic swarms. This hybrid apparatus combines both real differential drive robots and virtual agents to create a heterogeneous swarm of tunable size. We also extend continuification-based control methods for swarms to higher dimensions, and assess experimentally their validity in the new platform. Our study demonstrates the effectiveness of the platform for conducting large-scale swarm robotics experiments, and it contributes new theoretical insights into control algorithms exploiting continuification approaches.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper is mainly dedicated to solving two key problems: 1. **Experimental verification problems in large - scale multi - agent systems**: - Many new methods for controlling large - scale multi - agent systems rely on macroscopic representations, that is, under the assumption of an infinite number of agents, the collective behavior of the system is described by continuous approximation. However, these techniques are usually only verified by numerical simulation. - To address this limitation, the author introduced a mixed - reality platform, which combines real differential - drive robots and virtual agents to create a heterogeneous group with an adjustable scale. This hybrid device allows researchers to test swarm - robot technologies in a practical environment, especially their macroscopic collective motion. 2. **Extension and verification problems of high - dimensional continuous control methods**: - The author extended the continuous - based control method from one dimension to higher dimensions and experimentally verified its effectiveness on the new platform. Specifically, they derived the partial differential equations (PDE) that describe the emergent collective behavior of the swarm and designed a macroscopic control strategy to ensure that the swarm density converges to the expected value. - To achieve this, the author addressed the technical challenges from the one - dimensional case to the high - dimensional case, especially the new degrees of freedom emerging in the discretization process and the problem of ensuring well - definedness. ### Summary of specific content - **Mixed - reality environment**: The author constructed a mixed - reality environment that contains real robots and virtual agents. This environment can flexibly adjust the group size, avoiding the time and resource bottlenecks of large - scale experiments. - **High - dimensional continuous control**: The author proposed a theoretical framework to extend the one - dimensional continuous control method to higher dimensions and experimentally verified its effectiveness. The specific steps include: 1. **Continuousization**: Under the assumption of an infinite number of agents, reformulate the microscopic dynamics as a mass - conservation equation. 2. **Macroscopic control design**: Design the macroscopic control input so that the swarm density converges to the expected value. 3. **Discretization and microscopic control**: Discretize the macroscopic control input to generate specific control instructions for each agent. ### Experimental verification - **Experimental setup**: The author used four differential - drive robots and multiple virtual agents to conduct experiments through a mixed - reality platform. - **Experimental results**: The experimental results show that the proposed mixed - reality platform can effectively verify the high - dimensional continuous control method, and the swarm robots successfully achieved the desired density distribution. Through these efforts, the author not only demonstrated the potential of the mixed - reality platform in large - scale swarm - robot experiments but also provided new theoretical insights and technical implementations for high - dimensional continuous control methods.

Mixed Reality Environment and High-Dimensional Continuification Control for Swarm Robotics

High-dimensional continuification control of large-scale multi-agent systems under limited sensing and perturbations

A Continuification-Based Control Solution for Large-Scale Shepherding

Autonomous Swarm Robot Coordination via Mean-Field Control Embedding Multi-Agent Reinforcement Learning

Indirect Optimal Control of Advection-Diffusion Fields through Robotic Swarms

Indirect Swarm Control: Characterization and Analysis of Emergent Swarm Behaviors

Continuification control of large-scale multiagent systems under limited sensing and structural perturbations

Scalable Task-Driven Robotic Swarm Control via Collision Avoidance and Learning Mean-Field Control

Distributed and Decentralized Control and Task Allocation for Flexible Swarms

Spontaneous Emergence of Multitasking Robotic Swarms

Hierarchical Control of Smart Particle Swarms

Swarm Metaverse for Multi-Level Autonomy Using Digital Twins

Distributed Autonomous Swarm Formation for Dynamic Network Bridging

Human-guided Swarms: Impedance Control-inspired Influence in Virtual Reality Environments

Swarm-Enabling Technology for Multi-Robot Systems

Distributed control for geometric pattern formation of large-scale multirobot systems

Robust Multi-Robot Coverage of Unknown Environments using a Distributed Robot Swarm

Distributed reactive motion control for dense cooperative sweep coverage of corridor environments by swarms of non-holonomic robots

Decentralized Control of Quadrotor Swarms with End-to-end Deep Reinforcement Learning

Cognitive swarming in complex environments with attractor dynamics and oscillatory computing

MacroSwarm: A Field-based Compositional Framework for Swarm Programming