Decentralized Shepherding of Non-Cohesive Swarms Through Cluttered Environments via Deep Reinforcement Learning

TL;DR

This paper presents a decentralized reinforcement learning approach for shepherding non-cohesive swarms through cluttered environments, enabling scalable and collision-free navigation without retraining in complex scenarios.

Contribution

It introduces a hierarchical control architecture combining target assignment with a learned low-level steering policy that generalizes from single-target to multi-target environments.

Findings

Smooth, collision-free trajectories achieved in simulations

Method generalizes to multiple obstacles without retraining

Consistent convergence to goal regions demonstrated

Abstract

This paper investigates decentralized shepherding in cluttered environments, where a limited number of herders must guide a larger group of non-cohesive, diffusive targets toward a goal region in the presence of static obstacles. A hierarchical control architecture is proposed, integrating a high-level target assignment rule, where each herder is paired with a selected target, with a learning-based low-level driving module that enables effective steering of the assigned target. The low-level policy is trained in a one-herder-one-target scenario with a rectangular obstacle using Proximal Policy Optimization and then directly extended to multi-agent settings with multiple obstacles without requiring retraining. Numerical simulations demonstrate smooth, collision-free trajectories and consistent convergence to the goal region, highlighting the potential of reinforcement learning for scalable, model-free shepherding in complex environments.