Swarm self-organized multi-heterogeneous target trapping based on the distribution of agent movement influence

Abstract

In existing swarm-based self-organized trapping algorithms, the majority of studies have concentrated on single-target or homogeneous multi-target scenarios within two-dimensional spaces, with limited attention paid to the spatial trapping of multiple heterogeneous targets in three-dimensional environments. The problem of enabling trapping agents to self-organize through mutual perception to trap multiple dispersed and evasive heterogeneous targets in three-dimensional space presents a novel and challenging research direction. To address this gap, this paper proposes a self-organized swarm trapping approach for multiple heterogeneous targets based on the distribution of agent movement influence. First, evaluation criteria for trapping success are established through the definition of nearest-distance and morphological indices. Then, a single-step movement distribution field is constructed based on the incremental motion of neighboring agents to represent the local movement tendencies of trapping agents. This field is used to infer the trapping requirements of heterogeneous targets and to develop a target selection strategy that integrates target demands, density fields, and inter-target distances. Subsequently, the trapping agents achieve a swarm-level self-organized trap in three-dimensional space through the combined effects of four virtual forces: density regulation, target attraction, inter-agent repulsion, and self-propulsion. The feasibility of the proposed method is demonstrated through both algorithmic analysis and simulation-based validation. Results show that even under conditions with minimal redundancy in the number of trapping agents, the proposed strategy can adaptively select heterogeneous targets and form stable trap configurations with high efficiency.