Cooperative Target Enclosing Control of Multi-Agent Systems under Various Practical Constraints and Uncertainties

Abstract

Cooperative control of multi-agent systems (MASs) has drawn increasing interest in the control and robotics communities. Cooperative target enclosing control, in particular, has been a popular topic over the past two decades, owing to its wide range of applications, such as environmental monitoring and defense, spatial mapping and data acquisition, as well as support and escort operations. This control strategy, often addressed through distributed control protocols, aims to drive a group of mobile agents to track a static or moving target while achieving some user-defined formation pattern around it. Though considerable effort has been devoted to research on cooperative target enclosing control problems for various MASs, most of the existing studies are developed without full consideration of practical constraints and uncertainties, which are inevitable in real-world applications. Specifically, communication constraints, including communication delays and limited communication resources, often exist in practical systems and may lead to performance degradation or even instability. In practical scenarios, MASs are often subject to distance constraints due to limited interaction ranges and potential collisions between neighboring agents. Satisfying these distance constraints is crucial to achieving the objectives of cooperative control tasks and guaranteeing the safe deployment of MASs. Moreover, practical systems such as fixed-wing drones, differential drive ground robots, and underwater gliders are also often subject to nonholonomic constraints due to their inherent motion limitations, which significantly complicate the control process. In addition to these practical constraints, in applications such as patrol and escorting, target motion may not be known a priori, leading to target motion uncertainty that needs to be handled carefully.

This thesis aims to investigate the problems of cooperative target enclosing control for MASs under the aforementioned practical constraints and uncertainties. The main challenges include communication constraints caused by communication delays and limited communication resources, distance constraints raised from limited interaction ranges and collision avoidance, nonholonomic constraints from practical MASs, and uncertainties in the target motion. Robust and adaptive distributed control strategies are proposed to accommodate these complexities. The main results of this thesis are summarized as follows.

1. The static target enclosing control problem of MASs moving on a circle is investigated in the presence of both communication delays and distance constraints. To minimize the number of communication links, a novel distributed controller based on a cyclic pursuit strategy is developed in which each agent needs only its leading neighbor's information. In contrast to existing works, we propose a set of new potential functions to deal with communication delays and heterogeneous distance constraints simultaneously. A new framework based on the admissible upper bound of the formation error is established so that both connectivity maintenance and order preservation can be achieved at the same time. It is shown that the MAS can be driven to the desired formation asymptotically under the proposed controller.

2. The moving target enclosing control problem of MASs moving on a plane is investigated in the presence of uncertain target motion. Unlike existing studies, our approach handles an uncertain moving target whose trajectory is governed by an exosystem with unknown system dynamics and output functions. Specifically, an adaptive internal-model-based observer is developed to asymptotically estimate the target's velocity. Subsequently, a novel dynamic distributed control protocol that integrates prescribed performance control with a sliding-mode-inspired design is proposed. It is demonstrated that under this control scheme, the MAS achieves asymptotic tracking of the target and converges to the desired formation pattern characterized by specified circling radii and separation angles.

3. The moving target enclosing control problem of MASs is investigated with both uncertain target motion and limited communication resources. Compared with existing results, our method does not require the exact knowledge of target information and continuous monitoring among neighboring agents. In particular, a novel event-triggered adaptive internal-model-based observer is introduced to asymptotically estimate the target's position and velocity. To ensure a positive minimum inter-event time of each observer, a new event-triggering mechanism that incorporates three non-increasing resettable timer variables dependent on the observer states is proposed. In addition, a new distributed formation control law that only uses the event-triggered neighboring information is designed. It is shown that the MAS tracks the target's position and velocity asymptotically while achieving the prescribed formation pattern.

4. The moving target enclosing control problem of MASs is considered with both nonholonomic constraints and distance constraints. A novel control scheme is proposed to handle distance constraints arising from the need for network connectivity maintenance and collision avoidance. First, the requirements of the target enclosing problem are transformed into a distance-based formation framework. Compared with existing results, the newly constructed formation framework is guaranteed to be isostatic and does not require controlling the target's motion. Second, a fixed-time angular control law tailored to nonholonomic constraints using a barrier Lyapunov function is proposed. Third, a linear velocity control law is developed using the prescribed performance control approach and error transformation techniques. It is rigorously proved that our control laws enable the nonholonomic MAS to asymptotically achieve the desired angular formation pattern around a moving target while satisfying the established distance constraints.

These results lay a foundation and provide theoretical insights for advancing cooperative control strategies in more complex, uncertain, and resource-constrained target enclosing scenarios.

Date of Award	30 Jun 2025
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Lu LIU (Supervisor)