Riqing Chen

Estimator-based adaptive neural network control of leader-follower high-order nonlinear multiagent systems with actuator faults

The problem of distributed cooperative control for networked multiagent systems is investigated in this paper. Each agent is modeled as an uncertain nonlinear high‐order system incorporating with model uncertainty, unknown external disturbance, and actuator fault. The communication network between followers can be an undirected or a directed graph, and only some of the follower agents can obtain the commands from the leader. To develop the distributed cooperative control algorithm, a prefilter is designed, which can derive the state‐space representation to a newly constructed plant. Then, a set of distributed adaptive neural network controllers are designed by making certain modifications on traditional backstepping techniques with the aid of adaptive control, neural network control, and a second‐order sliding mode estimator. Rigorous proving procedures are provided, which show that uniform ultimate boundedness of all the tracking errors can be achieved in a networked multiagent system. Finally, a numerical simulation is carried out to evaluate the theoretical results.

Finite-time fault-tolerant coordination control for multiple Euler–Lagrange systems in obstacle environments

Abstract In obstacle environments, the problem of coordination tracking control for a team of Euler–Lagrange systems is investigated under modeling uncertainties, actuator faults and disturbances. First, syncretizing the Null-Space-based Behavioral (NSB) control, graph theory and finite-time control method, a novel desired velocity is predesigned to achieve the finite-time obstacle avoidance and coordination tracking. Then a set of finite-time fault-tolerant coordination control laws (FFCCLs) are presented to guarantee all of the agents to track a dynamic target while avoiding obstacles/collisions. To improve the robustness and control accuracy of the systems, an adaptive control gain is incorporated into the FFCCL so that the derived algorithm can be implemented without manual parameter adjustment. Both of the control architectures are distributed, model-independent and robust with respect to modeling uncertainties, actuator faults and disturbances. Finally, several numerical simulations are presented to demonstrate the efficacy of the control strategies, showing that the overall motion of the two tasks can be accomplished satisfactorily with high precision.

Estimator‐based adaptive neural network control of leader‐follower high‐order nonlinear multiagent systems with actuator faults

The problem of distributed cooperative control for networked multiagent systems is investigated in this paper. Each agent is modeled as an uncertain nonlinear high‐order system incorporating with model uncertainty, unknown external disturbance, and actuator fault. The communication network between followers can be an undirected or a directed graph, and only some of the follower agents can obtain the commands from the leader. To develop the distributed cooperative control algorithm, a prefilter is designed, which can derive the state‐space representation to a newly constructed plant. Then, a set of distributed adaptive neural network controllers are designed by making certain modifications on traditional backstepping techniques with the aid of adaptive control, neural network control, and a second‐order sliding mode estimator. Rigorous proving procedures are provided, which show that uniform ultimate boundedness of all the tracking errors can be achieved in a networked multiagent system. Finally, a numerical simulation is carried out to evaluate the theoretical results.

Estimator-based adaptive neural network control of leader-follower high-order nonlinear multiagent systems with actuator faults: Estimator-based adaptive neural network control of leader-follower high-order nonlinear multiagent systems with actuator faults

The problem of distributed cooperative control for networked multiagent systems is investigated in this paper. Each agent is modeled as an uncertain nonlinear high‐order system incorporating with model uncertainty, unknown external disturbance, and actuator fault. The communication network between followers can be an undirected or a directed graph, and only some of the follower agents can obtain the commands from the leader. To develop the distributed cooperative control algorithm, a prefilter is designed, which can derive the state‐space representation to a newly constructed plant. Then, a set of distributed adaptive neural network controllers are designed by making certain modifications on traditional backstepping techniques with the aid of adaptive control, neural network control, and a second‐order sliding mode estimator. Rigorous proving procedures are provided, which show that uniform ultimate boundedness of all the tracking errors can be achieved in a networked multiagent system. Finally, a numerical simulation is carried out to evaluate the theoretical results.

Adaptive Fuzzy Control of Leader-Follower High-Order Nonlinear Multi-agent Systems with Actuator Faults

In this paper, we aim to develop a set of distributed adaptive fuzzy controllers for a group of uncertain nonlinear high-order multi-agent systems in the presence of actuator faults. Firstly, a pre-filter is designed which can derive the state space representation to a newly constructed plant. Design of the control variable will be provided by making certain modifications on traditional backstepping technique with the aid of adaptive control, fuzzy control and a second-order sliding mode estimator. It shows that the effects due to actuator faults, external disturbances and the uncertainties of the model can be compensated with the proposed scheme, and the global boundedness of the tracking errors can also be ensured.

Multi-agent Network and Cellular Automata Based Resources Assignment for City Educational System

In the paper, a city educational resources assignment problem (CERAP) is raised. The combination of big data, multi-agent model and cellular automata (CA) methodologies are argued and developed. Then, the agent-based computational framework is discussed to the educational social science control (assignment) system. The proposed methodologies provides a powerful way to address nonlinearities, humanities, and especially interdisciplinary control problem. Agent-based modeling offers powerful new forms of hybrid theoretical-computational work. The agent-based approach invites the interpretation of society as a distributed computational network, and in turn the interpretation of social dynamics as a type of computation. Then, the simulation tools are mentioned. Finally, a short discussion and conclusion end this paper.