Robust fault‐tolerant consensus control for nonlinear multi‐agent systems with prescribed transient and steady‐state performance

Abstract

This paper investigates the finite‐time consensus tracking problem for second‐order nonlinear multi‐agent systems subject to the actuator bias fault. Based on the sliding mode control and prescribed performance control, two distributed fault‐tolerant consensus protocols are proposed for the sudden bias fault and the slowly varying bias fault of the actuator. The first protocol adopts a passive fault‐tolerant control scheme based on the robustness of the system, which has a simple structure, and is timely and effective against sudden faults. In the second protocol, a novel nonlinear recursive sliding mode surface is designed, and a distributed fault observer is proposed to compensate the influences caused by actuator faults actively. Both of the protocols can prescribe the transient and steady‐state performance of the system. Finally, the performances of the two protocols are verified by two simulation cases for the attitude systems of multiple unmanned underwater vehicles.

Leader‐following consensus control of second‐order nonlinear multi‐agent systems with applications to slew rate control

“…However, an intrinsic feature for consensus control of multi‐agent systems is that a distributed control law only uses the relative state information. In addition, the control laws designed in aforementioned works [5–7, 10, 17] used the eigenvalues of Laplacian matrix, which is global and not available to each agent. To avoid above three restrictions, other works [18–20] proposed distributed adaptive control laws for the leader‐following consensus problem of second‐order and high‐order integrator‐type nonlinear multi‐agent systems, respectively.…”

“…This allows us to design distributed control protocol that does not dependent on agent's own state information. Compared with other studies [18–20], where integrator type nonlinear multi‐agent systems were studied, we investigate a more general nonlinear multi‐agent systems in strict‐feedback form. The controller design is more challenging since virtual control laws need to be designed as nonlinear for systems in strict‐feedback form. Compared with other works [5–7,10,17], we construct a fully distributed control law without using any global information of Laplacian matrix.…”

“…Since physical systems are essentially nonlinear, many researchers have also investigated the consensus or containment problem of nonlinear multi-agent systems, especially for systems in strict-feedback form or a special case of integrator-type. For example, previous works [5][6][7][8][9][10] studied the consensus problem of systems with nonlinear functions satisfying globally Lipschitz or Lipschitz-like condition. Niu et al [11] proposed adaptive consensus protocols for first-order and second-order uncertain nonlinear systems with unknown control coefficients.…”

“…• Compared with other works [5][6][7][8][9][10], we consider a class of more general nonlinear multi-agent systems. The nonlinear functions are not required to satisfy globally Lipschitz nor Lipschitz-like condition in this paper.…”

“…The controller design is more challenging since virtual control laws need to be designed as nonlinear for systems in strict-feedback form. • Compared with other works [5][6][7]10,17], we construct a fully distributed control law without using any global information of Laplacian matrix.…”