This work presents a novel nonlinear trajectory tracking control framework for general planar models of underactuated vehicles with six states and two control inputs. In this framework, feasible state trajectories are derived through utilization of nonholonomic constraints and a transitional trajectory is introduced that reduces the sixth order system into a fourth order error dynamic stabilization problem. A nonlinear sliding mode control law is employed to stabilize the error dynamics. It is shown that the control law is uniformly asymptotically stable if unknown disturbances and modeling uncertainties are bounded. The framework is applied to differential drive mobile robots, air vehicles operating in the vertical plane, and marine vehicles. Simulations are presented for models of in-house mobile robots and surface vessels subject to unknown disturbances.
8 October 2008 in International Journal of Control
In this article, we introduce an approach based on sliding mode control to design full state feedback controllers for stabilisation of underactuated non-linear multibody systems. We define first order sliding surfaces as a linear combination of actuated and unactuated coordinate tracking errors. Lyapunov stability analysis guarantees that all system trajectories reach and remain on the sliding surfaces. However, stability of the sliding surfaces depends on the equilibrium manifold. If the system has isolated equilibrium points, it is linearly controllable and asymptotic stability can be guaranteed under certain conditions. Otherwise, the control system fails Brockett's necessary condition for existence of a smooth stabilising feedback. In the latter case, if the total momentum is conserved, the closed-loop control system will be marginally stable. Consequently, a procedure is proposed to achieve an asymptotically stable discontinuous control law through sliding surface redefinition and shape changes. It is proposed that repetitive application of shape changes will lead to asymptotic convergence of the system to the desired configuration. Simulation results are presented for an inverted pendulum as an example of a system with isolated equilibrium points and an existing communication satellite as an example of shape change control. In both cases, the control is shown to be effective and robust with respect to uncertainties and disturbances.
6 January 2022 in IEEE Transactions on Control of Network Systems
This article investigates distributed robust formation control and distributed obstacle avoidance problems for networks of heterogeneous underactuated surface vessels without global position measurements. We exploit the cascaded structure of the kinematics and dynamics of generic vessel models to develop structured reduced-order error dynamics for group cooperation. By incorporating graph theory, the supertwisting control technique, and persistence of excitation concept, a distributed robust formation control scheme is developed without requiring global position measurements, where agents in the network may possess completely different dynamic models. It is shown that the stabilization of the reduced-order error dynamics guarantees the stability of the entire vessel error system subject to modeling uncertainties and bounded disturbances. Distributed obstacle avoidance is achieved by surrounding obstacles with stable elliptical limit cycles. During the obstacle avoidance stage, a part of the formation deforms to allow vessels to follow transient trajectories around static and dynamic obstacles. Simulation results are provided to demonstrate that the proposed cooperative control scheme can prevent obstacle and interagent collisions while achieving robust formation in heterogeneous underactuated vessel networks.
We introduce a new exponentially convergent trajectory tracking control law based on the fully nonlinear quadrotor model subject to unknown disturbances and modeling uncertainties. The approach takes advantage of the system's differential flatness property to relate the second time derivative of linear accelerations (snap) to the roll and pitch moments. Using these model-based relations, the closed-loop system is transformed to follow an exponentially stable error dynamics that are robust to unknown disturbances and uncertainties bounded on a compact set. It is proven that the control law exponentially stabilizes the position and yaw tracking errors, while the roll and pitch angle tracking errors are ultimately bounded. Simulations and experimental results under windy conditions and without disturbance estimation, demonstrate the effectiveness and robustness of the approach.