Estimation Of Unknown Dynamics Research Articles

Onboard estimation of unknown dynamics of flexible spacecraft

Attitude control methods of highly flexible spacecraft have seen increased interest over the last decades thanks to the technological development of flexible solar panels and deployables, which improves the capabilities of satellites. However, a high-fidelity model of the flexible dynamics is hard to obtain on-ground testing because not all modes can be observed, complicating the controller design. This paper proposes a method to develop a high-fidelity model of a spacecraft with a flexible appendage subject to large deformations by modeling it as a finite series of rigid links connected by torsional springs and dampers. To overcome the uncertainties or unknowns in the flexible dynamics, an onboard estimation through an adaptive controller is performed for these unknowns while the spacecraft is maneuvered. The controller uses integral concurrent learning (ICL), an adaptive scheme that records inputs and outputs provided by sensors mounted on the flexible body. The novelty of this investigation is the development of self-adapting control gains for the learning matrix obtained from ICL. This approach achieves the objective of tracking a desired trajectory while accurately learning the unknown physical parameters of the flexible appendage by only using the recorded measurements of the appendage displacements and the spacecraft attitude and velocities. It was observed that for a finer discretization of the flexible appendage and therefore a higher fidelity model of the flexible dynamics, the estimation algorithm is able to observe all the frequencies necessaries to learn the unknown mechanical properties of the flexible body.

Robust Near-Optimal Coordination in Uncertain Multiagent Networks With Motion Constraints.

This article addresses the robust coordination problem for nonlinear uncertain second-order multiagent networks with motion constraints, including velocity saturation and collision avoidance. A single-critic neural network-based approximate dynamic programming approach and exact estimation of unknown dynamics are employed to learn online the optimal value function and controller. By incorporating avoidance penalties into tracking variable, constructing a novel value function, and designing of suitable learning algorithms, multiagent coordination and collision avoidance are achieved simultaneously. We show that the developed feedback-based coordination strategy guarantees uniformly ultimately bounded convergence of the closed-loop dynamical stability and all underlying motion constraints are always strictly obeyed. The effectiveness of the proposed collision-free coordination law is finally illustrated using numerical simulations.

Distributed Nash Equilibrium Computation With Uncertain Dynamics and Disturbances

Nash equilibrium computation problems under distributed communication scenarios are addressed in this paper. In particular, the players, including first- and second-order ones, are considered to be affected by unknown disturbances and dynamics. To accommodate this problem, the followings need to be simultaneously achieved: (1) Distributed estimation of unknown information needed for optimizing the players’ objective functions; (2) The optimization of the players’ objective functions; and (3) Dynamics stabilization for the players, especially for the accommodation of unknown disturbances and dynamics. Based on first-order low-pass filters, an unknown dynamics estimator is established (for the estimation of unknown dynamics and disturbances), by which distributed computing methods are constructed by further employing gradient-like optimization components (for optimizing the players’ objective functions) and consensus components (for distributed action estimation). The presented analytical investigation shows that the players’ actions and velocities (for second-order players) can converge into a small ball around the Nash equilibrium and zero respectively by choosing the control parameters properly according to the given rules. Moreover, the proposed algorithm is adapted to address connectivity control of mobile sensors. Numerical verification is provided by utilizing KUKA YouBot vehicles in V-REP.

A Disturbance Observer-Based Intelligent Finite-Time Sliding Mode Flight Controller Design for an Autonomous Quadrotor

This article presents an intelligent control strategy for the position and attitude tracking of a quadrotor using a nonsingular fast-terminal sliding mode and disturbance observer. The quadrotor system is divided into an underactuated position and the fully actuated attitude subsystem. A single hidden layer feed-forward neural network is used for the estimation of unknown dynamics associated with the attitude subsystem. Moreover, a nonlinear disturbance observer is designed to tackle the unknown bounded external disturbances acting on the attitude subsystem by supplying its estimation in the control laws that help in mitigating the chattering phenomena. The proposed methodology ensures the finite-time convergence of the tracking error and does not prone to the singularity problem in control. The closed-loop finite-time stability is investigated using the Lyapunov theory. Besides, an expression for the convergence time has been derived. The effectiveness of the proposed scheme is initially assessed using numerical simulations for the way-point tracking as well as circle tracking tasks and then validated in real-time using DJI Matrice 100 by performing a lemniscate tracking task. To show the superiority of the proposed methodology over existing methods, a comparative study has been done and found the improved tracking performance in the proposed approach.

Unknown Dynamics Estimator-Based Output-Feedback Control for Nonlinear Pure-Feedback Systems

Most existing adaptive control designs for nonlinear pure-feedback systems have been derived based on backstepping or dynamic surface control (DSC) methods, requiring full system states to be measurable. The neural networks (NNs) or fuzzy logic systems (FLSs) used to accommodate uncertainties also impose demanding computational cost and sluggish convergence. To address these issues, this paper proposes a new output-feedback control for uncertain pure-feedback systems without using backstepping and function approximator. A coordinate transform is first used to represent the pure-feedback system in a canonical form to evade using the backstepping or DSC scheme. Then the Levant's differentiator is used to reconstruct the unknown states of the derived canonical system. Finally, a new unknown system dynamics estimator with only one tuning parameter is developed to compensate for the lumped unknown dynamics in the feedback control. This leads to an alternative, simple approximation-free control method for pure-feedback systems, where only the system output needs to be measured. The stability of the closed-loop control system, including the unknown dynamics estimator and the feedback control is proved. Comparative simulations and experiments based on a PMSM test-rig are carried out to test and validate the effectiveness of the proposed method.

Robust Adaptive Sliding Mode Controller for Wearable Robots

Abstract This paper concerns adaptive sliding-mode control for wearable robots with a human in the loop. The exoskeletons are wearable robots in interaction with different users. The proposed approach supposes that the dynamic model of the exoskeleton-human system is unknown except for some classical bounded properties. The controller guarantees the closed-loop convergence with an embedded-in estimation of unknown dynamics and uncertainties. The stability analysis of the system is demonstrated using the Lyapunov theory. Experimentation on an upper arm exoskeleton was conducted in order to exhibit the effectiveness of the proposed control method. The results show good tracking of the desired trajectories, which can be used in the assistive phase of the functional rehabilitation.

Output Feedback Control of Uncertain Hydraulic Servo Systems

Most of the existing control methods for servo systems driven by hydraulic actuators have been developed by using a backstepping scheme and assuming that all system states (including internal hydraulic signals) are measurable. In this paper, we propose a new control design method for high-order servo systems with hydraulic actuator dynamics, where the backstepping scheme is avoided and only the system output (e.g., motion displacement) is required for the control implementation. For this purpose, the system model is first transformed into a canonical form, where the unknown dynamics in the system are lumped as one term. Then, we introduce a simple robust unknown dynamics estimator (UDE) that has only one tuning parameter but achieves exponential error convergence to accommodate the lumped uncertainties. Therefore, the function approximators (e.g., neural network and fuzzy systems) can be avoided, leading to reduced computational costs, simpler parameter tuning, and improved convergence as compared to backstepping methods. Extensive simulations and experiments based on a realistic test rig are conducted to show the efficacy of the proposed control.

Adaptive Tracking Control of an Exoskeleton Robot With Uncertain Dynamics Based on Estimated Time-Delay Control

In this paper, we present a backstepping approach integrated with time-delay estimation to provide an accurate estimation of unknown dynamics and to compensate for external bounded disturbances. The control was implemented to perform passive rehabilitation movements with a 7-DOF exoskeleton robot named ETS-Motion Assistive Robotic-Exoskeleton for Superior Extremity. The unknown dynamics and external bounded disturbances can affect the robotic system in the form of input saturation, time-delay errors, friction forces, backlash, and different upper-limb's mass of each subject. The output of the time-delay estimator is coupled directly to the control input of the proposed adaptive tracking control through a feed-forward loop. In this case, the control system ensures a highly accurate tracking of the desired trajectory, while being robust to the uncertainties and unforeseen external forces, and flexible with variation of parameters. Due to the proposed strategy, the designed control approach does not require accurate knowledge of the dynamic parameters of the exoskeleton robot to achieve the desired performance. The stability of the exoskeleton robot and the convergence of its state errors are established and proved based on Lyapunov–Krasovskii functional theory. Experimental results and a comparative study are presented to validate the advantages of the proposed strategy.

What can regularization offer for estimation of dynamical systems?

Estimation of unknown dynamics is what system identification is about and a core problem in adaptive control and adaptive signal processing. It has long been known that regularization can be quite beneficial for general inverse problems of which system identification is an example. But only recently, partly under the influence of machine learning, the use of well tuned regularization for estimating linear dynamical systems has been investigated more seriously. In this presentation we review these new results and discuss what they may mean for the theory and practice of dynamical model estimation in general.

Robust control with unknown dynamic estimation for multi-axial piezoelectric actuators with coupled dynamics

Piezoelectric actuators are widely used for precise micro-positioning. The ability of fine positioning is strictly under the effect of hysteresis nonlinear behavior. Simultaneous micro-positioning in multi-dimensions has also attracted much attention in recent years. In addition to hysteresis behavior, a nonlinear dynamic coupling exists between the different degrees of freedom in multi-axis piezoelectric actuators. The nonlinear coupling phenomenon is called the Axes Coupling Effect (ACE). A modified Prandtl–Ishlinskii (PI) operator and its inverse are utilized for both the identification and real time compensation of the hysteresis effect in this article. Considering the PI estimation error and probable un-modeled dynamics, a variable structure controller coupled with the neural network is proposed for position tracking. Due to the model-based structure of the proposed controller, the dynamic model of actuator should be identified. Coupling between nonlinear hysteresis behavior and the linear dynamic model causes a complicated identification. The ACE has also an unknown trend. Eliminating the necessity of dynamic parameters and ACE identification, a Radial Basis Function (RBF) neural network approach would estimate the unknown dynamics of the designed controller. Stability of the controller in the presence of estimated unknown dynamics is demonstrated analytically. Experimental results depict that the proposed approach can achieve precise tracking of multi-frequency trajectories and appropriate estimation of unknown dynamics.

Robust Decentralized Trajectory Tracking Control of Robot Manipulators Based on Extended State Observer

摘要: 针对模型不确定性多关节机器人的轨迹跟踪控制问题, 提出一种基于扩张状态观测补偿的关节独立分散鲁棒控制方法. 设计多个并联的扩张状态观测器, 分别对每个独立关节的未知模型动态与外部扰动进行动态估计和补偿, 采取非线性状态反馈控制以提高系统的控制性能. 各个关节的控制完全独立, 因而控制器结构简单、可靠、易于实现. 对所设计的关节分散鲁棒控制器的 Lyapunov 稳定性进行了分析, 证明系统是指数收敛全局一致最终有界稳定的. 对 PUMA560 机械手的大量控制仿真验证了本文方法的有效性. 关键词: 机器人 / 扩张状态观测器 / 自抗扰控制 / 分散控制 / 稳定性