Composite Learning Finite-Time Control With Application to Quadrotors

Reinforcement learning-based consensus control for MASs with intermittent constraints

In this article, an adaptive optimal consensus control problem is studied for multiagent systems in the strict-feedback structure with intermittent constraints (the constraints appear intermittently). More specifically, by designing a novel switch-like function and an improved coordinate transformation, the constrained states are converted into unconstrained states, and the problem of intermittent constraints is resolved without requiring "feasibility conditions". In addition, using the composite learning algorithm and neural networks to construct the identifier, a simplified identifier-actor-critic-based reinforcement learning strategy is proposed to obtain the approximate optimal controller under the framework of backstepping. Meanwhile, with the aid of the nonlinear dynamic surface control technique, the issue of "explosion of complexity" in backstepping is removed, and the requirements for filter parameters are loosened. Based on Lyapunov stability theory, it is demonstrated that all signals in the closed-loop system are bounded. Finally, two simulation examples are used to verify the effectiveness of the proposed method.

Predefined-Time Hierarchical Coordinated Neural Control for Hypersonic Reentry Vehicle

This paper investigates the predefined-time hierarchical coordinated adaptive control on the hypersonic reentry vehicle in presence of low actuator efficiency. In order to compensate for the deficiency of rudder deflection in advantage of channel coupling, the hierarchical design is proposed for coordination of the elevator deflection and aileron deflection. Under the control scheme, the equivalent control law and switching control law are constructed with the predefined-time technology. For the dynamics uncertainty approximation, the composite learning using the tracking error and the prediction error is constructed by designing the serial-parallel estimation model. The closed-loop system stability is analyzed via the Lyapunov approach and the tracking errors are guaranteed to be uniformly ultimately bounded in a predefined time. The tracking performance and the learning accuracy of the proposed algorithm are verified via simulation tests.

Secure predictor-based neural dynamic surface control of nonlinear cyber-physical systems against sensor and actuator attacks

This paper addresses a secure predictor-based neural dynamic surface control (SPNDSC) issue for a cyber-physical system in a nontriangular form suffering from both sensor and actuator deception attacks. To avoid the algebraic loop problem, only partial states are employed as input vectors of neural networks (NNs) for approximating unknown dynamics, and compensation terms are further developed to offset approximation errors from NNs. With introduction of nonlinear gain functions and attack compensators, adverse effects of an intelligent adversary are alleviated effectively. Furthermore, we present stability analysis and prove the ultimate boundedness of all signals in the closed-loop system. The effectiveness of the proposed control strategy is illustrated by two examples.

Quadcopter UAVs Extended States/Disturbance Observer-Based Nonlinear Robust Backstepping Control

A trajectory tracking control for quadcopter unmanned aerial vehicle (UAV) based on a nonlinear robust backstepping algorithm and extended state/disturbance observer (ESDO) is presented in this paper. To obtain robust attitude stabilization and superior performance of three-dimension position tracking control, the construction of the proposed algorithm can be separated into three parts. First, a mathematical model of UAV negatively influenced by exogenous disturbances is established. Following, an extended state/disturbance observer using a general second-order model is designed to approximate undesirable influences of perturbations on the UAVs dynamics. Finally, a nonlinear robust controller is constructed by an integration of the nominal backstepping technique with ESDO to enhance the performance of attitude and position control mode. Robust stability of the closed-loop disturbed system is obtained and guaranteed through the Lyapunov theorem without precise knowledge of the upper bound condition of perturbations. Lastly, a numerical simulation is carried out and compared with other previous controllers to demonstrate the great advantage and effectiveness of the proposed control method.

Quadrotor Real-Time Simulation: A Temporary Computational Complexity-Based Approach

The interaction of digital systems with dynamic systems requires synchrony and the accomplishment of time constrains, so the simulation of physical processes needs an implementation by means of real-time systems (RTS). However, as it can be expected, every simulation and/or implementation might demand too many computational resources, surpassing the capacity of the processor used by computational systems. This is the reason for the need to perform a temporary computational complexity analysis based on the study of the behavior of the execution times of the implemented simulation. In this regard, the Real-Time Operating Systems (RTOS) feature time managing tools, which allow their precise measurement and the establishment of scheduling criteria in process execution. Therefore, this research proposes accomplishing a temporary computational complexity analysis of the real-time simulation by an embedded system (ES) of an unmanned aerial vehicle (UAV) propelled by four rotors. Derived from this analysis, formal definitions are elaborated and proposed, which establish a close relationship between the temporary computational complexity and typical real-time temporary constraints. To the best of the author’s knowledge, the definitions presented in this article have not been reported in the literature. Furthermore, to perform the temporary computational complexity analysis of the UAV, the mathematical modeling based on the Euler–Lagrange approach is presented in detail. Finally, simulations were performed using a real-time system implemented on the Embedded Computer System (ECS) Raspberry Pi 2 Model B+, in order to validate the suggested definitions.

Disturbance Observer-Based Fault-Tolerant Control for Robotic Systems With Guaranteed Prescribed Performance

The actuator failure compensation control problem of robotic systems possessing dynamic uncertainties has been investigated in this paper. Control design against partial loss of effectiveness (PLOE) and total loss of effectiveness (TLOE) of the actuator are considered and described, respectively, and a disturbance observer (DO) using neural networks is constructed to attenuate the influence of the unknown disturbance. Regarding the prescribed error bounds as time-varying constraints, the control design method based on barrier Lyapunov function (BLF) is used to strictly guarantee both the steady-state performance and the transient performance. A simulation study on a two-link planar manipulator verifies the effectiveness of the proposed controllers in dealing with the prescribed performance, the system uncertainties, and the unknown actuator failure simultaneously. Implementation on a Baxter robot gives an experimental verification of our controller.

Adaptive Integral Sliding Mode Control Using Fully Connected Recurrent Neural Network for Position and Attitude Control of Quadrotor

This article proposes an adaptive integral sliding mode control (ISMC) strategy for quadrotor control that ensures faster and finite-time convergence along with chattering attenuation. Quadrotor dynamics are assumed to be unknown because of the high degree of parametric uncertainties, including external disturbances. The equivalent control law obtained by ISMC consists of quadrotor dynamics and, thus, cannot be applied to the quadrotor. A new fully connected recurrent neural network (FCRNN) controller has been proposed to mimic the equivalent control instead of estimating the Quadrotor dynamics separately. The proposed FCRNN architecture consists of output feedback to the input layer and the hidden layer, which enhances the approximation capability of FCRNN. All hidden layer neurons receive self-feedback and feedback from other hidden layer neurons, which further strengthens FCRNN's potential to capture complex dynamic characteristics. As learning should happen in finite time, the finite-time stability of the overall system has been guaranteed using the Lyapunov stability theory, and the update laws for FCRNN weights in real time are derived using the same. To show the effectiveness of the proposed approach, a comprehensive analysis has been done against existing SMC strategy and against well-known function approximation techniques, e.g., the radial basis function network (RBFN) and RNN.

Robust Intelligent Control of SISO Nonlinear Systems Using Switching Mechanism

In this article, a robust adaptive learning control strategy for uncertain single-input-single-output systems in strict-feedback form and controllability canonical form (CCF) is studied. For the strict-feedback system, the dynamic surface control is introduced while for the controllability canonical system, sliding-mode control is further constructed. The finite-time design is introduced for fast convergence. Under the switching mechanism, the intelligent design and the robust technique work together to obtain robust tracking performance. Once the states run out of the domain of intelligent control, the robust item will pull the states back while inside the neural working domain, the composite learning is developed to achieve higher approximation precision by building the prediction error for the weight update. The closed-loop system stability is analyzed via the Lyapunov approach. Especially for the CCF, the finite-time convergence is achieved while the system signals are globally uniformly ultimately bounded. Simulation studies on the general nonlinear systems and the flight dynamics show that the new design scheme obtains better tracking performance with higher precision and stronger robustness.

General Purpose Low-Level Reinforcement Learning Control for Multi-Axis Rotor Aerial Vehicles

This paper proposes a multipurpose reinforcement learning based low-level multirotor unmanned aerial vehicles control structure constructed using neural networks with model-free training. Other low-level reinforcement learning controllers developed in studies have only been applicable to a model-specific and physical-parameter-specific multirotor, and time-consuming training is required when switching to a different vehicle. We use a 6-degree-of-freedom dynamic model combining acceleration-based control from the policy neural network to overcome these problems. The UAV automatically learns the maneuver by an end-to-end neural network from fusion states to acceleration command. The state estimation is performed using the data from on-board sensors and motion capture. The motion capture system provides spatial position information and a multisensory fusion framework fuses the measurement from the onboard inertia measurement units for compensating the time delay and low update frequency of the capture system. Without requiring expert demonstration, the trained control policy implemented using an improved algorithm can be applied to various multirotors with the output directly mapped to actuators. The algorithm's ability to control multirotors in the hovering and the tracking task is evaluated. Through simulation and actual experiments, we demonstrate the flight control with a quadrotor and hexrotor by using the trained policy. With the same policy, we verify that we can stabilize the quadrotor and hexrotor in the air under random initial states.

Robust Neurooptimal Control for a Robot via Adaptive Dynamic Programming

We aim at the optimization of the tracking control of a robot to improve the robustness, under the effect of unknown nonlinear perturbations. First, an auxiliary system is introduced, and optimal control of the auxiliary system can be seen as an approximate optimal control of the robot. Then, neural networks (NNs) are employed to approximate the solution of the Hamilton-Jacobi-Isaacs equation under the frame of adaptive dynamic programming. Next, based on the standard gradient attenuation algorithm and adaptive critic design, NNs are trained depending on the designed updating law with relaxing the requirement of initial stabilizing control. In light of the Lyapunov stability theory, all the error signals can be proved to be uniformly ultimately bounded. A series of simulation studies are carried out to show the effectiveness of the proposed control.

Robust Quadrotor Control through Reinforcement Learning with Disturbance Compensation

In this paper, a novel control strategy is presented for reinforcement learning with disturbance compensation to solve the problem of quadrotor positioning under external disturbance. The proposed control scheme applies a trained neural-network-based reinforcement learning agent to control the quadrotor, and its output is directly mapped to four actuators in an end-to-end manner. The proposed control scheme constructs a disturbance observer to estimate the external forces exerted on the three axes of the quadrotor, such as wind gusts in an outdoor environment. By introducing an interference compensator into the neural network control agent, the tracking accuracy and robustness were significantly increased in indoor and outdoor experiments. The experimental results indicate that the proposed control strategy is highly robust to external disturbances. In the experiments, compensation improved control accuracy and reduced positioning error by 75%. To the best of our knowledge, this study is the first to achieve quadrotor positioning control through low-level reinforcement learning by using a global positioning system in an outdoor environment.

Fixed-time trajectory following for quadrotors via output feedback

A fixed-time trajectory following problem for quadrotors via output feedback is concerned. Based on the inner-outer separation design philosophy, the under-actuated quadrotor is formulated as a hierarchical structure composed by position and attitude dynamics. With an emphasis on removing the demand on unmeasured velocity and eliminating the negative effect of disturbances, fixed-time extended state observers utilizing two kinds of polynomial feedback terms are proposed to simultaneously identify unavailable velocity states and unknown uncertainties with a fixed-time estimation capability. With these observation results, a velocity free fixed-time control protocol is synthesized to enable a satisfied trajectory regulation with a uniform convergence time independent of initial positions, such that a prescribed fixed-time stability and enhanced robustness can be obtained with chattering-free inputs. By virtue of bi-limit homogeneity properties, all error variables of the resultant quadrotor system are demonstrated to be fixed-time convergent. Eventually, the benefits of developed algorithm are illustrated via simulations.

Universal Adaptive Neural Network Predictive Algorithm for Remotely Piloted Unmanned Combat Aerial Vehicle in Wireless Sensor Network

Remotely piloted unmanned combat aerial vehicle (UCAV) will be a prospective mode of air fight in the future, which can remove the physical restraint of the pilot, maximize the performance of the fighter and effectively reduce casualties. However, it has two difficulties in this mode: (1) There is greater time delay in the network of pilot-wireless sensor-UCAV, which can degrade the piloting performance. (2) Designing of a universal predictive method is very important to pilot different UCAVs remotely, even if the model of the control augmentation system of the UCAV is totally unknown. Considering these two issues, this paper proposes a novel universal modeling method, and establishes a universal nonlinear uncertain model which uses the pilot’s remotely piloted command as input and the states of the UCAV with a control augmentation system as output. To deal with the nonlinear uncertainty of the model, a neural network observer is proposed to identify the nonlinear dynamics model online. Meanwhile, to guarantee the stability of the overall observer system, an adaptive law is designed to adjust the neural network weights. To solve the greater transmission time delay existing in the pilot-wireless sensor-UCAV closed-loop system, a time-varying delay state predictor is designed based on the identified nonlinear dynamics model to predict the time delay states. Moreover, the overall observer-predictor system is proved to be uniformly ultimately bounded (UUB). Finally, two simulations verify the effectiveness and universality of the proposed method. The results indicate that the proposed method has desirable performance of accurately compensating the time delay and has universality of remotely piloting two different UCAVs.

Composite Neural Learning-Based Nonsingular Terminal Sliding Mode Control of MEMS Gyroscopes

The efficient driving control of MEMS gyroscopes is an attractive way to improve the precision without hardware redesign. This paper investigates the sliding mode control (SMC) for the dynamics of MEMS gyroscopes using neural networks (NNs). Considering the existence of the dynamics uncertainty, the composite neural learning is constructed to obtain higher tracking precision using the serial-parallel estimation model (SPEM). Furthermore, the nonsingular terminal SMC (NTSMC) is proposed to achieve finite-time convergence. To obtain the prescribed performance, a time-varying barrier Lyapunov function (BLF) is introduced to the control scheme. Through simulation tests, it is observed that under the BLF-based NTSMC with composite learning design, the tracking precision of MEMS gyroscopes is highly improved.

Hybrid finite-time trajectory tracking control of a quadrotor

In this paper, accurate trajectory tracking control problem of a quadrotor with unknown dynamics and disturbances is addressed by devising a hybrid finite-time control (HFTC) approach. An adaptive integral sliding mode (AISM) control law is proposed for altitude subsystem of the quadrotor, whereby underactuated characteristics can decoupled. Backstepping technique is further deployed to control the horizontal position subsystem. To exactly attenuate external disturbances, a finite-time disturbance observer (FDO) combining with nonsingular terminal sliding mode (NTSM) control strategy is constructed for attitude subsystem, and thereby achieve finite-time stability. Using the compounded control scheme, trajectory tracking errors can be stabilized rapidly. Simulation results and comprehensive comparisons show that the proposed HFTC scheme has remarkably superior performance.

Barrier Lyapunov Function Based Learning Control of Hypersonic Flight Vehicle With AOA Constraint and Actuator Faults

This paper investigates a fault-tolerant control of the hypersonic flight vehicle using back-stepping and composite learning. With consideration of angle of attack (AOA) constraint caused by scramjet, the control laws are designed based on barrier Lyapunov function. To deal with the unknown actuator faults, a robust adaptive allocation law is proposed to provide the compensation. Meanwhile, to obtain good system uncertainty approximation, the composite learning is proposed for the update of neural weights by constructing the serial-parallel estimation model to obtain the prediction error which can dynamically indicate how the intelligent approximation is working. Simulation results show that the controller obtains good system tracking performance in the presence of AOA constraint and actuator faults.