scholarly journals Data-Driven Model-Free Tracking Reinforcement Learning Control with VRFT-based Adaptive Actor-Critic

2019 ◽  
Vol 9 (9) ◽  
pp. 1807 ◽  
Author(s):  
Radac ◽  
Precup
Keyword(s):  
Process Model ◽  
State Feedback ◽  
Reference Model ◽  
Open Loop ◽  
Design Guidelines ◽  
Feedback Controller ◽  
Input State ◽  
State Feedback Controller ◽  
Learning Framework ◽  
Model Free

This paper proposes a neural network (NN)-based control scheme in an Adaptive Actor-Critic (AAC) learning framework designed for output reference model tracking, as a representative deep-learning application. The control learning scheme is model-free with respect to the process model. AAC designs usually require an initial controller to start the learning process; however, systematic guidelines for choosing the initial controller are not offered in the literature, especially in a model-free manner. Virtual Reference Feedback Tuning (VRFT) is proposed for obtaining an initially stabilizing NN nonlinear state-feedback controller, designed from input-state-output data collected from the process in open-loop setting. The solution offers systematic design guidelines for initial controller design. The resulting suboptimal state-feedback controller is next improved under the AAC learning framework by online adaptation of a critic NN and a controller NN. The mixed VRFT-AAC approach is validated on a multi-input multi-output nonlinear constrained coupled vertical two-tank system. Discussions on the control system behavior are offered together with comparisons with similar approaches.

Adaptive state-feedback stabilization of stochastic nonholonomic systems with an unknown time-varying delay and perturbations

2020 ◽  
pp. 014233122097417
Author(s):  
Qinghui Du
Keyword(s):  
State Feedback ◽  
Nonholonomic Systems ◽  
Feedback Stabilization ◽  
Feedback Controller ◽  
Input State ◽  
Time Varying ◽  
Initial State ◽  
State Feedback Controller ◽  
Time Varying Delay ◽  
Varying Delay

The problem of adaptive state-feedback stabilization of stochastic nonholonomic systems with an unknown time-varying delay and perturbations is studied in this paper. Without imposing any assumptions on the time-varying delay, an adaptive state-feedback controller is skillfully designed by using the input-state scaling technique and an adaptive backstepping control approach. Then, by adopting the switching strategy to eliminate the phenomenon of uncontrollability, the proposed adaptive state-feedback controller can guarantee that the closed-loop system has an almost surely unique solution for any initial state, and the equilibrium of interest is globally asymptotically stable in probability. Finally, the simulation example shows the effectiveness of the proposed scheme.

H∞ fuzzy proportional integral state feedback controller of photovoltaic systems under asymmetric actuator constraints

2020 ◽  
pp. 014233122092157
Author(s):  
K Houda ◽  
D Saifia ◽  
M Chadli ◽  
S Labiod
Keyword(s):  
State Feedback ◽  
Reference Model ◽  
Tracking Error ◽  
Control Design ◽  
Atmospheric Condition ◽  
Boost Converter ◽  
Feedback Controller ◽  
Duty Ratio ◽  
Linear Matrix ◽  
State Feedback Controller

This paper presents a new strategy for a robust maximum power point (MPP) tracking fuzzy controller for photovoltaic (PV) systems subject to actuator asymmetric saturation. A DC-DC boost converter is used to connect a PV panel with an output load. The output voltage of the DC-DC boost converter can be adjusted by duty ratio that is limited between 0 and 1. The aim of our control design is to track the MPP under atmospheric condition changes and the presence of the asymmetric saturation of the duty ratio. To minimize tracking error and disturbance effect, the dynamic behaviour of a PV system and its reference model are described by using Takagi–Sugeno fuzzy models. Then, a constrained control based on a fuzzy PI state feedback controller is proposed. The H∞ control approach is used in control design and stability conditions of the closed-loop system are formulated and solved in terms of linear matrix inequalities. Finally, simulation results are given to show the tracking performance of the control design.

scholarly journals Design and Analysis of a Variable Buoyancy System for Efficient Hovering Control of Underwater Vehicles with State Feedback Controller

2020 ◽  
Vol 8 (4) ◽  
pp. 263 ◽  
Author(s):  
Brij Kishor Tiwari ◽  
Rajiv Sharma
Keyword(s):  
State Feedback ◽  
Autonomous Underwater Vehicles ◽  
Weather Conditions ◽  
Underwater Vehicles ◽  
Open Loop ◽  
Rate Of Change ◽  
Feedback Controller ◽  
State Feedback Controller ◽  
Adverse Weather ◽  
Hovering Control

The design process for Variable Buoyancy System (VBS) is not known in full, and existing approaches are not scalable. Furthermore, almost all the small size Autonomous Underwater Vehicles/Gliders (AUVs/G’s) use very low capacity of buoyancy change (in the range of few milliliters) and the large size AUVs require large buoyancy change. Especially for adverse weather conditions, emergency recovery or defense-related applications, higher rate of rising/sinking (heave velocity) is needed along with an ability to hover at certain depth of operation. Depth of UVs can be controlled either by changing the displaced volume or by changing the overall weight and, herein, our focus is on the later. This article presents the problem of design and analysis of VBS for efficient hovering control of underwater vehicles at desired depth using the state feedback controller. We formulate and analyze the design and analysis approach of VBS using the fundamental of mechanics, system dynamics integration and control theory. Buoyancy is controlled by changing the overall weight of the vehicle using the ballasting/de-ballasting of water in ballast tanks through the use of Positive Displacement Pump (PDP) for control in heave velocity and hovering depth. Furthermore, detailed mass metric analysis of scalable design of VBS for different buoyancy capacities is performed to analyze the overall performance of the VBS. Also, the performances of AUVs integrated with VBS of different buoyancy capacities are investigated in both the open loop and closed loop with the LQR state feedback controller. Hovering performance results are presented for three Design Examples (DEs) of AUVs with 2.8 m, 4.0 m and 5.0 m length and they are integrated with various buoyancy capacities at 9 kg/min rate of change of buoyancy. Results indicate that the AUVs achieve the desired depth with almost negligible steady state error and when they reach the desired hovering depth of 400 m the maximum pitch angle achieved of 16.5 degree for all the Des is observed. Maximum heave velocity achieved during sinking is 0.44 m/s and it reduces to zero when the vehicle reaches the desired depth of hovering. The presented computer simulation results indicate good performance and demonstrate that the designed VBS is effective and efficient in changing the buoyancy, controlling and maintaining the depth, controlling the heave velocity and can be used in rescue/attack operations of both the civil and defense UVs.

Ball and Beam Balancing Mechanism Actuated With Pneumatic Artificial Muscles

2018 ◽  
Vol 10 (5) ◽  
Author(s):  
Željko Šitum ◽  
Petar Trslić
Keyword(s):  
State Feedback ◽  
Tracking Error ◽  
Design Procedure ◽  
Open Loop ◽  
The State ◽  
Feedback Controller ◽  
Reference Trajectory ◽  
State Feedback Controller ◽  
Beam System ◽  
Muscle Pair

The paper presents the results of modeling and control of an original and unique ball-on-beam system with a pneumatic artificial muscle pair in an antagonistic configuration. This system represents a class of under-actuated, high-order nonlinear systems, which are characterized by an open-loop unstable equilibrium point. Since pneumatic muscles have elastic, nonlinear characteristics, they are more difficult to control. Considering that an additional nonlinearity is added to the system which makes it harder to stabilize. The nonlinear mathematical model has been derived based on the physical model of the ball-on-beam mechanism, the beam rotating by using an antagonistic muscle pair and the pneumatic muscle actuated by a proportional valve. Based on the nonlinear model, the linearized equations of motion have been derived and a control-oriented model has been developed, which is used in the state feedback controller design procedure. The proposed state feedback controller has been verified by means of computer simulations and experimentally on the laboratory setup. The simulation and experimental results have shown that the state feedback controller can stabilize the ball-on-beam system around an equilibrium position in the presence of external disturbances and to track a reference trajectory with a small tracking error.

scholarly journals Robust Position Control of a Two-Sided 1-DoF Impacting Mechanical Oscillator Subject to an External Persistent Disturbance by Means of a State-Feedback Controller

Complexity ◽  
2019 ◽  
Vol 2019 ◽  
pp. 1-14 ◽  
Author(s):  
Firas Turki ◽  
Hassène Gritli ◽  
Safya Belghith
Keyword(s):  
State Feedback ◽  
Position Control ◽  
External Disturbance ◽  
Feedback Controller ◽  
Linear Matrix ◽  
Stability Conditions ◽  
Mechanical Oscillator ◽  
Impact Oscillator ◽  
State Feedback Controller ◽  
The Impact

This paper proposes a state-feedback controller using the linear matrix inequality (LMI) approach for the robust position control of a 1-DoF, periodically forced, impact mechanical oscillator subject to asymmetric two-sided rigid end-stops. The periodic forcing input is considered as a persistent external disturbance. The motion of the impacting oscillator is modeled by an impulsive hybrid dynamics. Thus, the control problem of the impact oscillator is recast as a problem of the robust control of such disturbed impulsive hybrid system. To synthesize stability conditions, we introduce the S-procedure and the Finsler lemmas by only considering the region within which the state evolves. We show that the stability conditions are first expressed in terms of bilinear matrix inequalities (BMIs). Using some technical lemmas, we convert these BMIs into LMIs. Finally, some numerical results and simulations are given. We show the effectiveness of the designed state-feedback controller in the robust stabilization of the position of the impact mechanical oscillator under the disturbance.

Sign in / Sign up

Export Citation Format

Share Document