Reinforcement Learning Tracking Control for Robotic Manipulator With Kernel-Based Dynamic Model

2020 ◽  
Vol 31 (9) ◽  
pp. 3570-3578 ◽  
Author(s):  
Yazhou Hu ◽  
Wenxue Wang ◽  
Hao Liu ◽  
Lianqing Liu
IEEE Access ◽  
2021 ◽  
Vol 9 ◽  
pp. 15592-15602
Author(s):  
Xueshan Gao ◽  
Rui Gao ◽  
Peng Liang ◽  
Qingfang Zhang ◽  
Rui Deng ◽  
...  

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Zafer Bingul ◽  
Oguzhan Karahan

Purpose The purpose of this paper is to address a fractional order fuzzy PID (FOFPID) control approach for solving the problem of enhancing high precision tracking performance and robustness against to different reference trajectories of a 6-DOF Stewart Platform (SP) in joint space. Design/methodology/approach For the optimal design of the proposed control approach, tuning of the controller parameters including membership functions and input-output scaling factors along with the fractional order rate of error and fractional order integral of control signal is tuned with off-line by using particle swarm optimization (PSO) algorithm. For achieving this off-line optimization in the simulation environment, very accurate dynamic model of SP which has more complicated dynamical characteristics is required. Therefore, the coupling dynamic model of multi-rigid-body system is developed by Lagrange-Euler approach. For completeness, the mathematical model of the actuators is established and integrated with the dynamic model of SP mechanical system to state electromechanical coupling dynamic model. To study the validness of the proposed FOFPID controller, using this accurate dynamic model of the SP, other published control approaches such as the PID control, FOPID control and fuzzy PID control are also optimized with PSO in simulation environment. To compare trajectory tracking performance and effectiveness of the tuned controllers, the real time validation trajectory tracking experiments are conducted using the experimental setup of the SP by applying the optimum parameters of the controllers. The credibility of the results obtained with the controllers tuned in simulation environment is examined using statistical analysis. Findings The experimental results clearly demonstrate that the proposed optimal FOFPID controller can improve the control performance and reduce reference trajectory tracking errors of the SP. Also, the proposed PSO optimized FOFPID control strategy outperforms other control schemes in terms of the different difficulty levels of the given trajectories. Originality/value To the best of the authors’ knowledge, such a motion controller incorporating the fractional order approach to the fuzzy is first time applied in trajectory tracking control of SP.


Complexity ◽  
2021 ◽  
Vol 2021 ◽  
pp. 1-7
Author(s):  
Xiaoyi Long ◽  
Zheng He ◽  
Zhongyuan Wang

This paper suggests an online solution for the optimal tracking control of robotic systems based on a single critic neural network (NN)-based reinforcement learning (RL) method. To this end, we rewrite the robotic system model as a state-space form, which will facilitate the realization of optimal tracking control synthesis. To maintain the tracking response, a steady-state control is designed, and then an adaptive optimal tracking control is used to ensure that the tracking error can achieve convergence in an optimal sense. To solve the obtained optimal control via the framework of adaptive dynamic programming (ADP), the command trajectory to be tracked and the modified tracking Hamilton-Jacobi-Bellman (HJB) are all formulated. An online RL algorithm is the developed to address the HJB equation using a critic NN with online learning algorithm. Simulation results are given to verify the effectiveness of the proposed method.


Sign in / Sign up

Export Citation Format

Share Document