Adaptive visual servoing with an uncalibrated camera using extreme learning machine and Q-leaning

Meng, Kui; Chen, Hao; Dong, Jie

doi:10.1016/j.neucom.2020.03.049

Cited by 35 publications

(25 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Kang et al [7] adopted a reinforcement learning method to adaptively adjust the servoing gain to improve the convergence rate and stability. Li et al [8] combined proportional derivative (PD) control with sliding mode control (SMC) to tackle the disturbance and uncertainties on a 6-degree-offreedom (DOF) VS manipulator.…”

Section: Related Workmentioning

confidence: 99%

Adaptive dynamic programming-based feature tracking control of visual servoing manipulators with unknown dynamics

Ren

2021

Complex Intell. Syst.

View full text Add to dashboard Cite

This paper investigates a feature tracking control method for visual servoing (VS) manipulators adaptive dynamic programming (ADP)-based the unknown dynamics. The major superiority of ADP-based optimal control lies in that the visual tracking problem is converted to the feature tracking error control with optimal cost function. Moreover, an adaptive neural network observer is developed to approximate the entire uncertainties, which are utilized to construct an improved cost function. By establishing a critic neural network, the Hamilton–Jacobi–Bellman (HJB) equation is solved, and the approximate optimal error control policy is derived. The closed-loop VS manipulator system is verified to be ultimately uniformly bounded with the developed ADP-based feature tracking control strategy according to the Lyapunov theory. Finally, simulation results under various situations demonstrate that the proposed method achieves higher tracking accuracy than other methods, as well as satisfies energy optimal requirements.

show abstract

Section: Related Workmentioning

confidence: 99%

Adaptive dynamic programming-based feature tracking control of visual servoing manipulators with unknown dynamics

Ren

2021

Complex Intell. Syst.

View full text Add to dashboard Cite

show abstract

“…As shown in Fig. 1, the innovation of this paper lies in the motion planning algorithm based on an error nondeterministic model that is added on the basis of the classical PBVS visual servo method [22]. By fusing motion error and sensing error, the correlation model between the nondeterministic error distribution at the end of the manipulator in the closed-loop visual servo process and the joint space state of the system is established.…”

Section: Introductionmentioning

confidence: 99%

An optimal visual servo trajectory planning method for manipulators based on system nondeterministic model

Tang

Zhang

2022

Robotica

View full text Add to dashboard Cite

When a manipulator captures its target by a visual servo system, uncertainties can arise because of mechanical system and visual sensors exist error. This paper proposes an intelligent method to predict the successful rate for a manipulator to capture its target with motion and sensor errors. Because the mapping between the joint space of the manipulator and the Cartesian space at the end of the manipulator is nonlinear, when there is a bounded error of the manipulator’s joint, the error range of the end motion is constantly changing with the different joint positions. And at the same time, the visual servo camera will also measure the target from different positions and postures, so as to produce measurement results with different error ranges. The unknown time-varying error property not only greatly affects the stability of the closed-loop control but also causes the capture failure. The purpose of this paper is to estimate the success probability of different capture trajectories by establishing the nondeterministic model of manipulator control system. First, a system model including motion subsystem and feedback subsystem was established with system error described by Gaussian probability. And then Bayesian estimation was introduced into the system model to estimate the executing state of the predefined trajectory. Linear least quadratic regulators (LQR) control is used to simulate the input correction in the closed-loop control between motion subsystem and feedback subsystem. At last, the successful probability of capturing the target is established by the Gaussian distribution at the end point of the trajectory with geometric relationship calculation between tolerance range and error distribution. The effectiveness and practicability of the proposed method are proved by simulation and experiment.

show abstract

“…Meng Kang et al proposed an integrated VS method with the extreme learning machine and Q-learning to address the control gain that is heuristically a constant. 20 This control gain is adjusted adaptively by a determined policy. However, these adaptive visual control methods encounter bottlenecks because they are defined in the low-dimensional space.…”

Section: Introductionmentioning

confidence: 99%

Visual servoing with deep reinforcement learning for rotor unmanned helicopter

Cao

Ning

2022

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

Visual servoing is a key approach to achieve visual control for the rotor unmanned helicopter. The challenges of the inaccurate matrix estimation and the target loss restrict the performance of the visual servoing control systems. This work proposes a novel visual servoing controller using the deep Q-network to achieve an efficient matrix estimation. A deep Q-network learning agent learns a policy estimating the interaction matrix for visual servoing of a rotor unmanned helicopter using continuous observation. The observation includes a combination of feature errors. The current matrix and the desired matrix constitute the action space. A well-designed reward guides the deep Q-network agent to get a policy to generate a time-varying linear combination between the current matrix and the desired matrix. Then, the interaction matrix is calculated by the linear combination. The potential mapping between the observation and the interaction matrix is learned by cascading the deep neural network layers. Experimental results show that the proposed method achieves faster convergence and lower target loss probability in tracking than the visual servoing methods with the fixed parameter.

show abstract

Adaptive visual servoing with an uncalibrated camera using extreme learning machine and Q-leaning

Cited by 35 publications

References 31 publications

Adaptive dynamic programming-based feature tracking control of visual servoing manipulators with unknown dynamics

Adaptive dynamic programming-based feature tracking control of visual servoing manipulators with unknown dynamics

An optimal visual servo trajectory planning method for manipulators based on system nondeterministic model

Visual servoing with deep reinforcement learning for rotor unmanned helicopter

Contact Info

Product

Resources

About