Neural-network control of mobile manipulators

Optimal robust constraint following control design and experimental validation for fuzzy robotic manipulator system

As a typically complex mechanical system, robotic manipulator has the characteristics of complex structure, parameter uncertainty and vulnerability to external interference. From the perspective of second-order servo constraints, this paper proposes a robust constraint following control algorithm, which can meet the control requirements of robotic manipulator system under system uncertainty, which is described by fuzzy set theory. It also guarantees the deterministic performance of uniform boundedness (UB) and uniform ultimate boundedness (UUB). The system performance index function based on fuzzy information is established, and by minimizing the system performance index function, the optimization design problem of the controller gain parameter is solved. Numerical simulations and experimental results verify the efficacy of the robust controller and control parameter optimization method.

Learning positioning policies for mobile manipulation operations with deep reinforcement learning

This work focuses on the operation of picking an object on a table with a mobile manipulator. We use deep reinforcement learning (DRL) to learn a positioning policy for the robot’s base by considering the reachability constraints of the arm. This work extends our first proof-of-concept with the ultimate goal of validating the method on a real robot. Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm is used to model the base controller, and is optimised using the feedback from the MoveIt! based arm planner. The idea is to encourage the base controller to position itself in areas where the arm reaches the object. Following a simulation-to-reality approach, first we create a realistic simulation of the robotic environment in Unity, and integrate it in Robot Operating System (ROS). The drivers for both the base and the arm are also implemented. The DRL-based agent is trained in simulation and, both the robot and target poses are randomised to make the learnt base controller robust to uncertainties. We propose a task-specific setup for TD3, which includes state/action spaces, reward function and neural architectures. We compare the proposed method with the baseline work and show that the combination of TD3 and the proposed setup leads to a $$11\%$$ 11 % higher success rate than with the baseline, with an overall success rate of $$97\%$$ 97 % . Finally, the learnt agent is deployed and validated in the real robotic system where we obtain a promising success rate of $$75\%$$ 75 % .

Rios J, Gomez-Avila J, Arana-Daniel N, Lopez-Franco C, Alanis AY. Adaptive neural PD controllers for mobile manipulator trajectory tracking

Artificial intelligence techniques have been used in the industry to control complex systems; among these proposals, adaptive Proportional, Integrative, Derivative (PID) controllers are intelligent versions of the most used controller in the industry. This work presents an adaptive neuron PD controller and a multilayer neural PD controller for position tracking of a mobile manipulator. Both controllers are trained by an extended Kalman filter (EKF) algorithm. Neural networks trained with the EKF algorithm show faster learning speeds and convergence times than the training based on backpropagation. The integrative term in PID controllers eliminates the steady-state error, but it provokes oscillations and overshoot. Moreover, the cumulative error in the integral action may produce windup effects such as high settling time, poor performance, and instability. The proposed neural PD controllers adjust their gains dynamically, which eliminates the steady-state error. Then, the integrative term is not required, and oscillations and overshot are highly reduced. Removing the integral part also eliminates the need for anti-windup methodologies to deal with the windup effects. Mobile manipulators are popular due to their mobile capability combined with a dexterous manipulation capability, which gives them the potential for many industrial applications. Applicability of the proposed adaptive neural controllers is presented by simulating experimental results on a KUKA Youbot mobile manipulator, presenting different tests and comparisons with the conventional PID controller and an existing adaptive neuron PID controller.

Neural Adaptive Backstepping Control of a Robotic Manipulator With Prescribed Performance Constraint

This paper presents an adaptive neural network (NN) control of a two-degree-of-freedom manipulator driven by an electrohydraulic actuator. To restrict the system output in a prescribed performance constraint, a weighted performance function is designed to guarantee the dynamic and steady tracking errors of joint angle in a required accuracy. Then, a radial-basis-function NN is constructed to train the unknown model dynamics of a manipulator by traditional backstepping control (TBC) and obtain the preliminary estimated model, which can replace the preknown dynamics in the backstepping iteration. Furthermore, an adaptive estimation law is adopted to self-tune every trained-node weight, and the estimated model is online optimized to enhance the robustness of the NN controller. The effectiveness of the proposed control is verified by comparative simulation and experimental results with Proportional-integral-derivative and TBC methods.

Robust adaptive motion/force control scheme for crawler-type mobile manipulator with nonholonomic constraint based on sliding mode control approach

In this paper, a robust adaptive motion/force control (RAMFC) scheme is presented for a crawler-type mobile manipulator (CTMM) with nonholonomic constraint. For the position tracking control design, an adaptive sliding mode tracking controller is proposed to deal with the unknown upper bounds of system parameter uncertainties and external disturbances. Based on the position tracking results, a robust control strategy is also developed for the nonholonomic constraint force of CTMM. According to the Lyapunov stability theory, the stability of the closed-loop control system, the uniformly ultimately boundedness of position tracking errors, and the boundedness of the force error and adaptive coefficient errors are all guaranteed by using the derived RAMFC scheme. Simulation and experimental tests on a CTMM with two-link manipulator demonstrate the effectiveness and robustness of the proposed control scheme.

Adaptive Neural Feedback Linearizing Control of Type (m,s) Mobile Manipulators with a Guaranteed Prescribed Performance

In this paper, a neural network (NN)-based tracking controller is proposed for a general class of type (m,s) wheeled mobile manipulators (WMMs) subjected to model uncertainties with prescribed transient and steady-state performance specifications. First, an input–output model of WMMs is derived by introducing proper output equations. Then, the prescribed performance technique is employed to propose a proportional integral derivative trajectory tracking controller for WMMs to ensure that the tracking errors converge to a smaller, arbitrary ultimate bound with a predefined maximum overshoot/undershoot and convergence speed. The learning capabilities of multilayer NNs are incorporated into the controller to approximate the uncertain nonlinear dynamics of the robot. An adaptive saturation-type controller is utilized to compensate NN estimation errors and external disturbances. A Lyapunov-based stability analysis is used to demonstrate that the tracking errors are uniformly ultimately bounded and converge to a small neighborhood of zero with a guaranteed prescribed performance. Numerical computer simulations are presented to show the effectiveness of the proposed controller.

Trajectory Tracking and Stability Analysis for Mobile Manipulators Based on Decentralized Control

Trajectory tracking of a mobile manipulator in the Cartesian space based on decentralized control is considered in this paper. The dynamic model is first rearranged to take the form of two interconnected subsystems with constraint flow, namely, a nonholonomic mobile platform subsystem and a holonomic manipulator subsystem. Secondly, using the inverse kinematics, the workspace desired trajectory of the mobile manipulator is transformed to the manipulator joint space as well as the platform desired trajectory. The kinematic control is developed from the desired trajectory of the platform. Then, the desired velocity is derived using the kinematic controller of the mobile platform, after which the velocity is used to obtain the control law of the mobile platform subsystem. Thirdly, the control law of the manipulator subsystem is developed based on the desired and real values of the manipulator, as well as the desired velocity. According to the Lyapunov stability theory, the proposed decentralized control strategy guarantees the global stability of the closed-loop system, and the tracking errors are bounded. Experimental results obtained on a 3-DOF manipulator mounted on a mobile platform are given to demonstrate the feasibility and effectiveness of the proposed approach. This is confirmed by a comparison with the computed torque approach.

Dual closed-loop sliding mode control for a decoupled three-link wheeled mobile manipulator

This paper presents a dual closed-loop sliding mode control strategy for a wheeled mobile manipulator with three-wheeled mobile platform (WMP) and three-link manipulator. The Euler-Lagrange method combined partially with the Newtonian method is applied to obtain full dynamic model and decoupled model is constructed in order to provide simple dynamic model for controller's structure to be simplified. Instead of the conventional velocity command trajectory based kinematic backstepping control method, a dual closed-loop control system is designed. A virtual velocity command based on sliding mode surface is generated in outer loop and the gap between a generated virtual command velocity and real velocity is compensated by an inner loop sliding mode controller. Outer loop helps to faster posture trajectory generation for locomotion of the WMP. Next, a finite-time sliding mode controller with an assumed feedforward dynamic gain method is designed for joint trajectory tracking for three-link manipulator by adding finite-time control terms in the designed controllers to obtain faster settling time and stronger robustness. The designed controllers were implemented into microprocessor connected to DC and dynamixel motor systems equipped in mobile platform and manipulator, respectively. Comparative simulation and experiment with a conventional sliding mode control show the effectiveness of the proposed dual closed-loop finite time sliding mode control scheme.

An automatic switching approach to teleoperation of mobile-manipulator systems using virtual fixtures

This work presents a novel command strategy developed to improve operator performance and minimize difficulties in teleoperation tasks for mobile-manipulator systems with a holonomic base. Aimed specifically at novice operators, virtual fixtures are introduced as a means to minimize collisions and assist in navigation. Using the 6-degree-of-freedom (DOF) Omnibot mobile-manipulator system (MMS), a command strategy is implemented such that the operator need only control a 3-DOF haptic joystick to achieve full control of the Omnibot MMS. The command strategy is used to coordinate control between the arm and the base of the system, prevent collisions with known obstacles, and alert the operator of proximity to those obstacles with haptic forces. Through experimental testing it is shown that operator performance improved with the use of virtual fixtures.

Learning from adaptive neural network output feedback control of a unicycle-type mobile robot

This paper studies learning from adaptive neural network (NN) output feedback control of nonholonomic unicycle-type mobile robots. The major difficulties are caused by the unknown robot system dynamics and the unmeasurable states. To overcome these difficulties, a new adaptive control scheme is proposed including designing a new adaptive NN output feedback controller and two high-gain observers. It is shown that the stability of the closed-loop robot system and the convergence of tracking errors are guaranteed. The unknown robot system dynamics can be approximated by radial basis function NNs. When repeating same or similar control tasks, the learned knowledge can be recalled and reused to achieve guaranteed stability and better control performance, thereby avoiding the tremendous repeated training process of NNs.

Robust adaptive tracking control for nonholonomic mobile manipulator with uncertainties

In this paper, mobile manipulator is divided into two subsystems, that is, nonholonomic mobile platform subsystem and holonomic manipulator subsystem. First, the kinematic controller of the mobile platform is derived to obtain a desired velocity. Second, regarding the coupling between the two subsystems as disturbances, Lyapunov functions of the two subsystems are designed respectively. Third, a robust adaptive tracking controller is proposed to deal with the unknown upper bounds of parameter uncertainties and disturbances. According to the Lyapunov stability theory, the derived robust adaptive controller guarantees global stability of the closed-loop system, and the tracking errors and adaptive coefficient errors are all bounded. Finally, simulation results show that the proposed robust adaptive tracking controller for nonholonomic mobile manipulator is effective and has good tracking capacity.

Embedded Vision-Based Autonomous Move-to-Grasp Approach for a Mobile Manipulator

This paper proposes a vision-based autonomous move-to-grasp approach for a compact mobile manipulator under some low and small environments. The visual information of specified object with a radial symbol and an overhead colour block is extracted from two CMOS cameras in an embedded way. Furthermore, the mobile platform and the postures of the manipulator are adjusted continuously by vision-based control, which drives the mobile manipulator approaching the object. When the mobile manipulator is sufficiently close to the object, only the manipulator moves to grasp the object based on the incremental movement with its head end centre of the end-effector conforming to a Bezier curve. The effectiveness of the proposed approach is verified by experiments.

A Methodology for tele-operating mobile manipulators with an emphasis on operator ease of use

An algorithm for the tele-operation of mobile-manipulator systems with a focus on ease of use for the operator is presented. The algorithm allows for unified, intuitive, and coordinated control of mobile manipulators. It consists of three states. In the first state, a single 6-degrees-of-freedom (DOF) joystick is used to control the manipulator's position and orientation. The second state occurs when the manipulator approaches a singular configuration, resulting in the mobile base moving in a manner so as to keep the end-effector travelling in its last direction of motion. This is done through the use of a constrained optimization routine. The third state is entered when the operator returns the joystick to the home position. Both the mobile base and manipulator move with respect to one another keeping the end-effector stationary and placing the manipulator into an ideal configuration. The algorithm has been implemented on an 8-DOF mobile manipulator and the test results show that it is effective at moving the system in an intuitive manner.

A new iterative learning controller using variable structure fourier neural network

A new iterative learning control approach based on Fourier neural network (FNN) is presented for the tracking control of a class of nonlinear systems with deterministic uncertainties. The proposed controller consists of two loops. The inner loop is a feedback control action that decreases system variability and reduces the influence of random disturbances. The outer loop is an FNN-based learning controller that generates the system input to suppress the error caused by system nonlinearities and deterministic uncertainties. The FNN employs orthogonal complex Fourier exponentials as its activation functions. Therefore, it is essentially a frequency-domain method that converts the tracking problem in the time domain into a number of regulation problems in the frequency domain. Through a novel phase compensation technique, this model-free method makes it possible to use higher-frequency components in the FNN to improve the tracking performance. In addition, the structure of the FNN can be reconfigured according to the system output information to make the learning more efficient and increase the convergent speed of the tracking error. Experiments on both a commercial gear box and a belt-driven positioning table are conducted to show the effectiveness of the proposed controller.

Adaptive-Fourier-neural-network-based control for a class of uncertain nonlinear systems

An adaptive Fourier neural network (AFNN) control scheme is presented in this paper for the control of a class of uncertain nonlinear systems. Based on Fourier analysis and neural network (NN) theory, AFNN employs orthogonal complex Fourier exponentials as the activation functions. Due to the clear physical meaning of the neurons, the determination of the AFNN structure as well as the parameters of the activation functions becomes convenient. One salient feature of the proposed AFNN approach is that all the nonlinearities and uncertainties of the dynamical system are lumped together and compensated online by AFNN. It can, therefore, be applied to uncertain nonlinear systems without any a priori knowledge about the system dynamics. Derived from Lyapunov theory, a novel learning algorithm is proposed, which is essentially a frequency domain method and can guarantee asymptotic stability of the closed-loop system. The simulation results of a multiple-input-multiple-output (MIMO) nonlinear system and the experimental results of an X - Y positioning table are presented to show the effectiveness of the proposed AFNN controller.

Robust Control of Wheeled Mobile Manipulators Using Hybird Joints

In this paper, robust force/motion control strategies are presented for mobile manipulators under both holonomic and nonholonomic constraints in the presence of uncertainties and disturbances. The proposed control strategies guarantee that the system motion converges to the desired manifold with prescribed performance, and constraint force control is developed using the passivity of hybrid joint rather than force feedback control. Experiment results validate that not only the states of the system asymptotically converge to the desired trajectory, but also the constraint force asymptotically converges to the desired force.

Adaptive Robust Motion/Force Control of Holonomic-Constrained Nonholonomic Mobile Manipulators

In this paper, adaptive robust force/motion control strategies are presented for mobile manipulators under both holonomic and nonholonomic constraints in the presence of uncertainties and disturbances. The proposed control is robust not only to parameter uncertainties such as mass variations but also to external ones such as disturbances. The stability of the closed-loop system and the boundedness of tracking errors are proved using Lyapunov stability synthesis. The proposed control strategies guarantee that the system motion converges to the desired manifold with prescribed performance and the bounded constraint force. Simulation results validate that the motion of the system converges to the desired trajectory, and the constraint force converges to the desired force.

An integrated agent-based software architecture for mobile and manipulator systems

Mobile manipulation involves the most important key issue in robotics: integration. While hardware integration seems to be nearly solved due to the increasing dominance of PC-compatible systems, software integration is still a challenge, since a lot of issues arise with the variety of operating systems, device drivers, application libraries, and programming languages which need to be merged in any real-world robotic system. This paper presents a software architecture, which seamlessly integrates robot arms, mobile bases, vision systems and sensing devices, in a distributed, homogeneous agent framework. Based on the Java platform, the agent-based architecture allows great flexibility in the integration of components, and provides a simple yet extensible and powerful software layer to develop further mobile manipulating environments. Detailed software issues, as well as preliminary results are shown, which pave the way towards the development of network-ready applications involving mobile and manipulating artifacts.

Design and Stabilization of Sampled-Data Neural-Network-Based Control Systems

This paper presents the design and stability analysis of a sampled-data neural-network-based control system. A continuous-time nonlinear plant and a sampled-data three-layer fully connected feedforward neural-network-based controller are connected in a closed loop to perform the control task. Stability conditions will be derived to guarantee the closed-loop system stability. Linear-matrix-inequality- and genetic-algorithm-based approaches will be employed to obtain the largest sampling period and the connection weights of the neural network subject to the considerations of the system stability and performance. An application example will be given to illustrate the design procedure and effectiveness of the proposed approach.

Robust Motion Control for Mobile Manipulator Using Resolved Acceleration and Proportional-Integral Active Force Control

A resolved acceleration control (RAC) and proportional-integral active force control (PIAFC) is proposed as an approach for the robust motion control of a mobile manipulator (MM) comprising a differentially driven wheeled mobile platform with a two-link planar arm mounted on top of the platform. The study emphasizes on the integrated kinematic and dynamic control strategy in which the RAC is used to manipulate the kinematic component while the PIAFC is implemented to compensate the dynamic effects including the bounded known/unknown disturbances and uncertainties. The effectivenss and robustness of the proposed scheme are investigated through a rigorous simulation study and later complemented with experimental results obtained through a number of experiments performed on a fully developed working prototype in a laboratory environment. A number of disturbances in the form of vibratory and impact forces are deliberately introduced into the system to evaluate the system performances. The investigation clearly demonstrates the extreme robustness feature of the proposed control scheme compared to other systems considered in the study.