Neural Network Compensation for Impedance Force Controlled Robot Manipulators

This article presents a comparative study between the hybrid impedance control approach and the impedance force control approach with neural networks compensation of manipulator modeling errors. The proposed algorithm consists of an outer hybrid impedance control loop that generates the reference (target) acceleration to an inner inverse dynamics control loop. In order to improve the controller robustness, a compensation action of the manipulator modeling errors is introduced, acting on the target acceleration. This compensation action is based on a neural network model achieved by minimizing the modeling errors along the manipulator trajectory. The neural network algorithm uses an error training signal to model errors, that is minimized along the trajectory. The performance of the hybrid impedance control system with neural network compensation, is compared with the impedance force controller with neural network compensation, and is illustrated by computer simulations with a two degree-of-freedom PUMA 560 robot, whose end-effector is forced to move along a frictionless surface located perpendicular to the horizontal plane. The results obtained reveal even better performance when the desired force profile is not constant along the trajectory. This situation is very important in fine motion tasks where the desired force must have a varying profile along the trajectory, namely at the beginning and the end of the task. The results have also shown accurate force tracking and position control, even when significant uncertainties in the robot dynamic model are assumed.

Visual pursuit tracking tasks developed to quantitatively assess neuroprosthetic hand grasp systems are discussed. The tests evaluate patient control of the hand grasp and measure the system input-output relationships which can be used to identify nonlinearities, deadbands, and inappropriate slopes in these relationships. Subjects track using the combined force and position outputs of the hand grasp. The feasibility of combined tracking was evaluated using three normal subjects. Two of the subjects performed significantly better at position tracking than at force or combined tracking. The third subject's position tracking was not significantly better than force tracking. All three performed significantly better at force tracking than at combined tracking. >

In this paper an intelligent prediction algorithm for robot force control is reviewed. The algorithm is not only applied to impedance controller, but also hybrid position/force controller. The experimental platform with open architecture controller is developed to test the force tracking effect when the environmental change in curvature and stiffness is taken into account. Using the system an easily open architecture is achieved in hardware and software, which is able to deal with the implementation of force tracking tasks of unknown contact environment performed by robot manipulators. The force tracking experiments for the irregular surface in different control model, different tracking velocity, different desired force, different contact stiffness are executed. The comparative experimental results show validity of the proposed method.

This paper presents a data-driven approach towards time-optimal trajectory generation for Unmanned Aerial Vehicles (UAV's) using a machine-learned trajectory generation mechanism for point-to-point time-optimal trajectories on-the-fly. To train this machine-learned black box trajectory generator off-line, a model-based optimization problem is first constructed for point-to-point time-optimal trajectory generation, with physical constraints on inputs, states, and rates. The formulated optimization problem is then solved off-line for a range of initial and terminal flight states to generate point-to-point data-sets that consist of the optimal state and input trajectories. This information is compressed by parameterizing the input and state trajectories using a set of basis functions. This data is then used to train the neural network-based trajectory planner. The output of the neural network is the basis function coefficient sets for the state and input trajectories (and the total flight time) which can then be used to reconstruct the flight trajectory. Once the neural network is trained, the data-driven on-board trajectory generator is ready to be deployed on the UAV for on-board planning. This approach is demonstrated for two scenarios: (1) the input to the neural network being the initial and terminal flight states and (2) the input to the neural network being initial and terminal flight states as well as physical constraints. To validate the performance of the machine-learned black-box trajectory generator, the root mean squared error between the neural network generated trajectories and the trajectories obtained from solving the optimization problem directly is statistically evaluated. These trajectories are also tested for violation of path constraints (which are not included explicitly in the training or input to the black box planner) by evaluating the mean constraint violation for each path-constrained variable.

Real-time motion planning and control of multi-joint robot manipulators are studied using neural networks. The proposed neural network approach consists of two modules: one for real-time collision-free motion planning, the other for real-time fine control of the robot manipulators. The motion planning module is a biologically inspired, parallel connected, topologically organised neural network, where each neuron is characterised by a shunting equation, whose state space is the robot configuration joint space. This module is capable of planning a real-time optimal path for robot manipulators through the dynamic activity landscape of the neural network. The motion control module involves of a feedforward neural network together with a traditional PD feedback loop. It is capable of achieving real-time fine motion control of robot manipulators under significant uncertainties and without any prior knowledge of the robot dynamics. This module can quickly compensate sudden changes in the robot dynamics. The real-time fine control of robot manipulators is achieved through the online learning of the neural network. Both the motion planning and motion control modules are computationally efficient, and their global stabilities are proved using Lyapunov stability theory.

In this paper, the effects of time delay and packet loss incurred in communication channel, are examined on the performance of bilaterally controlled hydraulic actuators under various control schemes. The controllers are force reflection (FR), position error (PE), combined force reflection and position error (FP), and four channels (FC). The performance of each control scheme is evaluated by performing a set of experiments on a test rig comprising haptic device, simulated valve-controlled hydraulic actuator, and communication channel. The performances are then compared using three criteria: position tracking, force tracking, and fidelity of the perceived stiffness. The experiments are performed under time delays ranging from 100 to 1,000 ms and percentage of packet losses up to 75%. It is shown that packet loss affects the signal oscillations, while time delay increases the steady-state tracking error. The FP controller shows the best performance with the least tracking error and signal oscillation in both position and force tracking. The FC scheme is shown to be least effective due to large tracking oscillations in the presence of packet loss and/or time delay. The value of stiffness perceived by the operator is found to decline with increasing time delays or packet losses in all controllers.

PurposeThis paper aims to present a smooth transition adaptive hybrid impedance control for compliant connector assembly.Design/methodology/approachThe dynamics of the manipulator is firstly presented with linear property. The controller used in connector assembly is inspired by human operation habits in similar tasks. The hybrid impedance control is adopted to apply force in the assembly direction and provide compliance in rest directions. The reference trajectory is implemented with an adaptive controller. Event-based switching strategy is conducted for a smooth transition from unconstrained to constrained space.FindingsThe method can ensure both ideal compliance behaviour with dynamic uncertainty and a smooth transition from unconstrained to constrained space. Also, the method can ensure compliant connector assembly with a good tolerance to the target estimation error.Practical implicationsThe method can be applied in the connector assembly by “pushing” operation. The controller devotes efforts on force tracking and smooth transition, having potential applications in contact tasks in delicate environment.Originality/valueAs far as the authors know, the paper is original in providing a uniform controller for improving force and position control performance in both unconstrained and constrained space with dynamic uncertainty. The proposed controller can ensure a smooth transition by only adjusting parameters.

This paper presents the feasibility study of an adaptive impedance force control application to deburring/milling process. The proposed control regulates desired force on nonhomogeneous workpiece in regardless of unknown location and stiffness of the workpiece. The adaptive technique of minimizing the force error between the desired and the actual force is used to compensate for uncertainties in unknown workpiece. Effect of selecting the impedance damping gain with different workpiece stiffness on force tracking is analytically determined. Analyzing the stability of the proposed adaptive force control yields the bound of the adaptive gain for stability. Simulation studies of deburring/milling process show the excellent force tracking and confirm stability bound for stable force tracking.

Contact force tracking of quadrotors based on robust attitude control

Robotic microinjection has been widely applied in biomedical engineering, but faces a great challenge on the precise force interaction with cells due to their inherently deformable, fragile, and nonlinear viscoelastic properties. This brief proposes a new neural-learning based adaptive sliding mode impedance force control scheme for the robotic microinjection system to improve the interaction performance. The key features of the developed method are as follows: i) Target impedance is derived by utilizing the nonlinear Hunt-Crossley model to match the microscale interaction with environmental cells; ii) An integral terminal sliding mode manifold based on the impedance error is designed to achieve finite-time convergence and accurate force tracking; iii) The proposed scheme relieves the burden of environmental model dependence by estimating the uncertain bound of external disturbances through an adaptive neural network compensator. The control system stability is analyzed by the Lyapunov theory, and the force tracking performance is also verified via a series of experiments.

Tele-impedance augments classic teleoperation by enabling the human operator to actively command remote robot stiffness in real-time, which is an essential ability to successfully interact with the unstructured and unpredictable environment. However, the literature is missing a study on benefits and drawbacks of different types of stiffness command interfaces used in bilateral tele-impedance. In this article, we introduce a term called coupling effect, which pertains to the coupling between human-commanded stiffness going to the remote robot and force feedback coming from the remote robot. We hypothesize that, whenever the operator's commanded stiffness and force feedback are subject to coupling effect (e.g., muscle activity based stiffness command interfaces), force feedback can invoke involuntary changes in the commanded stiffness due to human reflexes. Although the coupling effect takes away some degree of the operator's control over the commanded stiffness, these involuntary changes can be either beneficial (e.g., during position tracking) or detrimental (e.g., during force tracking) to the task performance on the remote robot side. We examined the coupling effect in an experimental study with <b xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">16</b> participants, who performed position and force tracking tasks by using both coupled type (muscle activity based) and decoupled type (external device based) of interface. The results demonstrate a benefit of the coupling effect when the remote robot is operating in presence of unexpected force perturbations, where lower absolute error in position tracking task was observed. On the other hand, the decoupled type of interface is beneficial for force tracking tasks on the remote robot side, such as establishing or maintaining a stable contact with objects. However, the coupling effect negatively influences the commanding of reference stiffness to the remote robot in both position and force tracking tasks for the coupled type of interface, compared to the decoupled type of interface, which is not affected.

Learning from adaptive neural output feedback control of robot manipulators

Sensor-based force decouple controller design of macro–mini manipulator

Force control of a single-rod electrohydraulic actuator is investigated. The model considers the friction effect of the actuator. First, a non-linear force controller using a feedback domination approach, which can globally finite-time stabilise the force tracking error system without the friction parameter uncertainties, is proposed. Then, another feedback domination controller is proposed to cope with the friction parameter uncertainties. To estimate the unmeasurable states, an observer with finite-time convergence is developed. Rigorous stability analysis is given. Compared with the conventional output feedback backstepping control method, the proposed output feedback domination control scheme provides a faster convergence rate and a higher tracking accuracy for force tracking. Numerical simulation results demonstrate the effectiveness of the proposed control design.

Simultaneous position and force tracking are two critical factors in teleportation systems to achieve transparency. In this paper, a new control scheme is proposed to guarantee asymptotic position and force tracking in teleportation systems with disturbances. To this end, two sliding mode controllers are designed for the master and the slave robots respectively and are incorporated into the 4-channel teleportation framework. In order to demonstrate the validity of the proposed control scheme, the performance of the overall teleportation control system is rigorously proved and analyzed in a mathematical way. The analysis results show that the designed teleportation control scheme can simultaneously achieve asymptotic position and force tracking performance of the master and the slave robots.