Rongxin Cui

Neural Network-Based Motion Control of an Underactuated Wheeled Inverted Pendulum Model

In this paper, automatic motion control is investigated for wheeled inverted pendulum (WIP) models, which have been widely applied for modeling of a large range of two wheeled modern vehicles. First, the underactuated WIP model is decomposed into a fully actuated second-order subsystem Σa consisting of planar movement of vehicle forward motion and yaw angular motions, and a passive (nonactuated) first-order subsystem Σb of pendulum tilt motion. Due to the unknown dynamics of subsystem Σa and universal approximation ability of neural network (NN), an adaptive NN scheme has been employed for motion control of subsystem Σa. Model reference approach has been used, whereas the reference model is optimized by finite time linear quadratic regulation technique. Inspired by human control strategy of inverted pendulum, the tilt angular motion in the passive subsystem Σb has been indirectly controlled using the dynamic coupling with planar forward motion of subsystem Σa, such that the satisfactory tracking of set tilt angle can be guaranteed. Rigorous theoretic analysis has been established, and simulation studies have been performed to demonstrate the developed method.

Mutual Information-Based Multi-AUV Path Planning for Scalar Field Sampling Using Multidimensional RRT*

Autonomous underwater vehicles (AUVs) have been widely employed in ocean survey, monitoring, and search and rescue tasks for both civil and military applications. It is beneficial to use multiple AUVs that perform environmental sampling and sensing tasks for the purposes of efficiency and cost effectiveness. In this paper, an adaptive path planning algorithm is proposed for multiple AUVs to estimate the scalar field over a region of interest. In the proposed method, a measurable model composed of multiple basis functions is defined to represent the scalar field. A selective basis function Kalman filter is developed to achieve model estimation through the information collected by multiple AUVs. In addition, a path planning method, the multidimensional rapidly exploring random trees star algorithm, which uses mutual information, is proposed for the multi-AUV system. Employing the path planning algorithm, the sampling positions of the AUVs are determined to improve the quality of future samples by maximizing the mutual information between the scalar field model and observations. Extensive simulation results are provided to demonstrate the effectiveness of the proposed algorithm. Additionally, an indoor experiment using four robotic fishes is carried out to validate the algorithms presented.

Adaptive Neural Network Control of AUVs With Control Input Nonlinearities Using Reinforcement Learning

In this paper, we investigate the trajectory tracking problem for a fully actuated autonomous underwater vehicle (AUV) that moves in the horizontal plane. External disturbances, control input nonlinearities and model uncertainties are considered in our control design. Based on the dynamics model derived in the discrete-time domain, two neural networks (NNs), including a critic and an action NN, are integrated into our adaptive control design. The critic NN is introduced to evaluate the long-time performance of the designed control in the current time step, and the action NN is used to compensate for the unknown dynamics. To eliminate the AUV’s control input nonlinearities, a compensation item is also designed in the adaptive control. Rigorous theoretical analysis is performed to prove the stability and performance of the proposed control law. Moreover, the robustness and effectiveness of the proposed control method are tested and validated through extensive numerical simulation results.

Extended State Observer-Based Integral Sliding Mode Control for an Underwater Robot With Unknown Disturbances and Uncertain Nonlinearities

This paper develops a novel integral sliding mode controller (ISMC) for a general type of underwater robots based on multiple-input and multiple-output extended-state-observer (MIMO-ESO). The difficulties associated with the unmeasured velocities, unknown disturbances, and uncertain hydrodynamics of the robot have been successfully solved in the control design. An adaptive MIMO-ESO is designed not only to estimate the unmeasurable linear and angular velocities, but also to estimate the unknown external disturbances. An ISMC is then designed using Lyapunov synthesis, and an adaptive gain update algorithm is introduced to estimate the upper bound of the uncertainties. Rigorous theoretical analysis is performed to show that the proposed control method is able to achieve asymptotical tracking performance for the underwater robot. Experimental studies are also carried out to validate the effectiveness of the proposed control, and to show that the proposed approach performs better than a conventional potential difference (PD) control approach.

Neural network based reinforcement learning control of autonomous underwater vehicles with control input saturation

In this paper, the trajectory tracking control of the autonomous underwater vehicle (AUV) has been investigated in discrete time, for ease of digital computer calculation. A reinforcement learning scheme is employed using two neural networks, whereas the first one is to compensate for uncertainties for the controller, and the second one is to estimate the evaluation function, such that optimal tracking performance could be achieve for the AUV. Simulation results show that the errors convergence to a adjustable neighborhood around zero, and optimization has been achieved in the sense of reinforcement learning.

A Sampling-Based Bayesian Approach for Cooperative Multiagent Online Search With Resource Constraints

This paper presents a cooperative multiagent search algorithm to solve the problem of searching for a target on a 2-D plane under multiple constraints. A Bayesian framework is used to update the local probability density functions (PDFs) of the target when the agents obtain observation information. To obtain the global PDF used for decision making, a sampling-based logarithmic opinion pool algorithm is proposed to fuse the local PDFs, and a particle sampling approach is used to represent the continuous PDF. Then the Gaussian mixture model (GMM) is applied to reconstitute the global PDF from the particles, and a weighted expectation maximization algorithm is presented to estimate the parameters of the GMM. Furthermore, we propose an optimization objective which aims to guide agents to find the target with less resource consumptions, and to keep the resource consumption of each agent balanced simultaneously. To this end, a utility function-based optimization problem is put forward, and it is solved by a gradient-based approach. Several contrastive simulations demonstrate that compared with other existing approaches, the proposed one uses less overall resources and shows a better performance of balancing the resource consumption.

Brain–Machine Interfacing-Based Teleoperation of Multiple Coordinated Mobile Robots

This paper describes the development of a teleoperation control framework of multiple coordinated mobile robots through a brain–machine interface (BMI). Utilizing the remote images of an environment, transferred to the human operator, visual compressive feedback loop produces imagine errors in nonvector space, where images are considered as a set without image processing of feature extraction. Given an initial set and a goal set, visual evoked potentials are used to generate EEG motion commands to make the image set converge to the goal set. The online BMI, utilizing steady-state visually evoked potentials, analyzes the human EEG data in such a format that human intentions can be recognized by AdaBoostSVM classifier and motion commands produced for the teleoperated robot. Bezier curve is utilized to parameterize the motion commands and leader–follower formation control is proposed to guarantee a good reference trajectory tracking performance. Extensive experimental studies have been carried out to assess the effectiveness of the proposed approaches.

Adaptive control of robot system of up to a half passive joints

In this paper, we study adaptive control of a robot system of \(2n\) joints with up to \(n\) joints being passive. By exploiting the dynamics couplings between the active joints and the passive joints, we have developed a method to use desired trajectories of active joints to indirectly “control” the motion of the passive joints. Optimal control techniques have been employed to control the active joints with smooth motion of minimized acceleration. Neural network (NN) has been used for block function approximation, in order to generate ideal desired trajectory of active joints. It has been theoretically established that under the developed adaptive controller and NN based trajectory generator, the passive joints can be effectively controlled to follow the predefined trajectory.

DMP and GMR based teaching by demonstration for a KUKA LBR robot

This paper investigates the problem of Teaching by Demonstration (TbD) on a KUKA lightweight robot (LBR). Motions are recorded by a human operator, and then the data is used to model a nonlinear system, i.e., the dynamic motor primitive (DMP). In order to learn from multiple demonstrations, Gaussian Mixture Models (GMM) are employed rather than using conventional Gaussian process for the evaluation of the non-linear term of the DMP. Then the Gaussian mixture regression (GMR) algorithm is applied to generate a synthesized trajectory with smaller position errors in 3D space. The proposed approach is tested and demonstrated by performing two tasks with KUKA iiwa robot.

Maximum Power Tracking Control for Current Power System Based on Fuzzy-PID Controller

In order to solve the problem of maximum energy capture in the dynamic changing current, the maximum power point tracking MPPT control strategy of an extensible vertical-axis blade current power generation system was developed. Consider the current speed with the characteristics of randomness and uncertainties. Affixed set of PID parameters can hardly achieve desired generator speed. Therefore, this paper presents a Fuzzy-PID control, which combines the characteristics of traditional PID and fuzzy control. The simulation results show that the Fuzzy-PID control is able to guarantee desired dynamic characteristics, such as fast response, robustness and stability, and is able to effectively track the maximum power point.