Guangjun Liu

Identification methods for Hammerstein nonlinear systems

This paper considers the identification problems of the Hammerstein nonlinear systems. A projection and a stochastic gradient (SG) identification algorithms are presented for the Hammerstein nonlinear systems by using the gradient search method. Since the projection algorithm is sensitive to noise and the SG algorithm has a slow convergence rate, a Newton recursive and a Newton iterative identification algorithms are derived by using the Newton method (Newton-Raphson method), in order to reduce the sensitivity of the projection algorithm to noise, and to improve convergence rates of the SG algorithm. Furthermore, the performances of these approaches are analyzed and compared using a numerical example, including the parameter estimation errors, the stationarity and convergence rates of parameter estimates and the computational efficiency.

Auxiliary model based multi-innovation extended stochastic gradient parameter estimation with colored measurement noises

For pseudo-linear regression identification models corresponding output error systems with colored measurement noises, a difficulty of identification is that there exist unknown inner variables and unmeasurable noise terms in the information vector. This paper presents an auxiliary model based multi-innovation extended stochastic gradient algorithm by using the auxiliary model method and by expanding the scalar innovation to an innovation vector. Compared with single innovation extended stochastic gradient algorithm, the proposed approach can generate highly accurate parameter estimates. The simulation results confirm this conclusion.

Gradient based and least-squares based iterative identification methods for OE and OEMA systems

Gradient based and least-squares based iterative identification algorithms are developed for output error (OE) and output error moving average (OEMA) systems. Compared with recursive approaches, the proposed iterative algorithms use all the measured input-output data at each iterative computation (at each iteration), and thus can produce highly accurate parameter estimation. The basic idea of the iterative methods is to adopt the interactive estimation theory: the parameter estimates relying on unknown variables are computed by using the estimates of these unknown variables which are obtained from the preceding parameter estimates. The simulation results confirm theoretical findings.

Multiinnovation Least-Squares Identification for System Modeling

A multiinnovation least-squares (MILS) identification algorithm is presented for linear regression models with unknown parameter vectors by expanding the innovation length in the traditional recursive least-squares (RLS) algorithm from the viewpoint of innovation modification. Because the proposed MILS algorithm uses p innovations (not only the current innovation but also past innovations) at each iteration (with the integer p > 1 being an innovation length), the accuracy of parameter estimation is improved, compared with that of the RLS algorithm. Performance analysis and simulation results show that the proposed MILS algorithm is consistently convergent. Moreover, a new interval-varying MILS algorithm is proposed, for which the key is to dynamically change the interval in order to deal with cases where some measurement data are missing. Furthermore, an auxiliary-model-based MILS algorithm is derived for pseudolinear models corresponding to output error moving average systems with colored noises. Finally, the proposed algorithms are applied to model an experimental water level control system.

Estimation of Battery State of Charge With

Battery state-of-charge (SOC) estimation is essential for a mobile robot, such as inspection of power transmission lines. It is often estimated using a Kalman filter (KF) under the assumption that the statistical properties of the system and measurement errors are known. Otherwise, the SOC estimation error may be large or even divergent. In this paper, without the requirement of the known statistical properties, a SOC estimation method is proposed using an H∞ observer, which can still guarantee the SOC estimation accuracy in the worst statistical error case. Under the conditions of different currents and temperatures, the effectiveness of the proposed method is verified in the laboratory and field environments. With the comparison of the proposed method and the KF-based one, the experimental results show that the proposed method can still provide accurate SOC estimation when there exist inexact or unknown statistical properties of the errors. The proposed method has been applied successfully to the robot for inspecting the running 500-kV extra high voltage power transmission lines.

H_{\infty}

In this paper, a linear parametric friction model is formulated by linearizing a nonlinear empirical friction model in parameters. A proposed decomposition-based control design framework is applied to synthesize the friction compensation scheme. A separate compensator is designed for each type of friction utilizing the most suitable control technique. The nominal friction is compensated by feed forward. An adaptive compensator is derived to compensate for parametric unmodeled friction with unknown but constant parameters, and a robust compensator is used to deal with friction model parameter variations, as well as non-parametric unmodeled friction. The combination of the compensators yields the overall compensation scheme. Adaptive and robust compensators complement each other in compensating the effects of model uncertainties. The analytical and simulation studies have confirmed the efficiency of the proposed friction compensation method.

Observer: Applied to a Robot for Inspecting Power Transmission Lines

A major technical challenge in controlling modular and reconfigurable robots is associated with the kinematics and dynamic model uncertainties caused by reconfiguration. In parallel, conventional model uncertainties such as uncompensated joint friction still persist. This paper presents a modular distributed control technique for modular and reconfigurable robots that can instantly adapt to robot reconfigurations. Under the proposed control method that is based on joint torque sensing, a modular and reconfigurable robot is stabilized joint by joint, and modules can be added or removed without the need to adjust control parameters of the other modules of the robot. Model uncertainties associated with link and payload masses are compensated using joint torque sensor measurement. The remaining model uncertainties, including uncompensated dynamic coupling and joint friction, are compensated by a decomposition-based robust controller. Simulation results have confirmed the effectiveness of the proposed method.

Decomposition-based friction compensation of mechanical systems

This paper analyzes track-stair interactions and develops an online tipover prediction algorithm for a self-reconfigurable tracked mobile robot climbing stairs, which is vulnerable to tipping-over. Tipover prediction and prevention for a tracked mobile robot in stair climbing are intractable problems because of the complex track--stair interactions. Unlike the wheeled mobile robots, which are normally assumed to obey the nonholonomic constraints, slippage is unavoidable for a tracked mobile robot, especially in stair climbing. Furthermore, the track-stair interactive forces are complicated, which may take the forms of grouser-tread hooking force, track--stair edge frictional force, grouser-riser clutching force, and even their compositions. In this paper, the track--stair interactions are analyzed systematically, and tipover stability criteria are derived for a tracked mobile robot climbing stairs. An online tipover prediction algorithm is also developed, which forms an essential part for autonomous and semiautonomous stair-climbing control. The effectiveness of the proposed algorithms are verified by experiments.

Distributed control of modular and reconfigurable robot with torque sensing

Conventional robot manipulators produce poor payload to weight ratio and limited manipulation ability, as a significant portion of available actuation force is used to balance their own weight. The same cause also limits the robots operation capability in terms of acceleration and manipulation force. Such problems become more severe for modular and reconfigurable robots (MRRs) when they are expanded by adding predesigned modules. Static balancing with counterweights and external springs can greatly improve a robots payload and manipulation capabilities, but require sophisticated mechanisms and restrict the working envelope of the robot. In this paper, an innovative spring-assisted MRR design and control framework is presented, which is developed based on a synergetic integration of robot control with a brake and an embedded spring at each modular joint. The spring is inserted between the brake and the motor shaft through a decoupling bearing. By activating the brake, static balancing can be established at any desirable position of each module and any configuration of the robot, allowing reinforced delicate operation in a neighborhood of the balanced configuration such as door opening, as well as spring-assisted lift of heavy payload. A distributed control method has been proposed to facilitate control of the spring-assisted MRRs, which does not rely on a priori dynamic models, and can suppress uncertainties caused by reconfigurations, eliminating the need to readjust control parameters of the lower modules when new modules are added or removed. Prototype modules have been developed, and the experimental results have confirmed the effectiveness of the proposed design and control.

Track--Stair Interaction Analysis and Online Tipover Prediction for a Self-Reconfigurable Tracked Mobile Robot Climbing Stairs

Nonlinear torsional compliance and hysteresis are associated with harmonic drives, and their accurate modeling is crucial for improving performance of the control system of harmonic drive-based devices such as robot joints. In this paper, a new approach is taken to model the torsional compliance and hysteresis behavior in harmonic drives. The proposed model is derived by modeling the compliance behavior of the flexspline and the wave generator instead of modeling the individual behaviors of the overall harmonic drive transmission. The hysteresis loss is captured by taking the wave generator torsional compliance into account. The proposed model is validated through numerical simulations and subsequently with experimental data.