Control Engineering Practice

For proper manual aircraft control, the pilot has to perceive the motion state of the aircraft. In this perception process both the visual and the vestibular systems play an important role. To understand this perception process and its impact on a pilot's control behavior a descriptive model was developed. The single-channel information-processor model was applied as the basic structure of the final model. Three groups of experiments were performed to refine the model structure and to define the majority of the model parameters. The model has been evaluated by measuring the control behavior in tracking tasks.

A pilot's perception of variables presented on the Electronic Flight Instrument System, EFIS, was investigated. A stimulus response technique was used to determine the accuracy and speed of the perception process. By varying the exposure time of the stimuli, it is shown that the perception of a variable's magnitude is faster and more accurate than the perception of the first derivative or rate of that variable. Results of experiments on roll and pitch attitude perception, the influence of scale division, and the perception of the indicated airspeed, are shown.

The availability of pressure information of a hydraulic actuator makes it possible to improve the quality of vehicle power transmission via precise feedback control and to realize on-board fault diagnosis. However, the high cost of a pressure sensor has not allowed its widespread deployment despite such apparent advantages. This paper presents an observer-based algorithm to estimate the pressure output of a hydraulic actuator in a vehicle power transmission control system. The proposed algorithm builds on more readily available slip velocity and the models of a hydraulic actuator and a mechanical subsystem. The former is obtained empirically via system identification, while the latter is derived physically. The resulting robust observer is guaranteed to be stable against possible parametric variations and torque estimation errors. The hardware-in-the-loop studies demonstrate the viability of the proposed algorithm in the field of advanced vehicle power transmission control and fault diagnosis

Magnetic levitation systems have recently become the focus of many research interests not only because they are most suitable for high precision engineering applications but also due to the fact that they represent a difficult challenge to control engineers. As a result, most previous studies have focused on the control stabilization problem. In this paper, we address the issue of performance with respect to uncertainty in order to achieve a desired rigidity. The proposed controller is an adaptive backstepping controller. The adaptive backstepping controller provides system stability under model uncertainty, and achieves the desired servo performance. The experiments show that the proposed control achieves a superior behavior than other control

Time-optimal control strategies of the on-off type can be used to bring batch reactor temperature to the set point in the minimum time. A practical controller implementing this control strategy in industry is the dual-mode controller. When well tuned, this controller shows excellent system performance for various batch reactors. However, because time-optimal control is a kind of open loop control strategy, the dual-mode controller may be sensitive to process variations. For robust control, the dual-mode controller is modified here with an iterative learning technique. This iterative learning dual-mode controller requires minimal information from the previous batch runs and can be incorporated in the existing dual-mode controller with minimal effort.

We propose two advanced control algorithms, the LMS adaptive filter and fuzzy CMAC neural network, to counteract the chatter problem for a lathe machine. Experimental results are also included. Approximately 20 dB reduction in chatter has been achieved. We have also developed a multi-DSP board which can be used to implement any type of intelligent controllers to machine systems. Other potential applications of the proposed methods are to milling and boring machines

A data-driven model-free control design method has been proposed in Hjalmarsson et al. (1994). It is based on the minimization of a control criterion with respect to the controller parameters using an iterative gradient technique. In this paper, we extend this method to the case where both the plant and the controller can be nonlinear. It is shown that an estimate of the gradient can be constructed using only signal based information. It is also shown that by using open loop identification techniques, one can obtain a good approximation of the gradient of the control criterion while performing fewer experiments on the actual system

Recently the Time Delay Control (TDC) method has been proposed as a promising technique in the robust control area, where the plants have nonlinear dynamics with parameter variations and substantial disturbances are present. TDC method, however, requires the measurements of all the state variables, together with their derivatives. This requirement imposes a severe limitation on the applications to most real systems. In order to solve this measurement problem, we proposed an observer design method that can stably reconstruct the state variables and their derivatives. Then, for a simulation study, the controller/observer based on our design method has been applied to a nonlinear plant, the result of which confirmed that the controller/observer performs satisfactorily as predicted. Finally we made experimentations on a DC servo motor that is subject to substantial amount of inertia variations and external disturbances. The results showed that the controller/observer performs quite robustly under those variations and disturbances, and is much less sensitive to sensor noise than the controller using numerical differentiations.

This paper describes how to control strip gauge, looper angle, and strip tension for a hot strip finishing mill and the application result. This is based on the optimal servo theory and the model decoupling method. In the finishing mill process, there exists mutual interaction among strip gauge, looper angle, and strip tension. Conventionally, each of them is controlled independently. To improve the control performance, multivariable control considering this interaction is applied to all production

Hybrid vehicles use two energy sources for their propelling. Usually an internal combustion engine (ICE) is used with one or more electric machine(s) (EM). The problem is then to split the driver power demand between the ICE and the EM in order to minimize a criterion, usually the fuel consumption. A global optimization algorithm based on optimal control theory is recalled. The obtained results are optimal but can only be obtained in simulation. For real time control purpose, this optimization algorithm is applied on a receding horizon. The main problem is then to choose the variables to be predicted on this horizon. By analyzing the optimization algorithm, it is shown that the prediction of the future driving conditions (vehicle speed and driver torque demand) is not necessary. Therefore, under some assumptions, a real time control is possible.

Pressure screens are primarily used for contaminant removal. More recently, they have been used for fibre fractionation to produce pulp that is more uniform and of higher quality. Careful control is required for the safe operation of pressure screens, and for providing the required flows and level of cleanliness without failure. This is particularly important in fibre fractionation where optimal separation occurs when screens are operated as close as possible to the failure limit. Current control strategies are based on linear univariate control techniques. However, because the process is nonlinear and strongly coupled, univariate linear controllers are likely to perform very poorly when tight control is required. This kind of control requires an accurate model of the system. We describe a nonlinear dynamic model of pressure screens. Different methods were used to estimate the unknown parameters of the model, which was then validated against real data. Furthermore, we propose a simple nonlinear multivariate control strategy based on a static nonlinear model. For performance evaluation, we compare a conventional control strategy to the proposed strategy in simulation, and in pilot plant experiments

It is widely and frequently observed that industrial robots conducting fast motion involve serious residual vibration, the period of which varies with time. To this problem, this paper presents a practical solution by providing a practical design and application of time-varying input shaping technique (TVIST) for an industrial robot. To suppress the time-varying vibration, at first, a guideline for designing practical TVIST is presented. Following the guideline, then, we design TVIST for a large size 6 degrees of freedom industrial robot. In doing so, a simple yet effective equation is derived from robot dynamic equations to estimate the time-varying period. Furthermore, a simple payload adaptation scheme is also included. The TVIST thus designed is experimented on the industrial robot under spatial motion and payload variation conditions. The experimental results show that the residual vibration is reduced to less than 10% of original one in magnitude, demonstrating that the efficiency of the TVIST does not compromise its effectiveness

The application of gain-scheduled control to a pilot-scale solar power plant is described. A field of parabolic collectors focus the solar radiation onto a tube where oil is pumped through in order to collect the solar power. The control problem is to keep the temperature of the oil leaving the field at its desired value by manipulating the oil pump flow rate. It is shown that gain-scheduling can effectively handle the plant nonlinearities, using high-order local linear ARX models that form the basis for the design of local linear controllers using pole placement.

An optimal-tuning nonlinear PID controller design strategy is proposed for hydraulic systems. After an analysis of these systems, an analytic physical dynamical model with dead-zone nonlinearity is derived. A nonlinear PID control scheme with the inverse of the dead zone is introduced to overcome the dead zone in the hydraulic systems. An optimal PID controller is designed to satisfy some desired time-domain performance requirements. Using an estimated process model, the optimal-tuning PID control provides optimal PID parameters even when the process dynamics are time variant. This strategy is implemented in an environment composed of dSPACE, MATLAB, SIMULINK and Real-Time Workshop. The performance of the controller is demonstrated on a hydraulic position control test rig.

The paper demonstrates that a self-learning neurofuzzy controller is able to regulate the temperature in a liquid helium cryostat. In order to simplify the task of commissioning the controller, a strategy for choosing the user-selected parameters from an equivalent proportional-plus-integral controller (PI) is derived. Experimental results which illustrate the potential of the proposed control scheme are presented. The performance of the self-learning neurofuzzy controller is also compared with that of a commercial gain-scheduled PI controller.

This paper deals with an efficient application of a model-based predictive control scheme in parallel mechanisms. A predictive functional control strategy based on a simplified dynamic model is implemented. Experimental results are shown for the H4 robot, a fully parallel structure providing 3 degrees of freedom (dof) in translation and 1 dof in rotation. Predictive functional control, computed torque control and PID control strategies are compared in complex machining tasks trajectories. Tracking performance and disturbance rejection are enlightened.

Although conventional PID-like SISO controllers are still most common in industry, there is a growing need for more advanced controller structures in order to comply with ever tighter performance requirements. In this paper we consider positioning devices in IC-manufacturing for which position-dependent plant dynamics are a performance limiting factor. We suggested to employ recently developed linear parameter varying (LPV) control techniques for designing position-dependent controllers that adapt themselves in order to achieve optimal closed-loop performance. Our main emphasis is on presenting a practical LPV design procedure which covers plant modeling, controller synthesis and actual implementation for an electromechanical positioning device, an advanced wafer-scanner. Our experimental results reveal that performance can be improved by LPV control if compared to a classical SISO design. We highlight a variety of troublesome aspects within the design cycle that lack a systematic theoretically founded solution and that limit the possible performance improvement achievable by LPV control.

This paper deals with fault detection and isolation (FDI) of smart actuators combining the benefits of bond graph modelling with external models. An external model is a generic method which can be used to specify and verify the functional specifications of smart equipment. It uses the concept of services provided to the user and the organization of operating modes as a state graph with the logical conditions (depending on available services) required to move from one operating mode to another. One drawback of the external model is that it describes the system in terms of functions without taking into account the dynamics of the equipment. In addition, the availability of the services is not determined by FDI procedures. The integration of the bond graph methodology as a graphical and multi-disciplinary modelling tool quantifies the services by associating one or more bond graph elements to each service. Furthermore, the causal and structural properties of the bond graph methodology can be used to design FDI algorithms (i.e. the generation of fault indicators) to determine the availability of the services. This information is used to determine the transition conditions in the state graph. This technique has been applied to monitor a complex pneumatic servo-positioner.

In this paper, a Lyapunov-based control algorithm is developed for force tracking control of an electro-hydraulic actuator. The developed controller relies on an accurate model of the system. To compensate for the parametric uncertainties, a Lyapunov-based parameter adaptation is applied. The adaptation uses a variable structure approach to account for asymmetries present in the system. The coupled control law and the adaptation scheme are applied to an experimental valve-controlled cylinder. Friction modeling and compensation are also discussed. The experimental results show that the nonlinear control algorithm, together with the adaptation scheme, gives a good performance for the specified tracking task. The original adaptive control law is then simplified in several stages with an examination of the output tracking at each stage of simplification. It is shown that the original algorithm can be significantly simplified without too significant a loss of performance. The simplest algorithm corresponds to an adaptive velocity feedback term coupled with a simple force error feedback.

A comparison between receding horizon control (RHC) approaches is presented for the longitudinal axis control of an F-16 aircraft. The results suggest that the flexibility provided by a scheduled RHC scheme based on flight condition-dependent linear prediction models is a necessary requirement for achieving good performance as opposed to a single LTI model-based method. The scheduled scheme offers an attractive alternative to a full nonlinear model-based RHC approach by trading off an acceptable degradation in performance to modest computational complexity and real-time implementability.

A backstepping approach is used in this paper to design a nonlinear controller for force control of a single-rod electrohydraulic actuator. The control design guarantees the convergence of the tracking error. The implementation of the control design requires system states for feedback, but in this case only the force output is available. To overcome this problem, a PI observer is used to estimate the states of the system. Experimental results have illustrated the success of the observer-based backstepping controller. The results are also compared to those obtained with conventional P and PI controllers. It can be shown that the observer-based backstepping controller has a relatively better tracking performance.

This paper describes the flight testing of a suite of controllers designed using linearisations extracted from a new nonlinear model of the Bell 205 helicopter. Details of the controller designs are presented, together with a comprehensive analysis of simulation and flight-test results. These are complemented by an evaluation of qualitative and quantitative handling qualities information against the design standard ADS-33. The most notable achievements of the work were: (1) stability was achieved for all controllers tested and, moreover, some of them yielded desirable handling qualities from the first test onwards, and (2) a high degree of consistency between desk-top simulation and flight-test results was observed. Achieved performance was generally satisfactory and future planned flight tests will build on these results with the aim to increase performance further. The paper discusses various aspects of controller design and presents an analysis of the results obtained. The likely direction of further work is also discussed.

This paper describes the development and validation of a nine-degree-of-freedom (DOF) helicopter model, and the use of this model in designing H∞ optimal controllers for testing on an experimental fly-by-wire helicopter. The model includes rotor flapping dynamics and an inflow correction factor. Its use has led to improvements (with respect to earlier designs based on a different model) in the robustness properties of the resulting controller. For example, the need for predictor-type filters in the feedback path has been eliminated. Flight-test results are presented. Discrepancies still remain between predicted and achieved performance, but overall handling quality ratings have improved with respect to a previous H∞ design, and clear improvements have been obtained in pitch and roll attitude control: e.g. an increase in roll axis bandwidth from 1.9 to 2.5 rad/s.

This paper presents a Year-2000 (Y2K) status report of mechatronics. The Y2K definition of mechatronics is “the synergetic integration of physical systems with information technology and complex-decision making in the design, manufacture and operation of industrial products and processes.” Mechatronics may be interpreted as the best practice for synthesis of engineering systems, and it covers a broad area and scope. Vehicle lateral control for automated highway systems, hard disk drives and media handling mechanisms for printing engines are reviewed as examples of mechatronics research. Engineering students should be exposed to mechatronics and to the culture of working in teams.

The current vehicle-highway system has reached a plateau in its ability to meet the demand for moving goods and people. This paper sketches an architecture for an automated highway system or AHS. The architecture can be realized by several designs that differ in terms of performance and sophistication. One design is described that could triple capacity and reduce travel time, guarantee collision-free operation in the absence of malfunctions, limit performance degradation in the case of faults, and reduce emissions by half. Evidence suggesting that the design can be implemented is summarized. It is indicated how the design can be adapted to different urban and rural scenarios and how a standard land-use model can show the impact of AHS on urban density. A summary of the progress of the National Automated Highway Systems Consortium is provided. The paper concludes with a critique of AHS.