Showing papers in "IEEE Control Systems Magazine in 1996"

Adaptive Control

Abstract: This book, written by two major figures in adaptive control, provides a wealth of material for researchers, practitioners, and students. While some researchers in adaptive control may note the absence of a particular topic, the book‘s scope represents a high-gain instrument. It can be used by designers of control systems to enhance their work through the information on many new theoretical developments, and can be used by mathematical control theory specialists to adapt their research to practical needs. The book is strongly recommended to anyone interested in adaptive control.

Anti-windup, bumpless, and conditioned transfer techniques for PID controllers

Abstract: Gives a simple and comprehensive review of anti-windup, bumpless and conditioned transfer techniques in the framework of the PID controller. We show that the most suitable anti-windup strategy for usual applications is the conditioning technique, using the notion of the realizable reference. The exception is the case in which the input limitations are too restrictive. In this case, we propose the anti-windup method with a free parameter tuned to obtain a compromise between the incremental algorithm and the conditioning technique. We also introduce the new notion of conditioned transfer, and we it to be a more suitable solution than bumpless transfer. All the discussions are supported by simulations.

Adaptive Control [Second edition, by Karl J. Astrom and Bjorn Wittenmark, Addison Wesley (1995)]

Abstract: This book, written by two major figures in adaptive control, provides a wealth of material for researchers, practitioners, and students. While some researchers in adaptive control may note the absence of a particular topic, the book‘s scope represents a high-gain instrument. It can be used by designers of control systems to enhance their work through the information on many new theoretical developments, and can be used by mathematical control theory specialists to adapt their research to practical needs. The book is strongly recommended to anyone interested in adaptive control.

Optimal control-1950 to 1985

Abstract: Optimal control had its origins in the calculus of variations in the 17th century. The calculus of variations was developed further in the 18th century by Euler and Lagrange and in the 19th century by Legendre, Jacobi, Hamilton, and Weierstrass. In the early 20th century, Bolza and Bliss put the final touches of rigor on the subject. In 1957, Bellman gave a new view of Hamilton-Jacobi theory which he called dynamic programming, essentially a nonlinear feedback control scheme. McShane (1939) and Pontryagin (1962) extended the calculus of variations to handle control variable inequality constraints, the latter enunciating his elegant maximum principle. The truly enabling element for use of optimal control theory was the digital computer, which became available commercially in the 1950s. In the 1980s research began, and continues today, on making optimal feedback logic more robust to variations in the plant and disturbance models; one element of this research is worst-case and H-infinity control, which developed out of differential game theory.

A brief history of automatic control

Abstract: The author looks at the history of automatic control in four main periods as follows: early control up to 1900; the pre-classical period 1900-1940; the classical period 1935-1960; and modem control post 1955. This article is concerned with the first three of the above; other articles in this issue deal with the more recent period.

Experimental study of biped dynamic walking

Abstract: An experimental study of a biped robot is presented. A new scheme named the "linear inverted pendulum mode" is utilized for controlling a biped walking on rugged terrain. The authors developed a six-DOF biped robot, "Meltran II," which has lightweight legs and moves in a two-dimensional vertical plane. To investigate the effects not well discussed in the theory, the authors carried out two experiments, the support phase experiment and the support exchange experiment. The support phase experiment was carried out to check the actual dynamics of a biped walking under the proposed control. It was shown that the dynamics of the robot can be regarded as linear even though the mass of the legs, which was neglected in the theory, exists. The support exchange experiment was performed to check leg support exchange. The authors found that a smooth leg support exchange is achieved by making the foot contact with a certain vertical speed and holding two-leg support for a certain short period. Based on these results, a whole biped control system was implemented. In the authors' experiment the robot walked over a box of 3.5 cm height at a speed of 20 cm/s.

Force regulation and contact transition control

Abstract: In this article, a new sensor-referenced control method using positive acceleration feedback together with a switching control strategy is developed for robot impact control and force regulation. The robot dynamic model is feedback-linearized and decoupled for the free-motion mode, the phase-transition mode, and the constrained-motion mode. Considering the detection of the impact as an event, the event-driven switching control strategy is used to deal with the inadvertent loss of contact of the robotic manipulator. Bouncing can be eliminated after finite switches. A stable transition can be achieved with a nonzero impact velocity; large impact forces can be avoided and the output force can be regulated after contact is established. Stability analysis based on the Lyapunov-like method is given for the proposed system. The scheme was implemented and tested on a 6 DOF PUMA 560 robot arm.

Vehicle following control design for automated highway systems

Abstract: Automatic vehicle following is an important feature of a fully or partially automated highway system (AHS). The on-board vehicle control system should be able to accept and process inputs from the driver, the infrastructure, and other vehicles, perform diagnostics, and provide the appropriate commands to actuators so that the resulting motion of the vehicle is safe and compatible with the AHS objectives. The purpose of this article is to design and test a vehicle control system in order to achieve full vehicle automation in the longitudinal direction for several modes of operation, where the infrastructure manages the vehicle following. These modes include autonomous vehicles, cooperative vehicle following, and platooning. The vehicle control system consists of a supervisory controller that processes the inputs from the driver, the infrastructure, other vehicles, and the on-board sensors and sends the appropriate commands to the brake and throttle controllers. In addition, the controller makes decisions about normal, emergency, and transition operations. Simulation results of some of the basic vehicle following maneuvers are used to verify the claimed performance of the designed controllers. Experiments on Interstate-15 that demonstrate the performance of the throttle controller with and without vehicle-to-vehicle communications in an actual highway environment are also included.

Hybrid systems in process control

Abstract: Modeling and control of hybrid systems, with particular emphasis on process control applications, are considered in this article. Based on a number of observations on typical mixed discrete and continuous features for such applications, a fairly general model structure for hybrid systems is proposed. This model structure, which clearly separates the open-loop plant from the closed-loop system, is suitable for analysis and synthesis of hybrid control systems. To illustrate this, three different approaches for control-law synthesis based on continuous and discrete specifications are discussed. In the first one, the hybrid plant model is replaced by a purely discrete event model, related to the continuous specification, and a supervisor is synthesized applying supervisory control theory suggested by Ramadge and Wonham (1987). The other two methods directly utilize the continuous specification for determination of a control event generator, where time-optimal aspects are introduced as an option in the last approach.

CAPC: A Road-Departure Prevention System

Abstract: he Crewman’s Associate for Path Control (CAPC) is an automatic road-departure warning system for motor vehicle drivers. This article presents the design of, and preliminary field results from, the CAPC prototype vehicle, a modified 1995 Ford Taurus SHO. The CAPC system is intended for highway drivers who are drifting off the roadway at a shallow angle of departure due to inattention, drowsiness, intoxication, or other causes. A camera senses the roadway ahead, and a suite of transducers provide measurements of vehicle motion and driver steering commands. These are used to anticipate unintended road departures and wam the driver. Warning and intervention functions for CAPC have been developed during ongoing research at the University of Michigan; the algorithms are reviewed here in the context of the prototype. The prototype hardware and software are described in detail, and preliminary experimental data from the field is presented.

Traffic flow modeling and control using artificial neural networks

Abstract: In this article we use an artificial neural network technique to model and control highway traffic in a single lane with no on- or off-ramps. The developed controllers generate the speed commands for each section of the lane that vehicles need to follow in order to achieve a desired traffic flow density distribution along the lane. In today's traffic, these speed commands could be communicated to drivers, who would then have to respond to them. This raise human factors issues that need further investigation in order to assess the possible benefits. In an automated highway, speed commands cn be communicated to the vehicle's computer control system and followed directly without human errors or delays. We simulate an automated highway environment with such a system to alleviate congestion. We demonstrate that the use of feedback control on the macroscopic level could bring dramatic improvements to traffic flow characteristics.

Filtering and stochastic control: a historical perspective

Abstract: We attempt to give a historical account of the main ideas leading to the development of nonlinear filtering and stochastic control as we know it today. We present a development of linear filtering theory, beginning with Wiener-Kolmogoroff filtering and ending with Kalman filtering. The linear-quadratic-Gaussian problem of stochastic control is considered and states that for this problem the optimal stochastic control can be constructed by solving separately a state estimation problem and a deterministic optimal control problem. Many of the ideas presented here generalize to the nonlinear situation. A reasonably detailed discussion of nonlinear filtering, again from the innovations viewpoint, is given. Finally, we deal with optimal stochastic control. The general method of discussing these problems is dynamic programming.

Air traffic management: evolution with technology

Abstract: One of the remarkable features of ATM systems from a control perspective is that there is no tolerance for system shutdown and start-up. The system literally must operate 24 hours per day, seven days per week. It is not feasible to simply shut down an old system and start up a new one. The economic and social costs of declaring that one day will be a day with no air traffic while the systems are switched are simply prohibitive. ATM, at least as far into the future as we can imagine, is a system with essential human and machine components. It is inconceivable that either component will ever disappear. Yet, as the system evolves, there are continual changes in the way work is shared between human and machine. Provision for such flexibility is an essential requirement for any proposed new system. New technologies are driving the ATM paradigm from command and control in the direction of more distributed decision making. This may be best illustrated by TCAS (the Traffic Alert and Collision Avoidance System). TCAS is a tactical system intended for use by pilots to avoid collisions caused by inadvertent violations of separation standards. It displays to the flight crew specific information about certain nearby aircraft and, depending on the likelihood of conflict, provides advice for actions needed to avoid collisions. Details of TCAS are given in the article.

Vibration control of a micro/macro-manipulator system

Abstract: Inertial force damping control by micro-manipulator modulation is proposed to suppress the vibrations of a micro/macro manipulator system. The damping controller, developed using classical control theory, is added to the existing control system. Real-time measurements of macro-manipulator flexibility are used to adjust the motion of the micro-manipulator to counteract structural vibrations. Experimental studies using an existing micro/macro flexible-link manipulator testbed demonstrate the effectiveness of the proposed control scheme for both vertical and horizontal plane vibration.

Identification and control of systems with friction using accelerated evolutionary programming

Abstract: This article proposes a novel evolutionary algorithm, called accelerated evolutionary programming (AEP), which improves evolutionary programming in terms of convergence speed and diversity. Comparison between the proposed algorithm and evolutionary programming is carried out for five widely used test functions to show the effectiveness of the proposed algorithm. The proposed algorithm is applied to the identification of a seven-parameter friction model of an X-Y table, which is adopted from the results of tribology studies. Based on the identified friction model, a compensator is designed for the control of the X-Y table without stick-slip motion at very low velocity. Experimental results on the X-Y table demonstrate the effectiveness of the proposed scheme, especially for very-low-speed tracking.

Application of discrete-event-system theory to flexible manufacturing

Abstract: The anticipated extensive use of flexible-manufacturing workcells in the future has encouraged recent research efforts on the development of automatic supervisory-control methodologies. However, despite intensive research on the theoretical aspects of the control of manufacturing workcells, modeled as discrete-event systems (DES), a limited amount of research has been reported on the implementation of DES-based supervisory controllers. In this article, such a generalized implementation methodology, that utilizes theoretical advances in conjunction with programmable-logic-controller (PLC) technology, is presented. The two primary advantages of the proposed methodology are: (i) the utilization of limited-size control strategies that can be efficiently generated on-line, and which are conflict- and deadlock-free by construction (via controlled-automata DES theory); and (ii) the use of PLCs, which are currently the most suitable and widely employed industrial process-control technology. In the authors' proposed methodology, a host personal computer (PC) possesses an on-line capability for the automatic generation of supervisory-control strategies, and their downloading to a PLC as required. The PLC, in turn, is responsible for monitoring the workcell, reacting to events and enforcing device behavior based on the current control strategy residing in its processor. A supervisory controller developed based on this approach was successfully implemented for a manufacturing workcell in the authors' laboratory.

Improving the performance of stabilizing controls for nonlinear systems

Abstract: There are a variety of tools for computing stabilizing feedback control laws for nonlinear systems. The difficulty is that these tools usually do not take into account the performance of the control, and therefore systematic improvement of an arbitrary stabilizing control law is extremely difficult and often impossible. The objective of this article is to present a design algorithm that addresses this problem. The algorithm that we present iteratively computes a sequence of control laws with increasingly improved performance. We also consider implementation issues and discuss some of the successes and difficulties that we have encountered. Finally, we present a number of illustrative examples and compare our algorithm with perturbation methods.

Input-output feedback stability and robustness, 1959-85

Abstract: The literature on input-output feedback is too large to do justice into a single article. Here we concentrate on the formative years ending in 1985 and leave the subsequent story to be told elsewhere. It is not a comprehensive survey of the literature or even of the most important papers. Rather, it is an attempt to describe events that marked the turning points. The period under scrutiny can be divided into roughly two parts; interest in nonlinear stability dominated the first part and robustness the second.

Automotive applications of a hybrid active noise and vibration control

Abstract: Fuel economy, international competition, and socioeconomic factors have forced manufacturers to design smaller and lighter automotive vehicles. Such vehicles are typically more susceptible to noise and vibrations, which increase as they age. For reasons of health, there has been a major increase of effort in the design of noise and vibration characteristics of vehicles that would better appeal to the occupants. This article presents some results from our past 12 years of research and development in the area of active/adaptive noise and vibration control in transportation systems, with specific results from activities related to automotive vehicles.

Experiments in adaptive model-based force control

Abstract: Reports experiments with a class of model-based adaptive force control algorithms for robot arms. The problem addressed in this article is the control of robots whose motion is constrained by point contact between the robot tool and a smooth rigid environment or workpiece. Manufacturing applications for force control include a great variety of commonplace tasks, such as grinding, polishing, buffing, deburring, and assembly operations currently performed either manually or by fixed automation equipment. The force control algorithm provides asymptotically exact tracking of both end-effector position and contact-force. This force control algorithm utilizes a sliding-mode control technique of a type first espoused for the case of free (non-contact) robot motion. The stability of the new force control algorithm can be proven with respect to the commonly accepted nonlinear rigid body dynamical equations of motion. Moreover its adaptive extension can be shown to adaptively compensate for unknown plant parameters such as link and payload inertia, joint friction, and friction arising at the contact point between the tool tip and the surface. In Naniwa et al. (1993) and Arimoto and Naniwa (1992) the authors report satisfactory performance of this force control algorithm in numerical simulation studies. This article demonstrates the comparative advantages and disadvantages of this control algorithm under a variety of conditions in actual working implementations.

The evolution of systems analysis and control: a personal perspective

Abstract: The foundations of systems analysis and control as we know them today were laid for the most part at MIT's Radiation Laboratory during World War II and a period thereafter. Most of the founders-both in the United States and abroad-are no longer with us. As one who had the privilege of knowing Wiener, Bode, Nyquist, Guillemin, Gordon Brown, Sam Mason, John Coales, Aizerman, Pontryagin, Letov, Bellman, and many others, I present some personal perceptions and reminiscences in this paper. However, in view of the vastness of the subject, I touch upon only a small subset of the issues and events that were at the center of attention.

Adaptive control around 1960

Abstract: This paper presents some of the developments of adaptive control during a 10-year period around 1960. This was a very fertile period when many ideas were developed that later proved useful. The motivation came from several sources-advances in control theory, demanding applications in flight control and process control, the desire to develop systems with learning capabilities, and problems related to decision-making under uncertainty. This article deals almost exclusively with development in the Western world.

RBF network feedforward compensation of load disturbance in idle speed control

Abstract: Automotive engine idle speed control is a disturbance rejection problem. An engine at idle is typically well away from its most favorable region of operation and exhibits significant nonlinearities. Control of such a system is complicated by delays of both physical (time between induction and power strokes) and computational origin. In the model used, there are two control variables, and these differ in both their range of effectiveness and their temporal characteristics: spark advance is fast-acting but limited in its effect, while throttle has a large range but a slower effect which results both from the dynamics of filling the intake manifold and from the induction-power delay. The spark variable also has a maximum effective value, i.e., a value beyond which it has an effect opposite to that expected. This article describes a nonlinear adaptive feedforward controller for compensation of external load disturbances in the idle speed control of an automotive engine. The controller is based on a radial basis function (RBF) network approximation of certain input-output mappings describing the system. An underlying assumption used in the controller design is that the external engine load is known to the controller. In particular, that might be achieved by putting an appropriate torque sensor in the powertrain or using other available information.

Dual control for an unstable mechanical plant

Abstract: The problem of applying direct dual control for positioning of a laboratory pilot-plant for the roll angle control of a vertical takeoff airplane by means of two propellers is described in the article. The bicriterial approach is applied for the synthesis of dual controllers, where two criteria are used which correspond to the two goals of dual control: control of the system output and exciting the system for speeding up the parameter estimation. The control law is derived after compromised optimization of two cost functions. The basic theory needed for synthesis of a direct dual control for linear single-input, single-output plants is discussed. After describing the experimental setup, the results of adaptive control of a mechanical laboratory pilot-plant are presented. It is pointed out that adaptive control is necessary and is the most suitable approach for this kind of plant. The new direct dual control algorithm is compared with indirect and direct adaptive control strategies based on the standard CE assumption.

Linear and neural network feedback for flight control decoupling

Abstract: Some experts are of the opinion that the task of flight training can become far less labor-intensive if the pilot can directly control each of the state variables of the aircraft individually. Yet complete decoupling of the aircraft as a nonlinear system is a formidable problem. Such a task requires accurate aircraft state information and rapid computing. The difficulties are compounded when the dynamics or the aerodynamics of the aircraft fall in the highly nonlinear regimes. The authors demonstrate the potential for an artificial neural network in conjunction with a linear compensator to perform such a function. The authors show that the linear compensator is unable to control the aircraft in the absence of the neural network. A neural network can be trained to produce the large nonlinear portion of the control inputs; however, a hybrid combination of the neural network and the compensator based on the linearized equations of motion gives the best results. Furthermore, The authors demonstrate that such a hybrid system can tolerate a large amount of noise in the network input. Several examples are shown, with and without the linear compensator. Finally, the authors demonstrate generalization within the training domain through accurately predicting a case that was absent in the training domain.

μ-synthesis control for a walking robot

Abstract: Proposes using two under-actuated robotic systems as two legs of a walking machine. Since the ankle joints of both legs have no actuation, the walking machine remains upright only when it is balanced via a computer-controller. The design of a stabilizing controller for one under-actuated leg, using a μ-synthesis approach, is explained in detail. The system at the equilibrium can be represented by a non-minimum phase system. The right half plane zero severely limits the achievable closed-loop bandwidth and robustness in modeling errors.

METAMODELLING: Bond Graphs and Dynamic Systems [BOOKSHELF]

Abstract: With the increasing complexity of processes to be analyzed, the control engineer often needs a model of the system to be controlled. However, in many practical cases, there is limited time for detailed system analysis, and the engineer may not be an expert in that particular system domain. The authors use the term “metamodelling” to describe the approach taken in this book, Le., a modeling methodology that transcends the accepted mathematical models for specific applications. This methodology abstracts general models from first principles by employing an existing notation (bond graphs) as a metalanguage for describing physical systems. This book is, therefore, concerned with separating out the model development process from the functions for which the model is developed, in order to enhance understanding of the essentials of the real physical systems. This book is organized in two parts. The first part describes modeling principles, based on system decomposition, first using classical dynamical analysis and then via the energy bond graph notation. Bond graphs are shown to provide a powerful core model representation from which a variety of mathematical models may be derived. Bond graphs provide a useful means of illustrating causality, which is a crucial aspect of system modeling. Part I consists of six chapters. In Chapter 1, the authors discuss why models are needed and, using an example industrial process, develop a requirement specification for a modeling tool. Chapter 2 deals with representation of elementary systems. In this chapter, the decomposition of a system into a structure linking elements representing its static and dynamic behavior is reviewed, first via classical dynamical analysis and then via the energy bond graph notation. Chapter 3 addresses the topic of causality. Although causality may appear to be an abstract notion, this chapter argues that causality is a crucial aspect of system modeling. Links to related areas such as constraint programming and qualitative modeling are drawn. The use of computers to aid modeling is a central theme of this book. Chapter 4 discusses the twin issues of representation and transformation. In particular, model transformations from the core (bond graph) representation to various derived mathematical models ( such as differential-algebraic equation, non-linear statespace, linearized state-space, frequency response, etc.) are given and illustrated. The art of modeling is, to a large extent, the art of abstracting the simplest model for the required purpose. Chapter 5 shows how a bond-graph methodology for system approximation can aid the system modeler. For the purpose of design, it is also sometimes useful to find the system input (as a function of time), which will produce a given system output. This leads to the idea of the inverse of a dynamic system. Chapter 6 shows how to obtain the bond-graph of an inverse system from the bond-graph of the system itself. The second part of the book uses specific case studies to illustrate the application of this methodology to systematic generation of the most widely used mathematical models. Part I1 consists of five chapters. In Chapter 7, the process of insulating copper wire using plasticating extruder is described. A systematic approach to modeling process systems is developed and illustrated in Chapter 8. In Chapter 9, models for inhaled drug uptake, with particular relevance to anesthesia, are derived based on physical principles encapsulated in bond graphs. Chapter 9 deals with mechanical systems and robotics. Bond graphs are used to model the dynamics of a two-dimensional mechanical link. This basic building block is used to systematically create dynamic models for a number of simple systems including a pendulum, a double pendulum, and a two-link manipulator. This process is repeated for three-dimensional systems, resulting in models for robotics manipulators, including the PUMA and Stanford arm architecture. Much of the current body of control achieves a generic coverage of application areas by having a generic representation of the systems to be controlled that are not well suited to partially known systems. In Chapter 1 1, the authors propose an alternative approach that, while achieving a generic coverage of systems, allows the use of particular representation (possibly partially known, possibly non-linear) of systems.

From conventional control to autonomous intelligent methods

Abstract: In this article we review the growth and development of control engineering, leading to modern adaptive methods and finally to autonomous intelligent control. Although the use of feedback control can be traced back to ancient and medieval times, it is really during the 20th century that control engineering has become a recognized discipline. Well-established methods to model and control plants with linear characteristics and unchanging parameters are already in existence. Nonlinear plants with time-varying internal parameters are more challenging and the so-called "adaptive" methods have been developed to address this issue. There are controllers that can "learn" by using AI techniques such as expert systems, genetic algorithms, neural networks, etc. These paradigms have evolved mostly from studying biological learning processes.

Early developments in nonlinear control

Abstract: This article presents a review of some of the early developments in nonlinear control engineering. It begins by briefly examining the status prior to World War II and then concentrates on the developments in approximately the following two decades. A significant amount of the work will be seen to have as its foundation the need to obtain methods to solve wartime problems related to servomechanisms, fire control, and missile control. The major analytical techniques of the phase plane, the describing function and Tsypkin's method for relay systems, are discussed at length. Two important aspects in obtaining control engineering solutions to problems are the existence of simulation facilities and the capabilities of the available hardware to implement specific solutions at an allowable cost. A brief review of some of the early papers on nonlinear control is also given, and finally some comments are provided on the continuing relevance of these early methods and some still unsolved problems.

Stability: the common thread in the evolution of feedback control

Abstract: Advances in feedback control theory have inextricably been linked to progress in stability theory, and conversely, the horizons of stability theory have greatly been expanded by the challenges posed in increasingly complex feedback control systems. In the present article we trace the development of stability theory in feedback control by making contact with some of the major milestones in this area.