Showing papers in "IEEE Control Systems Magazine in 1988"

Survey of model-based failure detection and isolation in complex plants

Abstract: Techniques to detect and isolate failures in complex technological systems, such as sensor biases, actuator malfunctions, leaks, and equipment deterioration are surveyed. The methods are based on analytical redundancy afforded by a mathematical model of the system. The main components of such techniques are residual generation using the model, signature generation by statistical testing, and signature analysis. Model-structural conditions for failure isolation are introduced together with transformation methods to implement them. Sensitivity and robustness considerations are presented, and a design framework based on model redundancy is proposed. >

A multilayered neural network controller

Abstract: A multilayered neural network processor is used to control a given plant Several learning architectures are proposed for training the neural controller to provide the appropriate inputs to the plant so that a desired response is obtained A modified error-back propagation algorithm, based on propagation of the output error through the plant, is introduced The properties of the proposed architectures are studied through a simulation example >

Hierarchical neural network model for voluntary movement with application to robotics

Abstract: In order to control voluntary movements, the central nervous system must solve the following three computational problems at different levels: determination of a desired trajectory in the visual coordinates; transformation of the trajectory from visual coordinates to body coordinates; and generation of motor commands. Based on physiological information and previous models, computational theories are proposed for the first two problems, and a hierarchical neural network model is introduced to deal with motor commands. The application of this approach to robotics is outlined. >

Application of generalized predictive control to industrial processes

Abstract: A novel algorithm called generalized predictive control (GPC) is shown to be particularly effective for the self-tuning control of industrial processes. The method uses long-range predictive control ideas with a carefully chosen controlled autoregressive and integrated moving average (CARMA) plant model and various horizons that allow for a rich variety of control objectives. The procedure can adapt to process dead time and model order, and a multivariable version gives tight control of complex plants without prior knowledge of the interactor matrix. Applications of GPC to a cement mill, a spray-drying tower, and a compliant robot arm give performance better than that of fully tuned proportional-integral-derivative regulators. >

Use of a rule-based system for process control

Abstract: A rule-based, digital, closed-loop controller that incorporates fuzzy logic has been designed and implemented for the control of power on the 5-MW Massachusetts Institute of Technology (MIT) Research Reactor under both steady-state and transient conditions. A comparison is made of the rule-based and analytic approaches. Differences in the division of labor between plant engineers and control specialists, the type of knowledge required and its acquisition, the use of performance criteria, and controller testing are discussed. The design, implementation, and calibration of rule-based controllers are reviewed, with specific examples taken from the completed work on the MIT Research Reactor. The possible role of rule-based technology in process control is evaluated. It is proposed that since rule-based systems are generally more robust than their analytic counterparts, they should be used both as backups to analytic controllers and as a means of improving the man-machine interface by providing human operators with the rationale for automatic control actions. >

Neural networks for routing communication traffic

Abstract: The use of neural network computational algorithms to determine optimal traffic routing for communication networks is introduced. The routing problem requires choosing multilink paths for node-to-node traffic to minimize loss, which is represented by expected delay or some other function of traffic. The minimization procedure is implemented using a modification of the neural network traveling-salesman algorithm. Illustrative simulation results on a minicomputer show reasonable convergence in 250 iterations for a 16-node network with up to four links from origin to destination. >

Introduction to neural networks for intelligent control

Abstract: Neural network architecture is presented as one approach to the design and implementation of intelligent control systems. Neural networks can be considered as massively parallel distributed processing systems with the potential for ever-improving performance through dynamical learning. The nomenclature and characteristics of neural networks are outlined. Two simple examples are presented to illustrate applications to control systems: one is fault isolation mapping, and the other involves optimization of a Hopfield network that defines a clockless analog-to-digital conversion. >

Modeling of an internal combustion engine for control analysis

Abstract: Recent activity in nonthermodynamic modeling of automotive internal combustion engines with spark ignition, which are inherently nonlinear, is reviewed A fundamental nonlinear model of the engine is presented, and a linear control-oriented model is derived from the nonlinear process Techniques for experimental verification are examined, and a practical linear engine example incorporating multirate sampling is illustrated >

Neural network architecture for control

Abstract: Two important computational features of neural networks are associative storage and retrieval of knowledge, and uniform rate of convergence of network dynamics independent of network dimension. It is indicated how these properties can be used for adaptive control through the use of neural network computation algorithms, and resulting computational advantages are outlined. The neuromorphic control approach is compared to model reference adaptive control on a specific example. It is shown that the utilization of neural networks for adaptive control offers definite speed advantages over traditional approaches for very-large-scale systems. >

Putting physics in control-the example of robotics

Abstract: It is argued that scalar summarizing properties, such as energy conservation or entropy production, often can be used effectively in the design of controllers for multi-input nonlinear physical systems. For instance, the conservation of total mechanical energy allows one to show simply the stability of simple proportional-derivative position controllers for robot manipulators, and it can also be systematically used to design trajectory controllers for these systems. Similarly, explicitly accounting for dissipative dynamics makes possible the use of the intrinsically stabilizing nature of certain disturbance terms such as viscous friction. Related approaches also can be of use in addressing problems of robustness to high-frequency unmodeled dynamics. The development points toward a more 'handcrafted,' physically motivated approach to nonlinear control system design. >

Object-oriented feature extraction method for image data compaction

Abstract: An online unsupervised feature-extraction method for high-dimensional remotely sensed image data compaction is proposed. This method is directed at the reduction of data redundancy in the scene representation of satellite-borne, high-resolution multispectral sensor data. The algorithm partitions the observation space into an exhaustive set of disjoint objects, and pixels belonging to each object are characterized by an object feature. The set of object features, rather than the pixel features, is used for data transmission and classification. Illustrative examples of high-dimensional image data compaction are presented, and the feature representation performance is investigated. Example results show an average compaction coefficient of more than 25 to 1 when this method is used; the classification performance is improved slightly by using object features rather than the original data, and the CPU time required for classification is reduced by a factor of more than 25 as well. The feature extraction CPU time is less than 15% of CPU time for original data classification. >

Automated robotic deburring using impedance control

Abstract: An automated deburring procedure using a robot manipulator is considered for the removal of burrs in the presence of robot oscillations and bounded uncertainties in the tool holder. Electronic compliancy (impedance control) is proposed as an adaptive mechanism to satisfy the requirements of this application. The development and implementation of the impedance control methodology on an active end-effector or the whole robot are examined for precision deburring and grinding tasks. >

Wood chip refiner control

Abstract: The authors treat the control of a chip refiner to produce wood pulp, where the incremental gain between the motor load and the plate gap is subject to sudden changes due to pad collapse. Passive adaptive control with fault detection, active suboptimal dual control, and dual-control solutions are discussed. For dual control, the choice of the control criterion is crucial, as it must reflect the peculiarities of the process. Two nonquadratic performance indices are proposed, and the myopic solutions are computed. Simulations show the superior performance offered by these indices. >

Design of tactile sensing systems for dextrous manipulators

Abstract: Preliminary work aimed at understanding the general issues and tradeoffs governing the design of extended tactile sensing systems is reviewed. General methods for estimating the bandwidths of line-addressed and matrix-addressed systems are presented. The proposed tactile sensing system incorporates four subsystems that permit the high-speed access of tactile data: (1) a transduction scheme; (2) a preprocessing scheme; (3) a multiplexing and transmission subsystem; and (4) tactile data selection techniques. Designs for implementation at each of these levels are presented. The designs emphasize practical necessities such as simplicity, reliability, and economy, along with plans to incorporate a tactile system into the Utah/MIT dextrous hand. >

Reducing residual vibration in systems with uncertain resonances

Abstract: Robots that perform rapid motions tend to excite system resonant frequencies. To perform a sequence of tasks more quickly, the settling time required for the vibration to decay should be minimal. Input functions are derived that produce rapid open-loop moves with greatly reduced residual vibration amplitude. To accommodate errors in the assumed system natural frequency, these forcing function are constructed so that the magnitude of their frequency spectra remains sufficiently small over a range of frequencies that bound the system natural frequency by +or-10%. These input functions are derived as a series expansion of ramped sinusoid functions with coefficients chosen to minimize spectral magnitude in this frequency band. Some simulations are performed to indicate that these functions can reduce residual vibration considerably even when the assumed natural frequency is in error by 10%. These inputs can then serve as the basis for a closed-loop implementation to generate reference trajectories that minimally excite system resonances. >

A view of automotive control systems

Abstract: Some current applications for engine control are described. Advanced control schemes for future nonlinear, time-varying automotive systems are discussed. Representative examples for plant modeling, parameter and state estimation, and adaptive control are presented. >

Evaluation of vision systems for teleoperated land vehicles

Abstract: A pilot study that compares vision systems that might be used to control a teleoperated land vehicle is described. The study compares three forward-looking vision systems, both for actual remote driving and for noninteractive video simulation. Remote driving has the advantage of realism but is subject to variability in driving strategies and can be hazardous to equipment. Video simulation provides a more controlled environment in which to compare vision-system parameters, but at the expense of some realism. Results demonstrate that relative differences in performance among the visual systems are generally consistent in the two test modes. A detection-range metric was found to be sensitive enough to demonstrate performance differences viewing large obstacles using black-and-white and color vision systems. Consequently, the author plans experimentation, aimed at optimizing vision-system parameters, that will rely to a greater extent on the more cost-effective, video-simulation approach. >

Vision-guided robotic fabric manipulation for apparel manufacturing

Abstract: Methods are described for vision-guided robotic control of fabric motion for performing simulated joining operations for apparel manufacturing. The determination of robot motion paths is based on visual information defining the position of the fabric edges in world coordinates. The usefulness of the shape analysis and motion control algorithms is demonstrated by experimentation with an integrated robot and vision fabric-handling system. Results of these tests show that using machine vision for planning robot motion provides an effective solution for implementing automated robotic fabric manipulation. >

Sensing and recognition of rigid objects using structured light

Abstract: Word directed toward the development of a vision system for bin picking of rigid 3D objects is reported. Any such system must have components for sensing, feature extraction, modeling, and matching. A structured light system which attempts to deliver a rich 2/sup 1///sub 2/D representation of the scene is described. Surface patches are evident as connected sets of stripes whose 3D coordinates are computed by means of triangulation and constraint propagation. Object edges are detected by the intersection of surface patches or by backprojecting image edges to intersect with the patches. Two matching paradigms are given for drawing correspondence between structures in the scene representation and structures in models. Three major contributions are reported: a method for sensing object surface patches without having to solve uniquely for stripe labels; the use of both an intensity image and a striped image, allowing scenes to be represented by detected edges along with 3D surface patches; and a pose-clustering algorithm, a uniform technique to accumulate matching evidence for recognition while averaging out substantial errors of pose. >

Laboratory facility for flexible structure control experiments

Abstract: A laboratory facility to study various control problems related to flexible mechanical structures is described. Computer, interfacing, and software issues are discussed. A novel proof-mass actuator is presented. A free-free suspended-beam experiment and a skewing-beam experiment are described. Three additional experimental setups at various stages of development are briefly considered. >

Rational linear algebraic tracking control system design

Abstract: Basic issues and methods in tracking control system design are considered, with emphasis on the logical nature of the design process. The controller transfer functions are not required to share the same poles, resulting in added design freedom. Classical root-locus-based design is placed in perspective, and practical ways of choosing the overall system transfer function to meet tracking performance objects are discussed in detail. >

Intelligent shape recognition for complex industrial tasks

Abstract: A knowledge-based shape representation and recognition system that can handle a large class of objects under less constrained situations than required for current machine vision system is proposed. Intelligent integration of different shape representation schemes and generation of the best shape recognition strategy are carried out using global shape properties. The proposed scheme effectively incorporates model-driven top-down and data-driven bottom-up approaches of shape analysis. By analyzing global shape properties, the essential features and their degrees of importance are determined quickly. In the representation phase, objects are described by using these essential features; in the recognition phase, the search for the best candidate is restricted to the models represented by these features, and the observed shape is matched to the candidate models in order of importance of the essential features. Systems are being developed for 2D and 3D shapes separately since they exploit different visual data, i.e. photometric and range, respectively. >

Computational stereo vision using color

Abstract: An investigation of the efficiency of color or chromatic information as an aid in solving the stereo correspondence problem is described. Two very simple stereo algorithms differing only in the use of chromatic information are applied to images containing various chromatic characteristics. The intensity-matching algorithm matches the zero crossings in a Laplacian-of-Gaussian filtered image, while the color-matching algorithm includes chromatic gradients that characterize the intensity zero crossings. The algorithms have been defined with a very simple matching structure in an effort to isolate the effects of including chromatic information in early stereo processing. The results indicate that the use of chromatic information can significantly reduce the ambiguity between potential matches while increasing the accuracy of the resulting matches. >

Application of modern control to a continuous anneal line

Abstract: A state-variable design approach is developed to control strip temperature in a continuous anneal process producing a variety of high-strength low-alloy steels. In this process, a steel strip passes continuously through five consecutive heat-treating operations to achieve desired metallurgical properties for the automotive industry market. Each operation is controlled by an individual controller designed to maintain the tight strip temperatures required for the product mix. Linear quadratic Gaussian control techniques are used to develop the control system. >

An overview of geometric modeling using active sensing

Abstract: A brief overview is given of techniques for constructing descriptions of 3D solid objects based on active sensing, whereby scene illumination is manipulated by controlling the light source or the imaging geometry. The light source can be controlled to cast light of a certain spatial pattern to reveal the surface structure of imaged objects. The pattern can be a single point, a single line segment, a set of parallel line segments, or orthogonal grids: and this spatially encoded scene is observed from one or more camera positions. Two types of analysis are possible to relate the encoded images to the 3D configuration of objects. In the first type, the relative position of the light source and the camera is obtained through a calibration process. The correspondence of features of the projector and the image is then established. The 3D positions of the encoded surface points are recovered by means of triangulation. The second type relates the orientation of the projected stripes in the image plane to the surface orientation and structure without the correspondence of features being established. >

Design of an autonomous navigation system

Abstract: The structural design of a system called ODYSSEUS is discussed. The system consists of a triangle base with three programmable wheels, a special digital compass based on the Grey code, a sonar system in a hexagonal configuration, a slope meter, and a microprocessor controller that selects information from the above components and determines the navigation paths. The system possesses learning capabilities, which are not treated here. >

Universal six-joint robot controller

Abstract: The authors describe the specifications, design, and implementation of a general-purpose six-axis robotic manipulator controller developed to serve as a research tool for investigating practical and theoretical aspects of control strategies in robotics. The 80286-based Intel System 310 was used for running the XENIX operating servo software as well as higher-level software that implements kinematics and path planning. A multibus-compatible interface board was designed and constructed to handle input/output signals from the joint motors of the robot manipulator. The universal controller is capable of driving robot manipulators equipped with electric joint motors and position optical encoders. To test functionality, the controller was connected to the joint motor DC power amplifier of a Unimate PUMA 560 arm, bypassing completely the manufacturer-supplied Unimation controller; proportional-integral-derivative (PID) control laws were installed into the XENIX operating system. Additional software drivers were implemented to allow application programs access to the interface board. All software was written in the C language. >

Computer control systems for automotive power trains

Abstract: The problems of idle speed control and transmission control are reviewed to illustrate how control-theoretic techniques are used in the design of power-train control systems. The structure of current automobile control systems is examined. Possible future system structures and the role of control theory in the development of such systems are discussed. >

Mechanism and control of a quadruped walking robot

Abstract: The quadruped walking robot called TURTLE-1 is described. A novel link mechanism named ASTBALLEM is used to construct highly rigid and easily controllable legs. Each leg has two degrees of freedom and is driven by two DC servomotors. The motion of the legs is controlled by a microcomputer, and various gaits are generated so that the robot walks not only statically but also quasidynamically. When the walking mode is static, the center of gravity of the robot is kept statically stable. When it is quasidynamic, a two-legged supporting period is required. >

Considerations in control scheme development for fermentation process control

Abstract: It is shown that more effective use of those measurements that are already available can provide an improvement in fermentation process control, which has been commonly viewed as almost insurmountable due to inherent measurement difficulties. Fermentation process problems that are significant in control system design are highlighted. Two industrial examples are given, showing how operation can be improved using currently available technology. One is a penicillin fermentation process, and the other is commercial mycelial fermentation. >