By Topic

Intelligent Control, 1989. Proceedings., IEEE International Symposium on

Date 25-26 Sept. 1989

Filter Results

Displaying Results 1 - 25 of 94
  • Proceedings. IEEE International Symposium on Intelligent Control 1989 (Cat. No.89TH0282-4)

    Publication Year: 1989
    Save to Project icon | Request Permissions | PDF file iconPDF (34 KB)  
    Freely Available from IEEE
  • Experiments in robot learning

    Publication Year: 1989 , Page(s): 561 - 565
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (420 KB)  

    The authors attempt to bridge the planning gap problem that exists between the AI (artificial intelligence) and robotics communities. the objective was determining how procedural and parameter knowledge could be represented and used in task planning. Because AI planners lack the ability to deal with kinematic and kinetic information of a world model, and robot planners possess poor reasoning ability, an advanced robotics research environment was considered the appropriate demonstrator. Experiments were conducted on two generic operations that are common to robotics work, grasping and pushing an object, working from a starting point that was randomly selected. Task parameter constraints are derived from rules induced from the small amounts of raw sensory data collected. These rules are then used to indicate whether a task, such as object grasp or push, could be successfully completed View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Path planning for an intelligent robot by the extended VGraph algorithm

    Publication Year: 1989 , Page(s): 544 - 549
    Cited by:  Papers (5)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (400 KB)  

    In many path-planning algorithms, attempts are made to optimize the path between the start and the goal in terms of Euclidean distance. Since the moving object is shrunk to a point in the configuration space, the findpath problem can be formulated as a graph-searching problem. This is known as the VGraph algorithm. However, its first drawback is related to the rotation of a moving object. This drawback has been solved by using the sliced projection method. The VGraph algorithm has serious drawbacks when the obstacles are three-dimensional. The extended VGraph algorithm is proposed to solve the drawbacks of the VGraph algorithm by using the recursive compensation algorithm. The latter is to find the collision-free shortest path in 3 D without increasing the complexity of the VGraph, and it is shown to converge the shortest path in 3 D. It is noted that, simplifying the VGraph, the extended VGraph algorithm can save not only the memory space to represent the VGraph but also the graph-searching time View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Path planning under uncertainty from a decision analytic perspective

    Publication Year: 1989 , Page(s): 556 - 560
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (384 KB)  

    Previous work has used a certainty grid for navigation and path planning. In the present work, the author attempts to formulate the path planning problem under uncertainty from a decision analytic perspective. Paths are generated based on the planner's preferences and expected utilities of actions. A proposed cost structure, controlled by a strategy index, can simulate attitudes ranging from risk aversion to risk seeking. Experimental results demonstrate how path length and path risk are traded off based on this index. The proposed framework for navigation under uncertainty has proven to be a powerful tool in encoding subjective preferences in path planning. With a solid theoretical foundation such as decision analysis and utility and game theory, uncertain spatial and other knowledge terms can be assimilated in the overall plan. The preliminary results so far have shown a promising outlook for further expansion of this technique to real-world problems View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Collision-free path planning for multirobot systems

    Publication Year: 1989 , Page(s): 537 - 542
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (408 KB)  

    The authors discuss a real-time algorithm for planning collision-free navigation or manipulation paths in dynamic, cluttered environments. A Newton-like iteration is used to determine a sequence of knot points that guide the motion of several robots in a velocity vector field toward their targets while avoiding collisions with a set of moving obstacles. Complex targets and obstacles are modeled with elliptical primitives. The velocity field can be updated online, and the path locally replanned in response to environmental changes. The velocity field approach is applied to two main problems: (i) two-robot path planning with curvature constraints, and (ii) coordinated redundant and bilateral manipulation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Model based production system for on-line design of minimum time control cycles

    Publication Year: 1989 , Page(s): 595 - 600
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (444 KB)  

    A nonlinear stochastic system is considered which is supposed to operate under conditions of constantly changing goal assignment as well as varying components of the mathematical model of the system and constraints. A method is proposed for using a model-based production system for designing minimum-time control cycles of mechanical motion for an application in which control assignment can be changed during the command execution. Based upon an enhanced Bellman optimality principle for stochastic systems, the string of commands should be recomputed as soon as the new goal of motion is assigned. The problem of computational efficiency is solved by using a model-based production system for online design of the motion cycle View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Planning 3-D collision-free paths

    Publication Year: 1989 , Page(s): 550 - 555
    Cited by:  Papers (4)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (472 KB)  

    In many robot applications, such as assembly and repair operations, the automatic development of collision-free paths is of vital importance. The authors present the results of obtaining such a path with a novel generate-and-test approach, which is both rapid and exact in a three-dimensional, dynamic environment. The approach uses successive spherical approximations (SSA) as a complete model for three-dimensional objects and a path planner which uses rules based on heuristic understanding of the three-dimensional relationships of objects in a known environment. It is shown that the SSA collision detection method can provide marked improvement over direct methods of detecting collisions. The average time required to find collisions with all obstacles along a path as required for path planning in a fairly clustered environment is under 10 s, an improvement over three times compared to a method using face-by-face comparisons View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Autonomous navigation to provide long-distance surface traverses for Mars Rover Sample Return Mission

    Publication Year: 1989 , Page(s): 362 - 367
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (508 KB)  

    Knowledge acquisition and representation techniques are developed to allow an autonomous vehicle to negotiate a path over partially unknown terrain, such as the long-distance traverses for the Mars Rover Sample Return Mission. Terrain navigation is accomplished by the following sequence of steps: the skylines from the visibility map and from the local terrain map are compared and the two 3-D curves in space are stochastically matched to update the vehicle location. A path segment is selected taking into consideration the goal, overall global topography, obstacles such as craters and boulders derived from the range-finder observations, and a priori constraints related to the mission, fuel consumption, and vehicle stability. The vehicle reaches a new vantage point with relatively broad visibility, but within sight of the last observation point. The skyline is then estimated and approximated again, and the cycle is reiterated for another advance View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neuromorphic learning of continuous-valued mappings in the presence of noise: application to real-time adaptive control

    Publication Year: 1989 , Page(s): 312 - 319
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (464 KB)  

    The ability of feedforward neural net architectures to learn continuous-valued mappings in the presence of noise is demonstrated in relation to parameter identification and real-time adaptive control applications. Factors and parameters influencing the learning performance of such nets in the presence of noise are identified. Their effects are discussed through a computer simulation of the back-error-propagation algorithm by taking the example of the cart-pole system controlled by a nonlinear control law. Adequate sampling of the state space is found to be essential for canceling the effect of the statistical fluctuations and allowing learning to take place View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parameter learning for performance adaptation in large space structures

    Publication Year: 1989 , Page(s): 589 - 594
    Cited by:  Papers (4)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (544 KB)  

    A parameter learning method is introduced which is used to broaden the region of operability of the adaptive control system of a space structure. The learning system guides the selection of control parameters in a process, leading to optimal system performance; the method is a form of learning by observation and discovery. It is applicable to any system where performance depends on a number of adjustable parameters. A mathematical model is not necessary as the learning system can be used whenever the performance can be measured via simulation or experiment. The results of a transient regulation experiment are presented. In this experiment, the learning system determines the best parameter values to use when given different disturbances View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Position and force control of robot arms

    Publication Year: 1989 , Page(s): 251 - 256
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (540 KB)  

    A stability analysis for robot manipulators under the influence of external forces is presented. Several control objectives are considered: rejecting the external force as a source of disturbance, complying with the external force as a generalized mass-spring-damper system, and actively controlling the external force when a dynamic model of the environment is available. In each of these cases, a stability analysis using the Lyapunov method is presented View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive processing with neural network controlled resonator-banks

    Publication Year: 1989 , Page(s): 306 - 311
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (376 KB)  

    The author describes a novel neuromorphic architecture for structurally adaptive control systems. The neural network controlled resonator-bank (NCRB) architecture consists of two main components, a resonator-bank filter structure and a neural network which controls the transfer characteristics of the filter. The architecture offers an attractive alternative to the approximation of nonlinear dynamic systems having a finite number of stable operating points with quasi-linear behavior. The first results of the experimental analysis of the NCRB structure are encouraging. The system is apparently able to approximate a wide range of nonlinear dynamic behaviors View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Task planning for Robotworld: a case study

    Publication Year: 1989 , Page(s): 348 - 353
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (356 KB)  

    The authors describe an implemented task planning system for the placement of electronic components on a printed circuit board (PCB). The task planner assumes an assembly workstation which is similar to V. Scheinman's Robotworld (1987), in which a number of small two-degrees-of-freedom robots slide on a horizontal platten above the PCB. After obtaining a PCB specification, the task planner generates a feeder layout, a feeder setup, a component assembly sequence, and a motion program for each of the robots. To test the task planner, the output is simulated graphically. For relatively small PCBs, all of this occurs within minutes, enabling quick turnaround from design to manufacture View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quasistatic sliding manipulation-on the transient response of sliding manipulation

    Publication Year: 1989 , Page(s): 438 - 443
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (424 KB)  

    Analytic models and numerical simulations of transient sliding are presented. The sliding motion of compliant grasps is governed by a first-order time-varying differential equation. The solution of such differential equations is characterized by a study-state and a transient solution. The transient solution is more important because it happens immediately after the sliding motion initiates so that it is necessary to control it more carefully, and the fingertips are usually in transient stages with little opportunity to establish steady-state sliding. Under most circumstances, the approximate solutions of the time-varying system were found to be close to those of direct numerical computation. Several experiments on the sliding motions of fingers were conducted using a model frame and disk that represent the fingertip. The experiments gave good qualitative and quantitative agreement with the theoretical and numerical solutions View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Dynamic modeling and control of nonlinear processes using neural network techniques

    Publication Year: 1989 , Page(s): 280 - 286
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (452 KB)  

    An adaptive network architecture of nonlinear elements and delay lines is proposed, which can be taught to model the time responses of a nonlinear, multivariable system. The structure has been applied to the modeling and control of a highly coupled multivariable process, namely, gas tungsten arc (GTA) welding. The authors present the architecture, learning algorithm, and experiments which showed the feasibility of the approach, and propose a controller architecture that can regulate a nonlinear, multivariable plant such as GTA welding View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An evidential reasoning extension of model-based failure diagnosis

    Publication Year: 1989 , Page(s): 520 - 525
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (492 KB)  

    A model-based failure detection and isolation methodology is considered which utilizes diagnostic models consisting of orthogonal parity equations. To improve the robustness of the algorithm, several models may be used in parallel, providing potentially incomplete and/or conflicting inferences. Dempster's rule of combination is used to integrate evidence from the different models. The probability measures are assigned on the basis of quantitative information extracted from the online model computations. A simulation example is presented View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Design of hierarchically distributed expert controllers for large-scale systems

    Publication Year: 1989 , Page(s): 12 - 17
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (396 KB)  

    An architecture with expert distributed controllers to control large functional systems hierarchically is considered. A structural hierarchy was formed based on the physical structure of the system and functions associated with the structure. The hierarchy and the functional associations are described completely by the structural and functional coordinability axioms. These axioms display a very specific type of hierarchical structure. One of the properties of the hierarchy is that the whole system is represented both at the top and at the tip levels. It has the ability to represent the whole system at any desired detail. Moreover, representational redundancies are deliberately allowed, but the top and tip levels are forced to be represented without redundancies. The authors define the functional behavior of the system and the expert controller via six primitives. These primitives express the functions, capabilities, restrictions, and information to control the behavior of the system. Finally, the control process is described as a uniform flow from top to tip View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Programming a damping matrix for error-corrective assembly

    Publication Year: 1989 , Page(s): 334 - 339
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (672 KB)  

    The author considers the situation in which the response of a manipulator to forces imposed on it is given by a user-selectable damping matrix (a six-by-six matrix which maps forces and torques into translational and rotational velocities). He then seeks to determine whether it is possible to choose the 36 matrix elements so that the forces which characterize every error condition map into the motions that correct it. He proposes three desired properties of the damping matrix: the bounded-forces, error-reduction, and nonorthogonal-motion properties. It is shown how the damping matrix most closely attaining these properties under all contact configurations can be constructed. For some tasks a `perfect' damping matrix possessing all three properties exists, and under its direction assembly must necessarily proceed to completion. For some tasks a perfect damping matrix does not exist, and it is conjectured that the three properties may still be useful as heuristics; the damping matrix most closely attaining them will likely be appropriate to the task. To test this conjecture, the author considers the canonical peg-into-chamfered-hole task and finds that no 'perfect' damping matrix is possible View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Implementation of intelligent controller using neural network state estimator

    Publication Year: 1989 , Page(s): 413 - 416
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (200 KB)  

    It is shown that the neural network can be used as a state estimator in a model-reference intelligent control system. Its learning capability and noise rejection characteristic are illustrated by the results of a simulation study. The implementation of the state estimator by a neural network was possible due to the development of a proper structure of the neural network which is capable of simulating the dynamic behavior of a linear or nonlinear system. This capability is achieved by use of a time-dependent learning process View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Entropy and outcome classification in reinforcement learning control

    Publication Year: 1989 , Page(s): 108 - 114
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (612 KB)  

    Several mathematical structures are proposed for learning control systems using a wide array of techniques from a variety of disciplines. Reinforcement learning offers the greatest degree of flexibility in utilizing process information in conjunction with concepts from optical control while maintaining the basic constructs used in mathematical learning theory at the direct control layer. The adaptive layer in the control hierarchy is developed based on two fundamental properties concerning the grouping of outcomes which must be satisfied if a control policy exists for the process in terms of the defined neighborhoods. The original control objective can be interpreted in light of these two objectives and a control policy will be synthesized once these conditions are satisfied by all neighborhoods constructed during the process of learning View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Autonomous planning under uncertainty: reset-identification strategies

    Publication Year: 1989 , Page(s): 465 - 470
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (432 KB)  

    The author investigates the problem of planning state-identification strategies. These strategies help an agent to reduce its uncertainty about the state of the environment through active exploration of the environment. Specifically, the author emphasizes the gathering of information from the environment where the agent utilizes the (entropic) actions of its control operations to provide maximum information gain. An interesting class of state-identification strategies is the class of those that use no sensing capabilities, i.e. sensorless strategies. Necessary and sufficient conditions for the existence of such strategies are presented. These conditions are exploited to develop an algorithm for planning sensorless state-identification strategies View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Path planning expert system for pipeline inspection robots

    Publication Year: 1989 , Page(s): 299 - 304
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (420 KB)  

    The authors discuss sensing planning based on system knowledge for pipeline inspection and intelligent control of a maintenance robot. The pipeline inspection robot can carry out inspection tasks to detect the malfunction location autonomously in a plant pipeline system. It can estimate and search failure causes. The path planning system is shown to find the optimal path for the robot from one point to another under specified criteria. Furthermore, the sensors, for example, infrared ray proximity sensors or ultrasonic sensors, are installed to detect obstacles such as flanges and T-joints, and the robot can move autonomously under microprocessor control. Using PRES (path planning expert system) and SPES (sensing planning expert system), the robot becomes intelligent enough to carry out given inspection tasks automatically View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of AI to real-time intelligent attitude control of spacecraft

    Publication Year: 1989 , Page(s): 287 - 292
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (408 KB)  

    Modern spacecraft demand from an attitude control system very high performance and accuracy, and many new features, such as self-modification, adaptability, more autonomy, and fault tolerance. It is noted that artificial intelligence and knowledge-based system techniques can be applied to spacecraft attitude control to meet the demanding requirements and to provide many new features. The author discusses the need for and the benefits of integrating artificial intelligence and automatic control techniques. This approach gives the attitude control system the power of reasoning, judgment, and learning and the capability to cope with unexpected and changing situations. An architecture for a real-time intelligent attitude controller is described, consisting of a traditional low-level controller and a knowledge-based system provides meta control to the low-level controller View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiple model-based control of robotic manipulators: theory and simulation

    Publication Year: 1989 , Page(s): 53 - 58
    Cited by:  Papers (8)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (504 KB)  

    The multiple-model-based control (MMBC) technique utilizes knowledge of nominal plant dynamics and principles of Bayesian estimation to provide parameter-independent trajectory tracking accuracy. The MMBC algorithm is formed by augmenting a model-based controller with a closed-loop form of multiple-model adaptive estimation (ΔMMAE). The ΔMMAE uses perturbation models of the combined plant and feedback control system, along with measurements of tracking error, to provide an estimate of the plant parameters. When MMBC is applied to the robotic manipulator control problem the ΔMMAE provides a payload estimate. The model-based controller combines the a priori knowledge of robot structure with the payload estimate to produce the multiple models of the manipulator dynamics required to maintain controller accuracy. MMBC provides a unique solution to the problem of maintaining trajectory tracking accuracy in uncertain payload environments View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • MIMIC: a robot planning environment integrating real and simulated worlds

    Publication Year: 1989 , Page(s): 459 - 464
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (448 KB)  

    The authors describe MIMIC, an interactive multirobot programming environment and robot vision system which allows program development and testing in an integrated real and simulated world. In this environment programs under development can easily be fed either to a real or to a simulated robot. When operating in the simulation mode the robot environment appears as a mix of the real and simulated worlds, called partial simulation. This is created by combining real image data obtained from the robot vision system with graphical simulation techniques, which thus allows simulation to mirror the real world more closely than completely graphical simulation. The advantages of this integrated approach over separate offline graphical robot simulators are discussed and demonstrated by a number of applications, including an interactive programming application and a two-robot task involving collision avoidance View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.