A hierarchical neurocontroller architecture consisting of two artificial neural network systems for the manipulation of a robotic arm is presented. The higher-level neural system participates in the delineation of the...
详细信息
A hierarchical neurocontroller architecture consisting of two artificial neural network systems for the manipulation of a robotic arm is presented. The higher-level neural system participates in the delineation of the robot arm workspace and coordinates transformation and the motion decision-making process. The lower one provides the correct sequence of control actions. The capabilities, including speed, adaptability, and computational efficiency, of the developed architecture are illustrated by an example.< >
A computational approach for solving the correspondence problem between different views of objects in range images is presented. This is modeled as a layered constraint satisfaction network which can be implemented on...
详细信息
A computational approach for solving the correspondence problem between different views of objects in range images is presented. This is modeled as a layered constraint satisfaction network which can be implemented on a parallel analog neural network. In this approach, each view of an object is represented by an attributed graph with nodes as surfaces and their bounding vertices, and links as relations between adjacent surfaces. The matching strategy is a two-step process. Each step is formulated with a constraint satisfaction network, and implemented on a Hopfield network. At each level, a set of local, adjacency and global constraints is specified, and an appropriate energy function to be minimized is defined. At the first level of this hierarchy, surface patches are matched and clusters of rotation transformations are hypothesized. At the second level, the computed rotation transformation is applied to the corresponding vertices, and the translation vector is computed.< >
A mechanism for planning that is based on a temporal model that emphasizes the possible relationships between tasks in the time domain is presented. A control loop is devised to build, validate, execute, and evaluate ...
详细信息
A mechanism for planning that is based on a temporal model that emphasizes the possible relationships between tasks in the time domain is presented. A control loop is devised to build, validate, execute, and evaluate task plans. Each agent in the control loop receives feedback to ensure that robust, valid plans are prepared for execution. Status and side-effects from plan executions are evaluated and fed into the planning agent for replanning, providing an experimental architecture for autonomous operations. This mechanism has been developed to help automate maintenance and repair operations for space station Freedom.< >
A methodology is presented for building integrated planning-reacting systems. The work is based on a formal process-based approach to building the reactive component; this makes it possible to formalize the concept of...
详细信息
A methodology is presented for building integrated planning-reacting systems. The work is based on a formal process-based approach to building the reactive component; this makes it possible to formalize the concept of a planner improving a reactor. The planner design emphasizes how the planner can use the reactor to focus its reasoning, as well as how the reactor is guided by the planner to improve its behavior. The reactive component (the reactor) is based on a process-based model of robot computation, the RS model. This gives a powerful representation for actions with precise formal semantics. The duty of the planning component (the planner) is to adapt the reactor to suit a set of objectives and the possibilities afforded by the environment. Planner and reactor both operate continually, separately and in a complementary fashion.< >
Intelligent computational systems can be described as an adaptive computational system integrating both traditional computational approaches and artificial intelligence (AI) methodologies to meet the science and engin...
详细信息
Intelligent computational systems can be described as an adaptive computational system integrating both traditional computational approaches and artificial intelligence (AI) methodologies to meet the science and engineering data processing requirements imposed by specific mission objectives. These systems will be capable of integrating, interpreting, and understanding sensor input information; correlating that information to the “world model” stored within its data base and understanding the differences, if any; defining, verifying, and validating a command sequence to merge the “external world” with the “internal world model”; and, controlling the vehicle and/or platform to meet the scientific and engineering mission objectives. Performance and simulation data obtained to date indicate that the current flight processors baselined for many missions such as Space Station Freedom do not have the computational power to meet the challenges of advanced automation and robotics systems envisioned for the year 2000 era. Research issues which must be addressed to achieve greater than giga-flop performance for on-board intelligent computational systems have been identified, and a technology development program has been initiated to achieve the desired long-term system performance objectives.
The issue of structural formulation of task plan generation for intelligent machines is investigated from the perspective of artificial intelligence. The knowledge base of the machines is represented by a relational d...
详细信息
ISBN:
(纸本)0818621087
The issue of structural formulation of task plan generation for intelligent machines is investigated from the perspective of artificial intelligence. The knowledge base of the machines is represented by a relational database, and the task-planning process is modeled by a condition/event net. It is shown that the task plans generated by the formulation proposed are Petri net languages and that any task-planning strategy can be achieved by a supervisory task planner. The objective of such a structural formulation is to establish a formal framework for the task plan generation of intelligent machines in terms of concepts in mathematical logic. This framework can serve as the foundation of analytical task plan generation at different levels of the hierarchical structure of intelligent machines. In other words, the framework may play the role of domain space upon which various analytical design approaches, such as the probabilistic method and neural network computing, can be integrated to specify the task plan generation in both architectural and computational aspects.
A prototypical computational model of constrained configurative synthesis (CCS), which involves combining a set of components into a functional whole, is described. A number of expert systems for widely diverse applic...
详细信息
ISBN:
(纸本)0818621087
A prototypical computational model of constrained configurative synthesis (CCS), which involves combining a set of components into a functional whole, is described. A number of expert systems for widely diverse applications are briefly reviewed. Common features of these systems are then abstracted to indicate the general heuristics employed by experts in various domains. Next, a generic computational framework is outlined for the CCS problems. The author concludes by suggesting CCS as a basic cognitive mechanism underlying the organization and processing of knowledge structures.
This conference proceedings volume contains 94 papers. The following topics are dealt with: hierarchical intelligent control;perception-representation-action triad;nonconventional cost functions;team decision making;s...
详细信息
ISBN:
(纸本)0818621087
This conference proceedings volume contains 94 papers. The following topics are dealt with: hierarchical intelligent control;perception-representation-action triad;nonconventional cost functions;team decision making;swarm intelligence;perceptual robotics;shape representation;shape extraction and analysis;active perception;acoustic perception;intelligent computing systems;knowledge representation for learning control;neurocontroller design;neural network applications;fault-tolerant control systems;noise identification in linear systems;evidential and fuzzy reasoning in intelligent control.
暂无评论