The authors investigate the generalization ability of the network generated by S.E. Fahlman and C. Lebiere's (FL) (1990) learning algorithm which has a distinctive feature to build any network topology. They train...
详细信息
The authors investigate the generalization ability of the network generated by S.E. Fahlman and C. Lebiere's (FL) (1990) learning algorithm which has a distinctive feature to build any network topology. They trained the same three-layer network by the FL algorithm and the backpropagation (BP) algorithm, and a comparison of recognition abilities shows that the FL network performs much better than the BP network. The FL network performs better because, in this network, hidden units use only saturated values and thus the hidden layer acts as a filter for noise. A two-layer network performs excellently if the training set is trainable by the two-layer network and if a pattern is recognized by detecting the maximum valued output. Since the FL algorithm begins with a minimal two-layer network which performs best under the condition stated, a designer can construct either a two-layer or a multilayer network according to which one best fits a particular application. Thus, it can be concluded that in all these respects the FL algorithm is preferable to the BP algorithm.< >
A type of miniature linear actuators called a cybernetic actuator for biomedical applications is described. This actuator has four driving states: free, increasing, decreasing, and lock. The free and lock states can r...
详细信息
A type of miniature linear actuators called a cybernetic actuator for biomedical applications is described. This actuator has four driving states: free, increasing, decreasing, and lock. The free and lock states can release excessive applied load from outside to provide safety characteristics for both human and micromechanisms. Miniaturized linear cybernetic actuators combining a piezoelectric impact drive with electromagnetic clamping were designed and developed. Feasibility and basic performance of the actuators were confirmed by theory and experiments. Improved performance in terms of maximum speed, energy efficiency, and silent drive as a miniature linear actuator were verified successfully by using a controlled friction impact drive.< >
A neuro fuzzy position controller for a servomotor, which is a nonlinear controller and gives much higher control performance for highly nonlinear systems than a linear controller, is proposed. The simulation results ...
详细信息
A neuro fuzzy position controller for a servomotor, which is a nonlinear controller and gives much higher control performance for highly nonlinear systems than a linear controller, is proposed. The simulation results show that it is robust and has self-tuning capabilities.< >
A neuro fuzzy position controller for a servomotor, which is a nonlinear controller and gives much higher control performance for highly nonlinear systems than a linear controller, is proposed. The DSP hardware archit...
详细信息
A neuro fuzzy position controller for a servomotor, which is a nonlinear controller and gives much higher control performance for highly nonlinear systems than a linear controller, is proposed. The DSP hardware architecture for implementation of neural networks has to be complemented by optimized program coding to avoid the latency effects due to its pipelining structure as to achieve higher throughput. The overall system, which involves tasks performed by the microcomputer and DSP, and the implementation of the sigmoid function using a lookup table are described.< >
This paper represents a physical model of two -dimensional transpiration cooling control system with surface heating and ablating. The governing equation is derived and the third boundary conditions are given. They co...
详细信息
This paper represents a physical model of two -dimensional transpiration cooling control system with surface heating and ablating. The governing equation is derived and the third boundary conditions are given. They constitute a two -dimensional variable -domain distributed parameter control system in which the control parameter appears in both the governing equation and the boundary conditions. Applying time semi-step alternating direction method to difference this mathematical model, the tridiagonal difference equations are obtained. For a flying shell of electromagnetic railgun, the mathematical simulation results agree with the realistic physical process, and the control parameter not only can control the temperature of heat shield, but also can control the boundary motion effectively.
A finite difference scheme offering second-order accuracy is introduced to solve numerically a system of two mixed-type coupled partial differential equations with variable coefficients. The stability conditions of th...
详细信息
＜正＞The unification problem is one of the most important and fundamental problems in many fields,such as artificial intelligence,knowledge engineering,intelligent robotics,deductive database,and new-type programming...
＜正＞The unification problem is one of the most important and fundamental problems in many fields,such as artificial intelligence,knowledge engineering,intelligent robotics,deductive database,and new-type programming *** paper presents an AND-OR terms graph for representing a power set of terms,and establishes a homomorphic mapping from the terms graph to a Harmony network for solving the first order unification problem under some equational ***,in order to construct the homomorphic Harmony network an improved competitive activation mechanism is presented,which allows more than one nodes within an inhibitory cluster of nodes to be activated simultaneously and avoids running the time-consuming Harmony network learning algorithm based on complicated stochastic *** algorithm,given by this paper,used for getting the maximum Harmony state of the homomorphic Harmony network,which is associated with the solution of unification problems,can update network states in such a way that several nodes corresponding to the same variable in a given set of terms always result in the same value through a stochastic dynamic *** approach and the algorithms in this paper are superior to the traditional symbolic logic methods in the high parallelism, the ability of solving the unification problem between uncertain and incomplete terms,the ease of treating combination of several equational theories,and the adaptability for many various types of the unification problems,especially,for solving many terms’ unification simultaneously and parallelly.
In this paper we discuss the overall goal, system design, application of key techniques and development of OIMS----Oil-mine Intelligent Management system.
In this paper we discuss the overall goal, system design, application of key techniques and development of OIMS----Oil-mine Intelligent Management system.
The authors discuss active tactile sensing for detecting a contact point between a multifingered robot hand and an unknown object. They define the concept of actively sensible meaning that an active sensing motion can...
详细信息
The authors discuss active tactile sensing for detecting a contact point between a multifingered robot hand and an unknown object. They define the concept of actively sensible meaning that an active sensing motion can be planned for a sensing event. Assuming more than two sensing events, conditional-minimum-external-sensor-realization is defined, where it is shown that it is possible to reduce the number of external sensors with a well-planned sensing algorithm. Based on these definitions, an active sensing algorithm for detecting the contact point with minimum-external-sensor-realization is proposed. This approach remedies several disadvantages found in conventional approaches. The approach was tested through experiments using a single-finger model with a fingertip tactile sensor.< >
暂无评论