Naoki Imasaki

Application of fuzzy neural network control to automatic train operation and tuning of its control rules

We have proposed two-degree-of-freedom fuzzy neural network control systems. It has a hierarchical structure of two sets of control knowledge, thus it is easy to extract and refine fuzzy rules before and after the operation has started, and the number of fuzzy rules is reduced. This paper shows an example application of automatic train operation and presents a rule tuning method.<<ETX>>

Elevator group control system tuned by a fuzzy neural network applied method

We have developed a high-performance elevator group control system with a performance tuning function, which employs a fuzzy neural network as a performance forecasting model of the elevator system. The fuzzy neural network, which is a structured neural network based on a fuzzy reasoning framework, stores the correlation between control-parameters and the response of the elevator group as a fuzzy rule set. It performs a fuzzy rule-based reasoning to forecast the system performance of the elevator group. The performance tuning function utilizes the forecasting model in order to search the optimal control parameters which give the best system performance in the present traffic situation. The fuzzy neural network applied system can automatically adapt itself to various traffic situations. This paper gives the overview of the elevator group control system with the fuzzy neural network and shows the validity of the methodology by some simulation results.<<ETX>>

Adaptive control of robot manipulators with anti-backlash gears

This paper proposes a control method for robot manipulators that have anti-backlash gears in the joints. The anti-backlash gear is modeled as a three segment flexible joint characteristic. A controller consists of a PD feedback part and a full dynamics feedforward part. This control method employs the sliding control scheme. In order to make the tip of the manipulator track desired trajectories and to eliminate steady state errors, this method utilizes two sliding surfaces for both links and actuators. For this purpose, positions and velocities of both links and actuators are fedback. An anti-backlash inverse characteristic is introduced to obtain the desired trajectories for the actuators from the desired trajectories of the links. An adaptation scheme to the variation of payloads is also investigated.

Elevator group control system with a fuzzy neural network model

We have developed a high-performance elevator group control system EJ-1000FN with a performance tuning function, which employs a fuzzy neural network as a performance forecasting model of the elevator system. The performance tuning function utilizes the forecasting model in order to search the optimal control parameters which give the best system performance in the present traffic situation.<<ETX>>

Distributed Self-Simulation of Holonic Manufacturing Systems

This paper details the development of a distributed simulation capability for holonic manufacturing systems based on the concept of self-simulation. Every holon in the system would function similar to an autonomous simulator that maintains its own clock and event execution. This paper discusses synchronization and other technical issues in designing such a simulator and puts forward the basic requirements and design options for its implementation. These are results of HMS project in IMS (Intelligent Manufacturing Systems) program.

Fuzzy artificial network and its application to a command spelling corrector

The authors propose a fuzzy artificial network (FAN) which utilizes associative memories and is constructed by a method which makes it easy to represent and to modify fuzzy rule sets. Whereas conventional fuzzy inference methods induce much fuzziness on multilayered fuzzy rule sets, the associative-memory-based FAN results in inferences which fit human senses better. This type of fuzzy inference is called associative inference. For memorizing fuzzy rule sets, the FAN system employs a correlation matrix which is constructed from a nominal correlation matrix, a bias matrix, and a scale parameter, so that it is easy to carry out refinement and cut-and-paste operations for rule sets. Using a FAN development system, a command spelling corrector is proposed which uses a multilayered fuzzy rule set. The spelling corrector application shows the eligibility of associative inference for multilayered fuzzy rule sets.<<ETX>>

Modeling Methodology for Systems with Ambiguity and Uncertainty

Abstract This paper discusses a modeling methodology for systems that contain ambiguity or uncertainty. Models built by the methodology are expected to handle the ambiguity and adapt to the uncertainty. The methodology is simple and versatile so that it can be applied to the description of various system architectures; e.g. multi-layered perceptron, fuzzy neural networks and so forth. A description language and a design tool that employ the method are also introduced.