Yukinori Kakazu

Genetic reinforcement learning for cooperative traffic signal control

Optimization of a group of traffic signals over an area is a large, multi-agent-type real-time planning problem without a precise reference model being given. To do this planning, each signal should learn not only to acquire its control plans individually through reinforcement learning, but also to cooperate with other signals. These two objectives-distributed learning of agents and cooperation among agents-conflict with each other, and a method that blends these two objectives together is required. In the method proposed in this paper, these two objectives correspond to localized reinforcement learning and global combinatorial optimization, respectively, and the method thus achieves cooperation in the long term without bothering with autonomy. The outline of the idea is as follows: each agent performs reinforcement learning and reports its cumulative performance evaluation, and combinatorial optimization is simultaneously carried out to find appropriate parameters for long-term learning that maximize the total profit of the signals (agents).<<ETX>>

EMG prosthetic hand controller discriminating ten motions using real-time learning method

We discuss the necessity of a learning mechanism for an EMG prosthetic hand controller, and the real-time learning method is proposed and designed. This method divides the controller into three units. The analysis unit extracts useful informations for discriminating motions from the EMG. The adaptation unit learns the relation between EMG and control command and adapts operators characteristics. The trainer unit makes the adaptation unit learn in real-time. Experiments show that the proposed controller discriminates ten forearm motions, which contain four wrist motions and six hand motions, and learns within 4/spl sim/25 minutes. The average of the discriminating rate is 91.5%.

Development of wearable exoskeleton power assist system for lower back support

In this paper, we propose a power assist device for lower back flexion and extension, when carrying a heavy load. To see the effect of the device, we model a human body and analyze a compression force in his lower back, as well as the evaluation of a supported force at a hand position. A prototype of the device is manufactured and a controller of the device is developed, which can follow a voluntary human motion assisting his strength.

Extended Depth-Buffer Algorithms for Hidden-Surface Visualization

Users of interactive CAD systems who are trying to cope with the hidden-surface visualization problem will want to investigate this new algorithm.

Evolution of Binary Decision Diagrams for Digital Circuit Design Using Genetic Programming

This paper proposes the methodology for hardware evolution by genetic programming (GP). By adopting Binary Decision Diagrams (BDDs) as hardware representation, larger circuits can be evolved, and they will be easily verified by utilizing commercial CAD software. The hardware descriptions specified in BDDs are improved by GP operators, to synthesize various combinatorial logical circuits.

Design of Compact and Lightweight Wearable Power Assist Device

In this paper, the development of a compact and lightweight power assist device for lower back flexion and extension, when carrying a heavy load, is presented. For investigating the effect of the device, we model a human body as a planar seven-link system, and we analyze a compression force in the lower back, particularly at the disc between the fifth lumbar vertebra and the first sacrum. We built two prototypes of the device, and we apply the model to real data acquired by the prototype. The analysis results show that the proposed device can reduce the compression force in the disc. A controller for the device is developed, which can follow a voluntary human motion using surface myosignals.Copyright

Generalization of the potential method for blending three surfaces

Abstract The potential method is one of the methods of generating a blending surface along the intersection curves of implicit surfaces. The paper describes a generalized method, for three surfaces, of yielding the formulae for the convex and nonconvex combinations, each of which retains the locality of the blending regions, and has the ability to control the ranges of the edge blends along the original sharp edges. This is analogous to variable-radius rolling-ball blends. To obtain the satisfactory blending functions, the polarity and special pencils of quadrics are applied. After these derivations, it is pointed out that the other quadratic blends for three surfaces correspond to the special cases of the described method. Further, the projective potential method is presented for a convex combination of the three surfaces.

Controlling Dynamics of GA through Filtered Evaluation Function

As a function optimizer or a search procedure, GAs are very powerful and have some advantages. Fundamental research concerning the internal behavior of GAs has highlighted their limitations as regards search performances for what are called GA-hard problems. The reason for these difficulties seems to be that GAs generate insufficient strategies for the convergence of populations. To overcome this problem, an extended GA, which we name the filtering-GA, that adopts the method of changing the effect of the objective function on the dynamics of the GA, is proposed.

Self-organized Control Of Traffic Signals Through Genetic Reinforcement Learning

This paper introduces a learning approach for traffic signal control. The reinforcement learning is intended to optimize the traffic flow around cross roads, while the Genetic Algorithms are intended to introduce a global optimization criterion to each of the local learning processes. It is shown that the combination of reinforcement learning and the Genetic Algorithm exhibits good performance for dense traffic conditions.

RESEARCH ON 3-D GEOMETRIC MODELING BY SWEEP PRIMITIVES

This paper proposes a new modeling method for not only ordinary mechanical part geometry but also special ones, such as a coil or a bend etc. The new method is called “Sweep Primitives”, because of constructing a volumetric primitives by sweep operation and combining these primitives by set operation. How to develop a sweep primitive is as follows, namely, named “secondary pattern” as spine of sweeping/trajectory curve and “primary pattern” as being the sweeped pattern along the secondary pattern. And these primary and secondary patterns are represented by “language” then we can use these with other volumetric primitives. As a result of developing and adapting this sweep primitive method, most of the mechanical part geometries are easily modeled with less effort.