Taku Okuno

Try to produce exceptional level skilled information engineers in the Information Engineering School of Hokkaido University

Many companies always demand exceptional level engineers from universities for the continual research and development of new techniques, improved production engineering and products systems. On the other hand, the recent rapid progress of engineering fields has been widening a gap between academic interests and practical activities, so the academic education system must be improved for the best engineers. Hokkaido University and several companies agreed to a plan to produce exceptional skilled software engineers, and started a program for Master Course of Software Engineering in April, 2003. The present paper reports major results on its unique curricula, how to integrate lectures and practices which cooperate engineers join to assist, self-management of our laboratory by students, a cooperative internship program and so on.

A Planning System for Precision Farming Based on an Autonomous Tractor

This paper describes the planning system developed for precision farming. Particularly, the system aims to support the cooperative works between global positioning system (GPS) based autonomous tractor and its applicators. The autonomous tractor can run precisely in fields according to GPS based navigation. The applicators can be controlled to perform variable outputs according to the precise position in fields. Using the tractor and its applicators, we can realize precision farming according to several conditions in fields. In order to use the autonomous tractor system, the planning system is developed for storing the field information, navigating the tractor and planning the variable farm-work. In this paper, we show the outline of the total systems, planning methods of navigation and optimizing variable farm-work. The results of experiments applied to the planning system will be shown.

Path Planning for Precision Farming Based on Autonomous Vehicles

Abstract This paper describes the path planning system for precision farming based on autonomous tractor with GPS. In order to use the autonomous tractor for the precision farming, we classify the path planning problems into three types. One is the path along the furrows in a crop field. Another type is the path for the replenishment. Third type is the path for working in a meadow or snowy field. Especially, to generate the optimum paths at the planning of meadow, we formulate the covering problem and apply the approach using genetic algorithm.

Field Information System for Map-Based Precision Farming

Abstract This paper describes the development of a field information system for the experimental implementation of map-based precision farming. The system is based on the geographic information system and is an aggregation of the application programs designed to help users make work plans. The system at present consists of five major applications: The maintenance of farm field maps, the planning of working paths, the drawing of application maps, the superposition of the application maps, and the planning of replenishment paths. These applications are integrated into a web-based application available via the Internet.

Integration of Learned Knowledge by Structured Boltzmann Machines.

This paper proposes a framework for integrating independently learned knowledge by Boltzmann machines. Auto-associative neural networks based on distributed representation have some advantageous properties of dealing with knowledge. However, as to the supplement of knowledge under a dynamically changing environment, or the merging of independently acquired knowledge, they do not work well enough because of difficulty in supplementary learning. Therefore we propose a model for integrating knowledge by connecting independently learned network modules without additional learning. First, localized representation, which is indispensable for realizing interaction between knowledge in different modules, is introduced. Although it will spoil the flexibility of distributed representation, competitive learning adopted for translating representations preserves it to some extent. Secondly, the configuration of the connected network is explicated, and is behavior is illustrated. It is completely realized by activation, competition, and inhibition of Boltzmann machine. Finally, to demonstrate the effectiveness of the proposed model, results of computer simulation on a navigation problem of obstacle avoidance are shown.

Merging Navigation Strategies For Vehicles By Structured Boltzmann Machine

In this paper, a network model for merging previously acquired strategies has been proposed and the results of some computer simulation shown. The simulation results show that the proposed framework works well for strategy merging in the vehicle navigation task.