Application of neuro-fuzzy system to agricultural vehicle for local map construction and navigation

in International Symposium (oral presentation paper), 國際研討會(全文口頭發表)
標題Application of neuro-fuzzy system to agricultural vehicle for local map construction and navigation
AuthorsBo-Han Wu(advisee), 吳柏翰(研究生), & Chung-Liang Chang 張仲良
出版日期Jun 18 2012 12:0
會議地點Jeonju, Koera

This paper proposes and independent navigation and crop detection function
equipped in agricultural vehicle for the orchard and the greenhouse’s crop production
system to reduce its, huge manual labor and costs. The vehicle equipped with a low
cost laser range finder and an odometer to build the environmental map and gains the
information of vehicle position. At the same time, we use the Neuro-Fuzzy method to
decrease the moving error caused by the odometry. The proposed system can
calculate the wheel speed difference of vehicle and to control the direction of vehicle.
First, we define the environment’s characteristic point and static state obstacle during the vehicle moving. A path planning of the shortest way the vehicle moved is also designed. After that, the optimal performance of system is to be achieved by training the Neuro-Fuzzy system during the vehicle moving, we use the laser range finder to carry out the movement and static state detection under unknown environment. Meanwhile, the odometer is to estimate the moving distance of vehicle. The vehicle can positioning and perform the path planning immediately. In addition, the vehicle equips with temperature/humidity (humiture) sensory and sprayer. The favors for vehicle to advance carry out the crop examination and automation perfusion, by reality test in the orchard. Under the 8 cm/sec vehicle’s moving speed, it will result average moving error 1.1 cm, the miscarriage probability of spraying is 5%. This design concept of proposed vehicle will apply to automatic cultivation system in the future.

校址:912 屏東縣內埔鄉學府路1號 總機:886-8-7703202 傳真:886-8-7740165 系統開發統維護單位:國立屏東科技大學 電算中心 版權所有