臺灣博碩士論文加值系統

本篇論文的目的在於描述一部能夠在大學校園中行進的自主導航車(
VIVA-I)視覺系統,我們提出一套整體性的規劃:包含系統架構的制定,作標
系統的轉換,道路環境的描述,各個模組的設計規格,VIVA-I整個系統分為
四大部份:視覺(感知)系統,導航系統,車輛控制系統,以及上層指令規劃,
經過近年來的努力,我們以經成功的發展一些模組,例如,路邊線偵測,夜間
高速公路邊線偵測,操場跑道偵測,斑馬線偵測,和一套導航模組,經過許多
的測試與整合,此項自主車初步發展完成,路試結果十分成功;於中正大學
校園內進行實車試驗,在有落葉,陰影干擾的情況下,我們的車子自行駕駛
且時速可達到二十幾公里,對於漆有邊線的道路,我們系統的處理時間需求
僅需0.1秒;若非受限於所用電動車的速度上限,在沒有其它車流的高速路
上時速超過100公里應可達到. 在此,我們提出一套整體性的視覺系統,
作為本實驗室未來設計的藍圖與發展的方向.

This thesis presents an integrated vision system modules
design for theautonomous vehicle VIVA-I developed in our
laboratory. The vehicle VIVA-I isdesigned to drive by itself in
our university campus. We proposed a completeplan: including
system configuration design, coordinate systemd transformation,
road environment descriptions, module specifications, and so on.
Our completevehicle navigation system can be divided into four
parts: perception system, navigator syatem, vehicle control
system and high-level navigation planner. In addition to the
navigator system, we have successfully developed quite a few
vision modules through our effort in the past years. These
include roadboundary detection, freeway road boundary detection
at night, multi-lane detectioin on stadium, and p[edestrian-
crossing detection. After a lot of testing and integration,
VIVA-I has complete its first stage of development,and the road
test of this autonomous vehicle has shown a great sucess.
According to the field experiments performed around Chung-Cheng
University campus, even under heavy interference of tree shadows
and falling leaves, our vehicle can drive autonomously with the
speed over 20 km/hr. For roads with painted boundary line, the
overall processing cycle of our electric vehicle,navigation
system is only around 0.1 sec. If not limited by the speed of
our electric vehicle, our navigation system should be able to
drive the car on freeway with speed over 100 km/hr. This
proposed integrated vision system can be the blue-print for the
future development in our laboratory.