在無人搬運車系統中防撞和導航為主要需求，本研究目的在於使用簡易架構的自走車以發展無人載具防撞系統，利用線影像感測器之信號，實施搬運車在速度控制和方向控制之決策，達到自動防撞之功能設計，並以三種不同類型車道進行雙車及三車的隨車與防撞之驗證。結果顯示，本研究所提出演算法可在複雜環境中，獲的隨車成功率100%與防撞成功率83%之實驗結果。因此，本研究所提出演算法能有助於實踐對各種不同車道與多載體之無人載具防撞系統。

Collision avoidance and guiding direction are major requirements for an automated guided vehicle system. The purpose of this study was to utilize intelligent cars with simple structure to develop the automated guided vehicle system. To achieve the function of automatic collision avoidance, the signal of linear CMOS image sensor was applied to make decision on speed and direction control. In addition, three different types of lane were applied in two and three cars to verify the effect of car-following and collision avoidance. The results showed that our proposed method allows vehicles to obtain 100% success rate in car-following situations in a complicated environment, and 83% success rate in collision avoidance situations. Therefore, our proposed method could help to implement an automated guided vehicle system to various lane and multiple vehicles.