- 學位論文
應用視覺資訊與Sugeno模糊推論規劃 自走車避障的運動軌跡
Obstacle Avoidance of Autonomous Land Vehicle Using Visual Information and Sugeno Fuzzy Inference for Trajectory Analysis
摘要
在無人自動導航車的研究中，閃避障礙物是一項重要的議題；該議題除了需要考慮障礙物在不同的相對距離和方向…等各項環境資訊，自動導航車需要考量自身的車體的各項限制因素。本論文考慮動態的前輪轉向角度以及車身寬度等參數，對初步產生的可能行進方向命令透過模糊理論之方法，改善車輛之運動控制命令，以獲得更為穩定且平滑的行進軌跡。至於本系統的輸入的環境資訊，是來自兩種圖型識別系統，分別是以單眼視覺和立體視覺兩種影像為基礎而產生的辨識結果。這些結果能表達出路面與障礙物等多種空間上的資訊，本論文再藉由Sugeno模糊模型之方法，根據多種不同的複雜環境資訊，對應建立出合適的模糊規則，全面性地考慮適合的行進方向，不但提高了系統的自適應性，也減少了運算上的複雜度，以獲得更為即時的計算結果。
並列摘要
In this paper, a fuzzy navigating system only based on image for obstacle avoidance similar to the capability of human vision is proposed. The system applies Sugeno type fuzzy model. to easily map many and various recognized information onto fuzzy rules for obstacle avoidance of the autonomous land vehicle (ALV) navigating. During the ALV navigating with the stereo vision camera, the fuzzy navigating system adopts the insufficient information (such as imprecise view angle and rough depth, etc.) to evaluate the best feasible steering direction. Even through one or more obstacles lie on a road surface and the road are separated into several side roads, the ALV still can avoid these obstacles and infer the best steering direction toward the wider side road.