車輛防撞器為智慧型車路系統的重要輔助裝置。在本論文中選用偵測性能佳、設計技術簡單且成本較低的頻率調變連續波雷達(FMCW radar)為偵測器,使用鋸齒波與三角波交替調變的頻率調變波形以避免接收訊號的混淆現象,解決多目標模糊的問題。然後,透過FFT訊號處理即可得知偵測器所量測目標之相對速度及相對距離,以供後級防撞控制處理器做進一步之分析計算,求得適當的座車加減速度。 於座車側邊加裝雷達,即可獲得即時的鄰車道車輛相對速度與相對距離值(和座車),以做為是否變換車道的依據。在本論文中所要探討的現象是,高速公路上跟車控制、變換車道行為。在防撞控制處理器方面,採用可適性模糊網路(ANFIS、FBFN)對前級訊號處理模組所獲得之資料來進行車輛防撞控制模擬,並產生一組參考信號藉由遞迴最小平方法則(RLS)來調整可適性模糊網路中之後項參數。最後,並和其他相關之研究(Hwu_變換車道模式)作一比較。經由模擬結果可以發現我們的方法確實可使車輛的行駛更符合安全性、舒適性、及合理性之三大要求。
The car collision avoidance detector is an important auxiliary equipment of the IntelligentVehicleHighwaySystem(IVHS). The Frequency Modulation Continuous Wave (FMCW) radar is used as a sensor. It has the advantages of small size, light weight and very low probability of interception. The combined triangular and sawtooth frequency-modulated waveforms are designed to resolve ambiguity among multiple targets. The Fast Fourier Transform (FFT) is used to implement the Doppler processor which measures the beat frequencies and computes the relative speed and distance between the host car and the leading car or cars on the adjacent lane. The control processor is based on the measured relative speed and distance to estimate a proper acceleration or deceleration for the host car. The behaviors of car-following and lane changing on the freeway are discussed. The performance of the adaptive fuzzy network (ANFIS、FBFN) car collision avoidance controller is simulated. The simulation results show that the proposed control processor can provide a safe and comfortable car driving.