CPEV自主避障控制策略研究

Abstract

本研究主要目標在於建立一套適用於工研院自行開發之智慧四輪電動車CPEV，可於單線同向雙車道上多障礙物的情境進行避障之自主避障控制策略。除了避障策略以外，本研究亦整合車輛穩定策略增加其穩定性，使其能在時速50公里以下擁有主動循跡轉向及主動煞停避障功能。 本研究除了建立避障控制所需之情境判斷與操作決策流程以及路徑循跡控制方法外，並針對車輛與兩車道障礙物間距離以及相對速度，調整控制參數，考量乘坐舒適度，規劃多條安全轉向避障路徑。另外亦整合車輛穩定策略包含防滑煞車控制、翻覆偵測警示與車輛轉向動態控制，以提升車輛於循跡避障過程之穩定性。 本研究首先利用MATLAB/Simulink建立控制策略模型，搭配以CarSim軟體建立之車輛動態模型進行模型迴路模擬( MIL )驗證與策略參數調整，並透過實車測試數據調整車輛模型動態特性。最後，透過自主避障策略使CPEV能夠在時速50公里以內，針對左右車道不同速度下的移動障礙車，順利以煞停或轉向避障方式閃避障礙車，並能夠針對不同情境修改轉向路徑。

The purpose of this research is to establish an active safety control strategy integrating obstacle avoidance and vehicle stability which fits the situation being on one-way two-lane and with multiple obstacles, and can be used on an intelligent electric vehicle developed by ITRI. We develop a car equipped with active steering, lane tracking and active braking control to dodge obstacles on the road or to stop the vehicle under 50km/h. This strategy not only consists of obstacle avoidance control including situation judgment flowchart with operational decision-making process and path following control method, but also can adjust the parametric path planning algorithm to plan several paths according to the distance and relative velocity between the vehicle and obstacle, and passenger’s comfort. And it integrates the vehicle stability strategy including anti-slip braking control, rollover detection warning and vehicle steering dynamic stability control in order to enhance the steadiness through the path tracking process. We establish the control strategy on MATLAB/Simulink and the vehicle dynamic model on CarSim to execute the Model-in-the-Loop simulation and parameter adjustment. And then we modify the dynamic characteristic of the vehicle model through experiment. Last, this research proposes a control strategy that can make CPEV within 50kph successfully avoid moving obstacles at one-way two-lane through active braking or active steering.