Technology On Vehicle Safety Driving Assist Based On Laser Measurement Scanner

The highway traffic, especially the freeway traffic has been developed rapidly since the eighties of the twentieth century. While traffic accident shows a tendency to increase, especially the rear-end collision accident which is the most representational accident form. Based on this premise, vehicle safety driving assist systems such as the rear-end collision avoidance warning of the collision and obstacle avoidance are studied in this paper with AS-R mobile robot platform. To finish the systems, two of the most important technologies are localization and obstacle avoidance. This paper is focus on localization and obstacle avoidance based on laser measurement scanner.Firstly, measuring range principle of laser measurement scanner and several factors which are closely-related to measure range are represented; then the paper analyzes the localization by odometer, but the far the robot moves, the more the locating error grows by this method.Secondly, in order to modify the error created by original odometer, the polar scan matching (PSM) is adopted in this paper. Polar Scan Matching (PSM) works in the laser scanner's polar coordinate system, therefore taking advantage of the structure of the laser measurements. Laser range measurements of current and reference scans are associated with each other using the matching bearing rule, which eliminates the search for corresponding points and makes translation estimation of the PSM approach O(n) complexity. The global consistent environment map has been built based on pose estimation in this thesis.Thirdly, emergency braking and obstacle avoidance are represented. We use laser range finder to detect the distance between the obstacle front of the robot and the robot, then based on the distance we decide whether give an alarm and to take emergency stop or slowdown strategy. A method of improved vector field histogram is proposed to resolve obstacle avoidance problem in the paper. Firstly, the algorithm decomposes the environment into a series of grids and taking into account the sensing uncertainties and the robot size, expends the grid which is detected as an obstacle point on the created grid map ;Secondly, creates primary polar histogram according to the robot and obstacle distribution relation, generate the free walking areas and obstacle avoidance areas, then creates the binary polar histogram by two threshold ,then taking account into the dynamics and the kinematics of the robot, creates the masked polar histogram; lastly, based on the masked polar histogram and a cost function selects the most suitable direction.Finally, we experimentalism with AS-R mobile robot and the experiment's results indicate that the proposed localization and obstacle avoidance are flexible and reliable.At the end of the dissertation, conclusions and outlooks are given.