Achieving the goal ‘Vision Zero,’ which aims to reduce traffic fatalities to zero by the year 2050, has become prime importance in many countries. Autonomous Driving is at the forefront of many technologies to achieve driverless transportation and make road transportation the safest medium. To accomplish this goal sensors like LiDAR, radar, and cameras are used to detect objects surrounding the ego-vehicle. LiDAR sensor is one of the most prominent sensors with the best achievable accuracy for range. A commonly used method to estimate objects with LiDAR sensors around the ego-vehicle is Oriented Bounding Box estimation. This method lacks the accuracy to detect the edge collision case, which has led to the further development of methods to detect the exact shape of the cars. One of such methods is a three-arc method. This paper investigates the three-arc fit contour prediction algorithm on LiDAR raw data acquired through driving on German roads (KITTI data set) and thus the implementation of the three-arc algorithm in real-world vehicle applications. It was found that the predicted three-arc geometry is in agreement with the vehicle contour derived from top view sketches, and it is feasible to apply this methodology in vehicles.
Investigation of Three-Arc Contour Prediction Algorithm on KITTI LiDAR Raw Data Set
2021-09-01
1698134 byte
Conference paper
Electronic Resource
English
COLOR-BASED ROAD DETECTION AND ITS EVALUATION ON THE KITTI ROAD BENCHMARK
British Library Conference Proceedings | 2014
|British Library Conference Proceedings | 2014
|British Library Conference Proceedings | 2019
|