Abstract
Moving-object tracking (estimating position and velocity of moving objects) is one of the key technologies for autonomous driving systems and driving assistance systems in mobile robotics and vehicle automation domains. To predict the possibility of a future collision, the tracking system has to recognize what objects are as accurately as possible. This paper presents a method for recognizing vehicles (cars and bicyclists) and people using highly-resolved multilayer laser scanner (MLS). Laser-scan data captured by MLS are clustered, and eight features are extracted from each of clustered laser-scan data, such as distance from MLS, velocity, object’s size, number of laser-measurement points, and distribution of the reflection intensities. A multiclass support vector machine is applied to classify cars, bicyclists, and people from the features. The experimental result using data set of “The Stanford Track Collection” showed that the success rate of people and vehicles classification by the proposed method was higher than 95%.
