Obstacle Detection Using Dynamic Particle-Based Occupancy Grids

Due to the complex nature of the driving environment, obstacle tracking systems are required to rely on intermediate dynamic information, before the obstacle is fully reconstructed. This paper presents an obstacle estimation system which uses the advantages of a particle-based occupancy grid tracking solution. The initial measurement data is a raw occupancy map extracted from dense stereovision-derived elevation maps. The occupancy grid tracking system is able to use the raw occupancy data to derive a filtered occupancy probability for each grid cell, along with dynamic information. The process of grouping the grid cells into individual obstacles takes advantage of the dynamic grid information to discriminate between nearby objects that have different speeds, and is able to produce oriented cuboids, having position, size and speed, without the need of an additional tracking step.