Stable Mapping Using a Hyper Particle Filter

Often Particle Filters are used to solve the SLAM (Simultaneous Localization and Mapping) problem in robotics: The particles represent the possible poses of the robot, and their weight is determined by checking if the sensor readings are consistent with the so far acquired map. Mostly a single map is maintained during the exploration, and only with Rao-Blackwellized Particle Filters each particle carries its own map. In this contribution, we propose a Hyper Particle Filter (HPF) – a Particle Filter of Particle Filters – for solving the SLAM problem in unstructured environments. Each particle of the HPF contains a standard Particle Filter (with a map and a set particles, that model the belief of the robot pose in this particular map). To measure the weight of a particle in the HPF, we developed two map quality measures that can be calculated automatically and do not rely on a ground truth map: The first map quality measure determines the contrast of the occupancy map. If the map has a high contrast, it is likely that the pose of the robot was always determined correctly before the map was updated, which finally leads to an overall consistent map. The second map quality measure determines the distribution of the orientation of wall pixels calculated by the Sobel operator. Using the model of a rectangular overall structure, slight but systematic errors in the map can be detected. Using the two measures, broken maps can automatically be detected. The corresponding particle is then more likely to be replaced by a particle with a better map within the HPF. We implemented the approach on our robot “Robbie 12”, which will be used in the RoboCup Rescue league in 2009. We tested the HPF using the log files from last years RoboCup Rescue autonomy final, and with new data of a larger building. The quality of the generated maps outperformed our last years (league’s best) maps. With the data acquired in the larger structure, Robbie was able to close loops in the map. Due to a highly efficient implementation, the algorithm still runs online during the autonomous exploration.