Map Building Based on a Xtion Pro Live RGBD and a Laser Sensors

Map Building; senzor fusion; Laser sensors; RGBD; Bayesian method

The main contribution of this paper is to show the feasibility to use the novel Xtion Pro Live RGBD camera into the field of sensor data fusion and map making based on the well established Bayesian method. This approach involves the combination of the Xtion Pro Live RGBD camera with the Hokuyo laser sensor data readings, which are interpreted by a probabilistic heuristic model that abstracts the beam into a ray casting to an occupied grid cell. Occupancy grid is proposed for representing the probability of the occupied and empty areas. In order to update the occupancy grid, the Bayesian estimation method is applied to both sensor data arrays. The sensor data fusion yields a significant improvement of the combined occupancy grid compared to the individual occupied sensor data readings. It is also shown by the Mahalanobis distance that by integrating both sensors, more reliable and accurate maps are produced. The approach has been exemplified by following a sensor data fusion method to building a map of an indoor environment robot.

The main contribution of this paper is to show the feasibility to use the novel Xtion Pro Live RGBD camera into the field of sensor data fusion and map making based on the well established Bayesian method. This approach involves the combination of the Xtion Pro Live RGBD camera with the Hokuyo laser sensor data readings, which are interpreted by a probabilistic heuristic model that abstracts the beam into a ray casting to an occupied grid cell. Occupancy grid is proposed for representing the probability of the occupied and empty areas. In order to update the occupancy grid, the Bayesian estimation method is applied to both sensor data arrays. The sensor data fusion yields a significant improvement of the combined occupancy grid compared to the individual occupied sensor data readings. It is also shown by the Mahalanobis distance that by integrating both sensors, more reliable and accurate maps are produced. The approach has been exemplified by following a sensor data fusion method to building a map of an indoor environment robot.