Hallucinating robots: Inferring Obstacle Distances from Partial Laser Measurements
This addresses a specific issue in mobile robot navigation where partial sensor data leads to inaccurate obstacle detection, though it is incremental as it builds on existing learning approaches.
The paper tackles the problem of robots incorrectly estimating distances to obstacles like glass panels using 2D laser scanners by proposing a method to infer obstacle distances directly from raw 2D laser data, achieving successful real-time demonstration on a Care-O-bot 4 robot.
Many mobile robots rely on 2D laser scanners for localization, mapping, and navigation. However, those sensors are unable to correctly provide distance to obstacles such as glass panels and tables whose actual occupancy is invisible at the height the sensor is measuring. In this work, instead of estimating the distance to obstacles from richer sensor readings such as 3D lasers or RGBD sensors, we present a method to estimate the distance directly from raw 2D laser data. To learn a mapping from raw 2D laser distances to obstacle distances we frame the problem as a learning task and train a neural network formed as an autoencoder. A novel configuration of network hyperparameters is proposed for the task at hand and is quantitatively validated on a test set. Finally, we qualitatively demonstrate in real time on a Care-O-bot 4 that the trained network can successfully infer obstacle distances from partial 2D laser readings.