The Auto-Complete Graph: Merging and Mutual Correction of Sensor and Prior Maps for SLAM

Abstract

Simultaneous Localization And Mapping (SLAM) usually assumes the robot starts without knowledge of the environment. While prior information, such as emergency maps or layout maps, is often available, integration is not trivial since such maps are often out of date and have uncertainty in local scale. Integration of prior map information is further complicated by sensor noise, drift in the measurements, and incorrect scan registrations in the sensor map. We present the Auto-Complete Graph (ACG), a graph-based SLAM method merging elements of sensor and prior maps into one consistent representation. After optimizing the ACG, the sensor map’s errors are corrected thanks to the prior map, while the sensor map corrects the local scale inaccuracies in the prior map. We provide three datasets with associated prior maps: two recorded in campus environments, and one from a fireman training facility. Our method handled up to 40% of noise in odometry, was robust to varying levels of details between the prior and the sensor map, and could correct local scale errors of the prior. In field tests with ACG, users indicated points of interest directly on the prior before exploration. We did not record failures in reaching them.

Robot exploration of indoor environments using incomplete and inaccurate prior knowledge

“…A recent work [22] uses prior maps taken from emergency maps to build graph-based SLAM representations. The method, which extends that of [23], merges the information from sensors and a prior map into a representation of the environment and corrects errors both in sensor readings and the prior map.…”

“…Our focus is on exploration performance and not on the quality of the produced map (as in [22]), but the relationships between the two approaches could be further investigated.…”

“…Note that, although prior knowledge PK E is known, the map M E for safe navigation of a robot in E should be built and an exploration is still required (as also discussed in [22]).…”

URSIM: Unique Regions for Sketch Map Interpretation and Matching

“…In another scenario, sketch maps can be used to enhance existing prior maps, e.g., emergency maps. Working with Dortmund's firemen, we identified that they have access to emergency maps that they use to direct their operations [39]. However, while firemen are proficient with emergency maps, other people may not be, especially in situations of stress.…”

“…Furthermore, we previously developed the Auto-Complete Graph (ACG) method [39] to integrate an emergency map into the robot's Simultaneous Localization and Mapping (SLAM). By first matching a sketch map to an emergency map using URSIM and then integrating the emergency map in the SLAM using the ACG method, the information of the sketch map could be included in the robot's map during exploration and used to navigate to points of interest, both by sketching a trajectory or autonomous navigation.…”