Location utility-based map reduction

Abstract

Abstract-Maps used for navigation often include a database of location descriptions for place recognition (loop closing), which permits bounded-error performance. A standard posegraph SLAM system adds a new entry for every new pose into the location database, which grows linearly and unbounded in time and thus becomes unsustainable. To address this issue, in this paper we propose a new map-reduction approach that pre-constructs a fixed-size place-recognition database amenable to the limited storage and processing resources of the vehicle by exploiting the high-level structure of the environment as well as the vehicle motion. In particular, we introduce the concept of location utility -which encapsulates the visitation probability of a location and its spatial distribution relative to nearby locations in the database -as a measure of the value of potential loop-closure events to occur at that location. While finding the optimal reduced location database is NPhard, we develop an efficient greedy algorithm to sort all the locations in a map based on their relative utility without access to sensor measurements or the vehicle trajectory. This enables pre-determination of a generic, limited-size place-recognition database containing the N best locations in the environment. To validate the proposed approach, we develop an open-source street-map simulator using real city-map data and show that an accurate map (pose-graph) can be attained even when using a place-recognition database with only 1% of the entries of the corresponding full database.

Two-Stage Focused Inference for Resource-Constrained Collision-Free Navigation

“…, such that approximate entropic information over the focused hidden variablesX is maximized, subject to the same constraint as (1): max…”

“…However, the formulation allows for more complex mappings from the unfocused to the focused set. 1 The˜notation is used throughout to refer to the focused set of variables. .…”

“…In the indoor navigation scenario, it is more important to have accurate estimates of landmark locations in tight corridors to avoid collisions. Similarly on the road network, landmarks on more highly traveled roads are more useful [1] and in the underwater scenario it is more important that the robot localize the mines as opposed to the clutter. Consequently, the robot can save resources, such as memory and computational effort, by focusing the mapping operation to more explicitly support the task.…”

Unsupervised trajectory compression

“…For example, it is evident from Figure 3 that the method not only determines a suitable basis for compression but also recovers the structure of the underlying transportation network. Steiner et al [39] demonstrated that knowledge of this underlying network can facilitate map reduction in the context of localization.…”

Graph-based terrain relative navigation with optimal landmark database selection

“…This landmark selection approach extends upon our previous work, in which we presented a method to optimally select locations for a place-recognition database for vehicle localization and pose-graph-based simultaneous localization and mapping (SLAM) [27]. A detailed literature review of the localization database selection problem is provided in that paper, and we here focus specifically on landmark selection.…”