Making use of what you don't see: negative information in Markov localization

Abstract

Abstract-This paper explores how the absence of an expected sensor reading can be used to improve Markov localization. This negative information usually is not being used in localization, because it yields less information than positive information (i.e. sensing a landmark), and a sensor often fails to detect a landmark, even if it falls within its sensing range. We address these difficulties by carefully modeling the sensor to avoid false negatives. This can also be thought of as adding an additional sensor that detects the absence of an expected landmark. We show how such modeling is done and how it is integrated into Markov localization. In real world experiments, we demonstrate that a robot is able to localize in positions where otherwise it could not and quantify our findings using the entropy of the particle distribution. Exploiting negative information leads to a greatly improved localization performance and reactivity.

Mapping UHF RFID tags with a mobile robot using a 3D sensor model

“…Negative information has been successfully applied for object tracking [13] and localization [14] of mobile robots. In this paper, we treat non-detections as negative information for our stereo antenna configuration to improve 3D mapping accuracy.…”

Use of negative information in positioning and tracking algorithms

“…Most of the work targeted problems for mobile robot localization [21][22][23]. In Markov localization for mobile robots, the absence of an expected measurement can be used to improve localization.…”

“…One difficulty in implementing a system that uses negative information is that there are two main reasons for the lack of an expected measurement reading: the target may not be there or the sensor may not be able to detect the target. To avoid false negatives, the model needs to consider possible obstructions [21]. Nevertheless, even a false attempt to detect a target can be exploited in tracking applications, based on Bayesian approach to target tracking [22].…”

Cooperative Distributed Sensors for Mobile Robot Localization

“…In robot localization domain, the work of Hoffmann et al [22] on negative information in ML considers negative information the absence of landmark sensor measurements. Occlusions are identified using a visual sensor that scans colors of the ground to determine if there is free area or obstacle.…”