Erik Schaffernicht, Marco Trincavelli and Achim J. Lilienthal
Bayesian Spatial Event Distribution Grid Maps for Modeling the Spatial Distribution of Gas Detection Events
Sensor Letters, Volume 12(6-7), 2014, pp. 1113 - 1118.
Abstract: In this paper we introduce a novel gas distribution mapping algorithm, Bayesian Spatial Event Distribution (BASED), that, instead of modeling the spatial distribution of a quasi- continuous gas concentration, models the spatial distribution of gas events, for example detection and non-detection of a target gas. The proposed algorithm is based on the Bayesian Inference framework and models the likelihood of events at a certain location with a Bernoulli distribution. In order to avoid overfitting, a Bayesian approach is used with a beta distribution prior for the parameter mu that governs the Bernoulli distribution. In this way, the posterior distribution maintains the same form of the prior, i.e. will be a beta distribution as well, enabling a simple approach for sequential learning. To learn a map composed of beta distributions, we discretize the inspection area into a grid and extrapolate from local measurements using Gaussian kernels. We demonstrate the proposed algorithm for MOX sensors and a photo ionization detector mounted on a mobile robot and show how qualitatively similar maps are obtained from very different gas sensors.
Keywords: gas distribution mapping, statistical modeling, Bernoulli distribution, beta distribution
Download: [PDF (0.4 MB)]
Bibtex:
@ARTICLE{Schaffernicht_etal:SensorLetters:2014,
  AUTHOR = {Schaffernicht, Erik and Trincavelli, Marco and Lilienthal, Achim~J.},
  TITLE = {Bayesian Spatial Event Distribution Grid Maps for Modeling the Spatial Distribution of Gas Detection Events},
  JOURNAL = {Sensor Letters},
  YEAR = {2014},
  VOLUME = {12},
  NUMBER = {6-7},
  PAGES = {1142--1146},
  KEYWORDS = {gas distribution mapping, statistical modeling, Bernoulli distribution, beta distribution},
}