Gas Source Tracing with a Mobile Robot Using an Adapted Moth Strategy

Abstract

Abstract. As a sub-task of the general gas source localisation problem, gas source tracing is supposed to guide a gas-sensitive mobile system towards a source by using the cues determined from the gas distribution sensed along a driven path. This paper reports on an investigation of a biologically inspired gas source tracing strategy. Similar to the behaviour of the silkworm moth Bombyx mori, the implemented behaviour consists of a fixed motion pattern that realises a local search, and a mechanism that (re-)starts this motion pattern if an increased gas concentration is sensed. While the moth uses the local airflow direction to orient the motion pattern, this is not possible for a mobile robot due to the detection limits of currently available anemometers. Thus, an alternative method was implemented that uses an asymmetric motion pattern, which is biased towards the side where higher gas sensor readings were obtained. The adaptated strategy was implemented and tested on an experimental platform. This paper describes the strategy and evaluates its performance in terms of the ability to drive the robot towards a gas source and to keep it within close proximity of the source.

Simulation Experiments with Bio-inspired Algorithms for Odor Source Localization in Laminar Wind Flow

“…In recent work [14] [13], we have shown through experiments with real robots that the surge-spiral [6] [7] [2] [4] and the surge-cast [13] algorithms are faster and more reliable than pure casting [10] [9] [20] [12] [11] [1] in laminar wind flow. This result was insofar surprising, as the casting algorithm got much more attention by the research community up to date.…”

Mapping multiple gas/odor sources in an uncontrolled indoor environment using a Bayesian occupancy grid mapping based method

“…In this setting, some researchers have proposed methods that either not rely or only partially rely on wind measurements [6], [10], [16], [17].…”

Theoretical analysis of three bio-inspired plume tracking algorithms

“…In recent work, we compared the surge-spiral algorithm [7] [8] [9] [10], the surge-cast algorithm and pure casting [11] [12] [13] [14] [15] [16] [17] in both simulation [20] and with real robots [21] [22] in laminar wind flow. The results of both studies revealed that upwind surge algorithms are more efficient than pure casting.…”