Learning to Rapidly Re-Contact the Lost Plume in Chemical Plume Tracing

Abstract

Maintaining contact between the robot and plume is significant in chemical plume tracing (CPT). In the time immediately following the loss of chemical detection during the process of CPT, Track-Out activities bias the robot heading relative to the upwind direction, expecting to rapidly re-contact the plume. To determine the bias angle used in the Track-Out activity, we propose an online instance-based reinforcement learning method, namely virtual trail following (VTF). In VTF, action-value is generalized from recently stored instances of successful Track-Out activities. We also propose a collaborative VTF (cVTF) method, in which multiple robots store their own instances, and learn from the stored instances, in the same database. The proposed VTF and cVTF methods are compared with biased upwind surge (BUS) method, in which all Track-Out activities utilize an offline optimized universal bias angle, in an indoor environment with three different airflow fields. With respect to our experimental conditions, VTF and cVTF show stronger adaptability to different airflow environments than BUS, and furthermore, cVTF yields higher success rates and time-efficiencies than VTF.

Robust pollution source parameter identification based on the artificial bee colony algorithm using a wireless sensor network

“…A timely perception of the released pollution and identification of its source parameters can save a lot of time for the subsequent elimination of the pollution source. Feasible solutions to this problem include deploying an array of stationary pollution sensor nodes [2] or dispatching mobile robots [3,4] to the concerned area. The deployed sensor nodes can construct a WSN in a self-organized manner, which is capable of rapidly covering a large area.…”

Chemical Source Searching by Controlling a Wheeled Mobile Robot to Follow an Online Planned Route in Outdoor Field Environments

“…Most methods for this subprocedure are biologically inspired, such as the gradient-following algorithms [11,12,13,14], the zigzagging algorithms [2,13], the upwind algorithms [8,15], adapted ant colony optimization algorithm [16], particle swarm optimization algorithm and its variants [17,18,19,20], the SPIRAL algorithm [21], and genetic programming algorithm [22]. During the plume tracing procedure, how to re-contact a lost plume is also an important issue [23]. In the final phase, the robot locates the source [24,25].…”

A Robot Equipped with a High-Speed LSPR Gas Sensor Module for Collecting Spatial Odor Information from On-Ground Invisible Odor Sources

“…Imitating biological olfactory tracking behavior to proactively determine the locations of gas sources using a mobile robot with an artificial olfactory system has been actively researched over the past two decades. − In contrast to a passive gas-sensing system, such an active sensing system can smell, track, and find the source of an odor. In early work, several groups reported robots that can track a gas plume in the air and find odor sources using a semiconductor gas sensor and search algorithms. − In mapping the gas distribution, the spatial resolution is determined by combining the sensor response speed and speed of motion of the gas plume. − Many methods have been applied to improve the performance of an odor robot: a multirobot olfactory system, − a bioinspired robot design, − and a combination with different kind of sensors. − Some of the research focused on the improvement of tracing algorithm. − Most of these robots were designed for odor source localization. The response speed of a semiconductor gas sensor based on a chemical reaction, which is commonly used for these robot systems, is not high (typically longer than 10 s).…”