Study On The Realization Method Of Active Olfaction Based On Multi-Agent

The goal of the research work about mobile robot odor localization (MROL, also called active olfaction in this paper) is to detect and finally declare the source of relevant odor/gas in the circumstance based on robot and its sensor systems. There are lots of potential applications for the active olfaction, for example, dangerous environment exploring, disaster rescue as well as looking for the leakage sources of toxic and harmful gases.This thesis focuses on the realization method of active olfaction. The relatively detailed work can be concluded as follows.1. The state-of-the-art of active olfaction is summarized, especially the methods based on multi-agent. The difficulties and emphases are analyzed and the new methods are found.2. Dynamic air environment model which simulates the real environment more accurately are established using FLUENT. The current dynamic air environment model is analyzed and modified.3. The Improved Ant Colony Algorithm and the Evolution Gradient Algorithm are introduced to solve the MROL problem. The simulation results are given and the shortage and advantage are analyzed.4. The realization of active olfaction is summarized and the foreground is analyzed.