Study On The Active Olfaction Of Robot Based On Ant Colony Algorithms

The objective of the active olfaction research is to make use of the mobile robots and their sensors to trace the plume and declare the odor/gas source actively. The current research achievement indicates that the search strategies used are mainly based on chemical gradient which is called"chemotaxis"or up-steam which is called"anemotaxis", and the search task is mainly realized by single robot or multiple robots. On the basis of the summary of the current research achievement, the following research work has been done:1) The sate-of- the- art of active olfaction is reviewed in detail. The three main steps included in the active olfaction, i.e., plume finding, plume tracking and odor/gas source declaration, and associated strategies applied in each step are summarized. Swarm based active olfaction is also summarized. Two difficulties included in the swarm based active olfaction, i.e., the search strategy and the collaboration of the multi-agent are also pointed.2) A software & hardware platform of a small mobile robot has been designed and accomplished. The concrete modules of the mobile robot are presented in detail, together with lots of schematic diagrams. The embedded operating system uc/OS2 embedded system is used to improve the real-time performance of the mobile robot. The existing problems are explained, and some potential improvements are suggested.3) The ant colony algorithms are introduced to solve the swarm based active olfaction problem. Simulation results have been achieved by FLUENT software. The basic theory of the ant colony algorithms is introduced, based on which several improvements have been made. Simulation results validate its feasibility. The influence of the several concrete parameters to simulation results is analyzed and the problems included are also pointed.4) Swarm based active olfaction is summarized and the future research perspective is analyzed at the end.