Study On The Method Of Active Odor Source Localization Based On Fruit Fly Algorithm

In the face of frequent chemical gas leaks,locating leak sources quickly and accurately is very important for rescue work.Locating odor source with robot effectively can minimize the loss of life and property.Therefore,more and more attention have been paid for the active odor source localization of robot,but most researches were limited to simple 2D experiment environments.However,the scenes where leakage accidents occurred are usually complex and changeable,and the robot should complete the task of locating odor source with high accuracy at the early stage of gas leakage.For the gas diffusion and odor source location in complex environment,the method and strategy of active odor source localization in both 2D and 3D environment are studied.The main works of this thesis are summarized as follows:Firstly,the analysis of the plume model is a prerequisite for developing an effective odor source localization strategy.Both 2D and 3D plume model of indoor ventilation environment are constructed.The obstacles and changing wind speed are considered in plume model with FLUENT.Secondly,the fruit fly optimization algorithm(FOA)is used to locate the odor source in 2D indoor ventilation environment.An odor source localization strategy based on Fixed-step Fruit Fly Optimization Algorithm(FS-FOA)is proposed.The feasibility and effectiveness of the strategy based on FS-FOA for locating odor source in unknown environments are verified in different conditions of the simulation environment.Thirdly,for improving the low efficiency as fixed step value in FS-FOA,a Concentration-Adaptive-Step Fruit Fly Optimization Algorithm(CAS-FOA)is proposed and used in 2D different complex indoor ventilation simulation environments.Compared with other mehods,the CAS-FOA proposed in this thesis has higher success rate and higher accuracy.Finally,to be closer to actual application,the strategy of odor source locating based on CAS-FOA in 3D environment is proposed.The simulation results in 3D environment show that the proposed strategy is feasible and valid.