Information Fusion Applied In Localization Of Gas Leakage Source

The localization of gas leakage is using devices or other ways to localize the situation of harmful gas leakage,in order to avoid widespread pollution damage by long-time leakage and minimize the finicial losses.A number of researchers have presented some strategies and methods for the research of gas source searching,which improved the development of the study of leakage localization.As the intelligentialization of mobile robot is going deep,the task of seeking of gas source will be finished by device instead of by person.In this paper,the idea of Multi-source Information Fusion(MSIF)which based on many former researches is used in the task of gas leakage searching,thereby,the information,such as situation and rate,of leakage can be fully estimated through multi-approach and multi-level ways in searching process.All works are unfolded as the following.Firstly,detailedly introduce some common gas diffusion models which existed in many articles by researchers.But some models are too complex to be used in the leakage finding,so,these models are compared in their theories and usable range,and then used to simulate gas distribution in Matlab in this paper.Secondly,in order to select the most suitable gas diffusion model from many ones,this article proposes a new way which integrated with Fuzzy Control.Fuzzy has a reasoning system and can realize the transfer from Expert Knowledge to digital system,then get the best result from limited evidence by sensors.A simulation experience in use of Fuzzy Control toolbox of Matlab is showed finally and the result testifies the feasibility of this method.Thirdly,a developing technology which called Multi-source Information Fusion is placed.MSIF includes a lot of data-processing algorithms summarized by authors and is beneficial to fuse the data of sensors,lead to make a best decision to judge the gas leakage souce.This paper recommends some frequently-used algorithms and gives corresponding example to make verification.Finally,Multi-source information fusion algorithms are used in the searching of gas leakage in this paper.Different from traditional ways,gas sensors and vision sensor are mixed to get information,single gas sensor measurement is replaced by the multiple gas sensors,and the measurement method is changed from single-point to multipoint.In the application of algorithms,the Weighted Average Method which was used to fuse the data of gas sensors can reduce the influence of the noises and other troubles,the Least Square Method is applied to estimate the location and flow of the leakage,the Bayesian Inference can accumulate different information source to judge the leakage leads to the robot can search the target reasonable.