基于HSIC的烘烤场景多源异构传感器优化配置策略OA

[Objective]This research aims to effectively explore the non-uniform distribution of tempera-ture and humidity inside the dense oven and realize the accurate perception of the baking scene,to solve the prob-lem of multi-source data redundancy caused by excessive sensor points.[Method]A mobile intelligent dense oven was used as the research object to obtain a total of 861×24 discrete time series data of 24 temperature and relative humidity sensors during the baking cycle.With HSIC as the basis,the Bayesian data fusion principle was introduced.Based on the principle of information maximization,information gain rate,and configuration redun-dancy constraints,the priority ranking and quantity selection basis for sensor configuration were clarified,thereby establishing an optimal configuration strategy for multi-source heterogeneous sensors.[Result]There was a sig-nificant non-uniformity in the distribution of temperature and relative humidity field in the bulk curing barn,and there was a strong non-linear correlation among different regions.Based on the sensor optimization configuration strategy,a total of 12 sensor objects,namely S1,S5,S4,S6,S7,S8,S11,S16,S13,S14,S21 and S24,were se-lected under the condition of reducing the sensors by 50%.Compared with the original data,the RMSE after site optimization was 0.663 8,and the information gain rate(gr)of temperature and relative humidity were 0.018 2 and 0.001 8,respectively.[Conclusion]The sensor configuration strategy developed in consideration of the non-uniform distribution of temperature and relative humidity in the bulk curing barn can effectively ensure the per-ception accuracy and information abundance of temperature and relative humidity,and can provide theoretical ba-sis and technical reference for intelligent monitoring and control of the curing environment,the development of curing process for tobacco leaves of varying qualities,and the optimized design of curing barn structures.