Research And Application Of Data Fusion Algorithm Based On Deep Learning

Data fusion technology combines information from multiple sensors to achieve higher accuracy and more specific inferences than can be achieved by using a single sensor alone,so it is widely used in the filed of military,medical,human-computer interaction,target tracking and so on.As a classic artificial intelligence algorithm,neural network has a good effect in dealing with non-linear problems.Convolutional neural network,as one of the representative algorithms of deep learning,has the advantages of strong feature representation and parallelism.Therefore,the algorithms based on convolutional neural network are discussed and applied to the eye tracking data fusion in this paper.The main work of this paper is as follows:1.A time series data fusion algorithm based on convolutional neural network is proposed.In order to solve the problems of deep convolutional neural network(DCNN)’s insufficient dependence on long-distance information and low initial input feature robustness of the network,this paper proposes a combined denoising autoencoder(DAE),DCNN and self_attention mechanism Improved data fusion algorithm(denoted as IDCNN algorithm).firstly,the algorithm uses a denoising autoencoder to denoise and reconstruct the input signal,secondly adds an self_attention mechanism module to the DCNN network,and finally compares the IDCNN algorithm with the DCNN and BPNN algorithms on two datasets in the UCI database.Experimental results show that the fusion performance of the IDCNN algorithm is more outstanding.2.An eye movement and tracking data fusion algorithm based on deep learning is proposed.For traditional data fusion algorithms,the fusion effect of eye movement and tracking data in multiple scenes is poor.This paper proposes a new eye movement and tracking data fusion algorithm based on deep learning,namely Eye-CNN-BLSTM algorithm.firstly,the algorithm adds new artificial features based on the spatial position information of the original eye movement and tracking data;secondly,it combines a convolutional neural network and bidirectional long-short-term memory network to design a new fusion structure;experimental results show that it is compared with six classic data fusion algorithms,the proposed algorithm has better fusion performance on the OTB-100 dataset.Overall,the improved algorithms proposed in this paper have achieved good results,but the algorithm needs to be optimized in terms of network parameter settings and operating efficiency,which is further research work.