Non-Intrusive Load Identification Model Based on 3D Spatial Feature and Convolutional Neural Network

Load identification method is one of the major technical difficulties of non-intrusive composite monitoring. Binary V-I trajectory image can reflect the original V-I trajectory characteristics to a large extent, so it is widely used in load identification. However, using single binary V-I trajectory feature for load identification has certain limitations. In order to improve the accuracy of load identification, the power feature is added on the basis of the binary V-I trajectory feature in this paper. We change the initial binary V-I trajectory into a new 3D feature by mapping the power feature to the third dimension. In order to reduce the impact of imbalance samples on load identification, the SVM SMOTE algorithm is used to balance the samples. Based on the deep learning method, the convolutional neural network model is used to extract the newly produced 3D feature to achieve load identification in this paper. The results indicate the new 3D feature has better observability and the proposed model has higher identification performance compared with other classification models on the public data set PLAID.

TCL: a taxi trajectory prediction model combining time and space features

Vehicle trajectory modeling is an important foundation for urban intelligent services. Trajectory prediction of cars is a hot topic. A model including convolutional neural network(CNN) and long short-term memory(LSTM) was proposed, which is named trajectory-CNN-LSTM(TCL). CNN can extract the spatial features of the trajectory in the input image. Besides, LSTM can extract the time-series features of the input trajectory. After that, the model uses fully connected layers to merge the two features for the final predicting. The experiments on the Porto dataset of The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases(ECML-PKDD) show that the average prediction error of TCL is reduced by 0.15 km, 0.42 km, and 0.39 km compared to the trajectory-convolution(T-CONV), multi-layer perceptron(MLP), and recurrent neural network(RNN) model, respectively.