Search

With the development and popularization of smart-phones, human activity recognition methods based on contact perception are proposed. The smartphones which are embedded with various sensors can be used as a platform of mobile sensing for human activity recognition. In this paper, we propose an automated human activity recognition network HDL with smartphone motion sensor units. The HDL network combines DBLSTM (Deep Bidirectional Long Short-Term Memory) model and CNN (Convolutional neural network) model. The DBLSTM model is first used to model long sequence data and ultimately generate a bidirectional output vector in a abstract way. The DBLSTM model is good at dealing with serialization tasks but poor in the ability to extract features. Hence, the CNN model is then used to extract features from the abstract vector. Finally, the output layer employs a softmax function to classify human activities. We conduct experiments on the Public domain UCI dataset. The experimental results show that the proposed HDL network achieves reliable results with accuracy and F1 score as high as 97.95% and 97.27%. Compared with other networks based on the same smartphone dataset, the accuracy of HDL is higher than S-LSTM and Dropout CNN network by 2.14% and 6.97% respectively.

In the smart cities, sharing electric vehicles have many advantages such as environmental protection, low carbon emission, and high efficiency, which will greatly facilitate human life. However, whether the shared electric vehicle can be charged in time directly affects the users' experience. Based on this fact, we study that how to deploy the charging piles in the parking station of the shared electric vehicle, and propose a charging pile deployment model. First, the number of shared electric vehicles that check out from parking stations is predicted based on the random forest algorithm. Then, a mobile model of the shared electric vehicle is established. Based on the mobile model, the number of shared electric vehicles that check in within the target period is calculated and whether they reach the starvation state are determined. After that, the number of charging piles to be deployed is determined. Finally, the effects of hunger rate, battery capacity and the distance between parking stations on the number of charging piles were simulated by experiments. The experimental results verify the validity and feasibility of the proposed model. The prediction model proposed in this paper can provide a certain decision basis for the sharing of electric vehicle charging pile planning.

Identifying binding sites recognized by transcription factors (TFs) is one of major challenges to decipher complex genetic regulatory networks encoded in a genome. A set of binding sites recognized by the same TF, called a motif, can be accurately represented by a position frequency matrix (PFM) or a position-specific scoring matrix (PSSM). Very often, we need to compare motifs when searching for similar motifs in a motif database for a query motif, or clustering motifs possibly recognized by the same TF. In this paper, we have designed a novel metric, called SPIC (Similarity between Positions with Information Contents), for quantifying the similarity between two motifs using their PFMs, PSSMs, and column information contents, and demonstrated that this metric outperforms the other state-of-the-art methods for clustering motifs of the same TF and differentiating motifs of different TFs.

—Security is one of key problems in RFID application. This research is to find a security algorithm for RFID with very low computing requirement. Minimalist Cryptography only needs exclusive Or (XOR) and simple logical control, and it can protect privacy and provide mutual authentication. However, the algorithm has hidden security problems while assuming that the attacker couldn’t monitor the data between RFID readers and tags for continuous sessions. The paper proved this weakness and put forward three ways to improve the weakness, i.e. channel encryption etc. The attacker could only crack the improved algorithm by brute-force attack. The experimental result showed that even by brute-force attack the attack needs more continues sessions after improvement, the cracking time and the record number needed for cracking increased with improved algorithm, and the cracking difficulty increases exponentially when the length of channel key increases, and the encryption complication increases in liner. The improved Minimalist Cryptography provided a safer lightweight security algorithm for RFID.