To reduce the increasingly congestion in cities, it is essential for intelligent transportation system (ITS) to accurately forecast the short-term traffic flow to identify the potential congestion sites. In recent years, the emerging deep learning method has been introduced to design traffic flow predictors, such as recurrent neural network (RNN) and long short-term memory (LSTM), which has demonstrated its promising results. In this paper, different from existing work, we study the temporal convolutional network (TCN) and propose a deep learning framework based on TCN model for short-term city-wide traffic forecast to accurately capture the temporal and spatial evolution of traffic flow. Moreover, we design the model with the Taguchi method to develop an optimized structure of the TCN model, which not only reduces the number of experiments, but also yields high accuracy of forecasting results. With the real-world traffic flow data collected from highways in Birmingham City of U.K., we compare our model with four deep learning based models including LSTM models, GRU models, SAE models, DeepTrend and CNN-LSTM models in terms of the mean absolute error (MAE) and mean relative error (MRE) regarding the actual flow data. The experimental results demonstrate that our framework achieves the state-of-art performance with superior accuracy in short-term traffic flow forecasting.

Clustering organizes nodes into groups in order to enhance the connectivity and stability of cognitive radio sensor networks. Mainly depending on the channel availability, many existing spectrum-aware clustering algorithms may not achieve the most satisfactory clustering. Taking into account the various influence factors to establish the optimal clustering is a challenge to enhance the network performance. This paper proposes a novel spectrum-aware clustering algorithm based on weighted clustering metric to obtain the optimal clustering by solving an optimization model. The new weighted clustering metric, simultaneously evaluating temporal-spatial correlation, confidence level and residual energy, is used to elect clusterheads and ally member nodes. After clustering, the clusterheads sensing spectrum instead of all member nodes greatly reduces the energy consumption of spectrum sensing and increases the opportunity of data transmission. The performance comparison between the traditional spectrum-aware clustering algorithms and our proposed algorithm has been highlighted with the experiments.

This paper investigates the direction identification of micro-cracks with nonlinear components generated by Lamb wave with frequency-mixing technique. Three-dimensional finite element simulations were carried out to investigate the interaction mechanism between Lamb wave signals and micro-cracks. Upon re-visiting the conventional Lamb wave excitation signal with two kinds of fundamental frequencies (f₁ and f₂), it was found to be possible to generate new types of frequencies (f₁ ± f₂) at the sideband if nonlinear sources existed in the plate. A pulse inversion method was used to extract the sideband frequency for nonlinear ultrasonic detection. By arranging piezoelectric chip arrays around the micro-crack, the acoustic nonlinearity parameter β related to the fundamental frequency and the sideband frequency for different micro-crack directions was calibrated. It was shown that β varied for different crack directions, which provides useful information about the scattering features of the nonlinear Lamb wave interacting with the micro-crack to characterize its directivity. Moreover, the scattering degree defined with the relative nonlinear parameter β′ of the micro-crack in different directions was investigated in detail by changing the size of the micro-crack. The outcomes showed that the forward scattering signal of the crack had a greater amplitude, whereas the backscattering signal had a smaller amplitude compared with the scattering signals in other directions from micro-cracks. In addition, the signal scattering degree in the forward direction from micro-cracks increased with the increasing micro-crack length, but decreased with increasing crack width. Furthermore, for the buried crack, the forward scattering degree of Lamb wave from micro-crack decreased as crack was buried deeper in plate. In summary, the findings of this study can help to further advance the use of nonlinear Lamb wave with the frequency-mixing technique for identifying the direction of micro-cracks.

Text-based person retrieval is the process of searching a massive visual resource library for images of a particular pedestrian, based on a textual query. Existing approaches often suffer from a problem of color (CLR) over-reliance , which can result in a suboptimal person retrieval performance by distracting the model from other important visual cues such as texture and structure information. To handle this problem, we propose a novel framework to E xcavate A ll-round I nformation B eyond C olor for the task of text-based person retrieval, which is therefore termed EAIBC. The EAIBC architecture includes four branches, namely an RGB branch, a grayscale (GRS) branch, a high-frequency (HFQ) branch, and a CLR branch. Furthermore, we introduce a mutual learning (ML) mechanism to facilitate communication and learning among the branches, enabling them to take full advantage of all-round information in an effective and balanced manner. We evaluate the proposed method on three benchmark datasets, including CUHK-PEDES, ICFG-PEDES, and RSTPReid. The experimental results demonstrate that EAIBC significantly outperforms existing methods and achieves state-of-the-art (SOTA) performance in supervised, weakly supervised, and cross-domain settings.