The Internet of Vehicles (IoVs) makes communications between numerous devices that use various protocols susceptible to hacker incursions and attacks, which can compromise privacy and seriously jeopardize driving safety. Many studies have been proposed to detect intrusions hitherto, but two major limitations remain. First, traditional Vehicles-to-Cloud (V2C) have difficulty in figuring out the decentralized distribution of data and computational power in IoVs. Second, the majority of studies suffer from unbalanced data in which the attacks only make up a small part and fail to detect low-probability attacks. To address these limitations, we design a Federated Learning-Edge Cloud (FL-EC) communication architecture for IoVs with a Feature Select Transformer (FSFormer) for effective intrusion detection: In FL-EC, mobile users collect and encrypt data before uploading it to edges for training, with edges and cloud functioning as clients and servers in FL, ensuring privacy and efficient data transmission. In FSFormer, we propose a Feature Attention mechanism to search and promote significant features. Furthermore, the Feed-Forward Network is replaced with a Routing module for a deeper but less-parameter network. Extensive experiments show that our model effectively boosts the detection rate of low-probability attacks and outperforms five baseline models in almost all scenarios.

Fringe projection profilometry (FPP) using a digital video projector is widely used for three-dimensional shape measurement. However, the gamma nonlinearity, system vibration, and noise cause the captured fringe patterns to be nonsinusoidal waveforms and have a grayscale deflection from their true value. This leads to an additional phase measurement error for a general phase-shifting algorithm. Based on the theoretical analysis, we propose a method to eliminate the phase error considering two factors. In this method, four-step phase-shifting is done four times with an initial phase offset of 22.5 deg and the average of these four phase maps precisely results in the real phase. As a result, phase error caused by gamma nonlinearity can be effectively suppressed. In addition, every image in phase shifting is replaced by the average of 20 fringe images continuously captured at the same state to avoid the phase error caused by system vibration and noise. Experimental results show that this method is effective in eliminating the phase error in practical phase-shifting FPP. In general, more than 90% of the phase error can be reduced.

Obtaining the distribution of home and work locations is essential for city planning, as it defines the structure and mobility pattern of a city. With the development of telecommunication networks, mobile network data, having the advantages of large coverage and strong followability, have produced large amounts of information about human activities. Thus, it has become a popular research subject for human position detection. In this study, we proposed a new method to detect home and work locations based on the extraction of focal points in traces, identifying an individual’s working and resting hours, and analyzing the characteristics of city grids using mobile phone cellular signaling data (CSD). At the individual level, we validated the algorithm on ground-truth volunteer data and achieved a small deviation of under 500 and 565 m for home and work location detection 85% of the time. At the aggregate level, we tested it on a city-wide anonymized CSD set and found a high Pearson correlation between our result and the census data of 0.93. Compared to existing studies, this study improved the granularity and location accuracy of home and work location detection, as well as validated the method using both individually labeled ground-truth data and aggregate data for the first time. Applying the algorithm in a city, we captured the population distribution, commuting patterns, and job-housing balance of the city and demonstrated the potential in using mobile network data for urban planning and policy formulation.