Human Pose Estimation (HPE) is a core component in real-time decision systems, supporting critical applications such as healthcare monitoring, autonomous driving, and sports analytics. While deep learning models—particularly CNNs and Transformer-based architectures—have significantly improved HPE accuracy, they remain vulnerable to adversarial perturbations that subtly distort keypoint localization, thereby undermining system reliability. To address this challenge, we propose regularization and pruning transferable adversarial attack (RPA), a novel framework designed to enhance the transferability of adversarial samples in Transformer-based HPE models. RPA integrates two synergistic strategies: gradient regularization, which suppresses dominant feature correlations to reduce overfitting, and adaptive weight pruning, which removes redundant parameters to reduce model-specific noise. This dual mechanism enables the generation of transferable adversarial attacks that are effective across diverse model architectures. Extensive experiments on state-of-the-art HPE networks demonstrate that RPA consistently outperforms existing attack methods. In white-box settings, RPA reduces average precision (AP) by 0.05-0.30; in black-box scenarios, it yields AP drops of 0.01-0.04. These findings expose critical vulnerabilities in IoT-enabled HPE applications and establish a new benchmark for evaluating adversarial robustness in real-time perception systems.

Summary With the popularity of mobile smart devices and functional diversification, the infection ways of mobile smartphones worm also become diverse. In mobile networks, mobile devices are suffering from the threat of worms all the time. Once the worm outbreaks, it will bring huge losses to mobile phone users and mobile network operators. Researches on mobile worm propagation model allow us to control and detect potential worm threat, according to the characteristics of worm's outbreak. In the paper, we put forward a worm propagation model based on the mobile network environment. After analyzing the model, we then give the simulation for controlling factors affecting worm propagation. This model allows us to have a certain understanding for the spread on the size and speed of the mobile worm and provide effective methods to control the spread of the mobile worm. Copyright © 2015 John Wiley & Sons, Ltd.

Machine‐to‐machine (M2M) networks often serve mission‐critical applications in our daily life. Maintaining nodes connectivity is crucial for M2M networks to accomplish their tasks effectively. The fault of a critical node may break the network into several blocks, and thus hinders the operation. Rapid recovery is highly required in such a case. In this study, the authors focus on designing an effective algorithm for restoring damaged network topologies by moving available nodes. Unlike contemporary algorithms that only consider network connectivity or coverage, the proposed algorithm not only takes into account network connectivity, coverage and energy consumption, but also concerns about the shortest path between any pair of nodes. Simulation results demonstrate that the proposed algorithm outperforms the existing scheme in terms of network lifetime.

Smartphone is broadly applicable to the human activity recognition HAR mobile devices. However, energy consumption becomes a big obstacle to such mobile devices of real-time monitoring. In order to solve this problem, this paper presents a method of activity recognition based on energy-efficient schemes. In terms of data acquisition and processing, energy-efficient schemes adopt the best sample rate and extract the most effective feature combinations in accordance with the different activities, so as not to increase energy consumption; while in terms of recognition algorithm, we adopt the improved structure of multi-class support vector machine, combine it with the probability of activity occurrence, so as to reduce the time complexity of recognition. This method can minimize energy consumption greatly under the premise of maintaining higher recognition accuracy. Moreover, this paper adopts mobile cloud security technology to reduce potential risk of the smartphone's data transmission and processing. We present our work with an experimental study, and our experiments show that the accuracy of activity recognition based on energy-efficient schemes we proposed is up to 90.6%. In addition, this method will save 51.0% energy than that when sample rate and extracted features are, respectively, fixed at 100Hz and combined features. Copyright © 2016 John Wiley & Sons, Ltd.

This research focuses on two aspects of Intelligent Transportation System (ITS): Intelligent Traffic Light and Dynamic Route Guidance (DRG). The paper aims to make traffic light and route guidance to be smarter. In this paper, the authors apply Grey-Markov Model which combines Grey Model and Markov Model together to predict short-time traffic and then build Intelligent Traffic Light Model (ITLM). For purpose of realizing DRG, the authors improve ant colony optimization (ACO) by putting forward a new feedback pheromone and changing the probabilistic formula, which would make ACO feasible for solving the DRG in reality transportation. Simulations show that the model do have a better performance on short-time traffic predicting and improved ACO is suitable for DRG.

SUMMARY This paper presents two routing games, a unicast and a multicast routing game, for wireless sensor networks. We analyze the actions of nodes inside/outside lowest cost path (LCP) and draw their payoff functions. Our simulation shows that this evolutionary game theory scheme has several advantages over a widely used collusion–resistant routing scheme. All nodes, either out of LCP or in LCP, will ultimately choose the strategy ‘transmit’. To improve the payoff, nodes in LCP should either minimize their actual forwarding cost or maximize their claimed cost as long as it remains in the LCP, whereas minimizing the actual claimed cost is not an optimal option for nodes out of LCP. Copyright © 2013 John Wiley & Sons, Ltd.

In the m-out-of-n oblivious transfer (OT) model, one party Alice sends n bits to another party Bob, Bob can get only m bits from the n bits. However, Alice cannot know which m bits Bob received. Y. Mu and Naor presented classical m-out-of-n oblivious transfer based on discrete logarithm. As the work of Shor, the discrete logarithm can be solved in polynomial time by quantum computers, so such OTs are unsecure to the quantum computer. In this paper, we construct a quantum m-out-of-n OT (QOT) scheme based on the transmission of polarized light and show that the scheme is robust to general attacks, i.e. the QOT scheme satisfies statistical correctness and statistical privacy.