One of the most widespread and important applications in wireless sensor networks (WSNs) is the continuous data collection, such as monitoring the variety of ambient temperature and humidity. Due to the sensor nodes with a limited energy supply, the reduction of energy consumed in the continuous observation of physical phenomenon plays a significant role in extending the lifetime of WSNs. However, the high redundancy of sensing data leads to great waste of energy as a result of over-deployed sensor nodes. In this paper, we develop a structure fidelity data collection (SFDC) framework leveraging the spatial correlations between nodes to reduce the number of the active sensor nodes while maintaining the low structural distortion of the collected data. A structural distortion based on the image quality assessment approach is used to perform the nodes work/sleep scheduling, such that the number of the working nodes is reduced while the remainder of nodes can be put into the low-power sleep mode during the sampling period. The main contribution of SFDC is to provide a unique perspective on how to maintain the data fidelity in term of structural similarity in the continuous sensing applications for WSNs. The simulation results based on synthetic and real world datasets verify the effectiveness of SFDC framework both on energy saving and data fidelity.

In a noisy manufacturing environment, range-based target localization for wireless sensor networks (WSNs) experiences various variations in range measurements, thus causing the unsatisfactory results. Starting from an empirical observation on the existing various industrial noise distribution, in this paper, a diffusion Gauss-Newton (GN) algorithm with cooperation strategy is proposed for solving target localization Non-Linear Least Squares (NLLS) problem in a WSN. The proposed algorithm has an equalization effect on the unbalance noise distribution over the network by aggregating the global estimates into local GN update via diffusion strategy. When facing a hostile industrial environment where the ambient noise is heterogeneous or has the sudden changes across partial nodes, the significant performance degradation is produced by diffusion GN. To solve the problem, we propose further an improved version of diffusion GN, which is adaptive to sudden changes on noisy range measurements. Instead of using a static combiner, the new algorithm leverages the evolutionary game theory (EGT) to assign a time-varying weight for the estimate from each neighboring node based on the individual range error. Consequently, the good estimates from the neighbors with high SNR have a larger weight in the combiner than the bad estimates caused by the low SNR or high noise. Some simulation examples show that the standard diffusion GN is effective in stationary noise and its improved version provides the adaptation to changing noisy environment.

With the rise of the processing power of networked agents in the last decade, second-order methods for machine learning have received increasing attention. To solve the distributed optimization problems over multiagent systems, Newton’s method has the benefits of fast convergence and high estimation accuracy. In this article, we propose a reinforced network Newton method with $K$ -order control flexibility (RNN-K) in a distributed manner by integrating the consensus strategy and the latest knowledge across the network into local descent direction. The key component of our method is to make the best of intermediate results from the local neighborhood to learn global knowledge, not just for the consensus effect like most existing works, including the gradient descent and Newton methods as well as their refinements. Such a reinforcement enables revitalizing the traditional iterative consensus strategy to accelerate the descent of the Newton direction. The biggest difficulty to design the approximated Newton descent in distributed settings is addressed by using a special Taylor expansion that follows the matrix splitting technique. Based on the truncation on the Taylor series, our method also presents a tradeoff effect between estimation accuracy and computation/communication cost, which provides the control flexibility as a practical consideration. We derive theoretically the sufficient conditions for the convergence of the proposed RNN-K method of at least a linear rate. The simulation results illustrate the performance effectiveness by being applied to three types of distributed optimization problems that arise frequently in machine-learning scenarios.

Due to the negative impact from spatial correlation, spatially correlated cognitive radio (CR) based devices participating in cooperative spectrum sensing may be harmful to the detection performance. In this paper, we propose an energy-efficient cooperative spectrum sensing scheme based on spatial correlation for cognitive Internet of Things (CIoT). To mitigate the communication overhead and ensure sufficient sensing accuracy, the CR-based devices (CRDs) can be grouped into several clusters. The member nodes undertake cooperative spectrum sensing tasks in turn by rotating, and send the local test statistic to their cluster head nearby. Then, by exploiting the spatial correlation of the members, the cluster head combines the sensing results and make use of likelihood ratio test to obtain the cluster decision. After receiving the decisions from all clusters, the fusion center employs hard scheme to make the final decision about spectrum occupancy. The simulation results show that our scheme not only provides the better sensing performance, but also improve the energy efficiency.

Automated facial gender and age classification has remained a challenge because of the high inter-subject and intra-subject variations. We addressed this challenging problem by studying multi-instance- and multi-scale-enhanced multi-task random forest architecture. Different from the conventional single facial attribute recognition method, we designed effective multi-task architecture to learn gender and age simultaneously and used the dependency between gender and age to improve its recognition accuracy. In the study, we found that face gender has a great influence on face age grouping; thus, we proposed a random forest face age grouping method based on face gender conditions. Specifically, we first extracted robust multi-instance and multi-scale features to reduce the influence of various intra-subject distortion types, such as low image resolution, illumination and occlusion, etc. Furthermore, we used a random forest classifier to recognize facial gender. Finally, a gender conditional random forest was proposed for age grouping to address inter-subject variations. Experiments were conducted by using two popular MORPH-II and Adience datasets. The experimental results showed that the gender and age recognition rates in our method can reach 99.6% and 96.14% in the MORPH-II database and 93.48% and 63.72% in the Adience database, reaching the state-of-the-art level.