In recent years, artificial intelligence has played an increasingly important role in the field of automated control of drones. After AlphaGo used Intensive Learning to defeat the World Go Championship, intensive learning gained widespread attention. However, most of the existing reinforcement learning is applied in games with only two or three moving directions. This paper proves that deep reinforcement learning can be successfully applied to an ancient puzzle game Nokia Snake after further processing. A game with four directions of movement. Through deep intensive learning and training, the Snake (or self-learning Snake) learns to find the target path autonomously, and the average score on the Snake Game exceeds the average score on human level. This kind of Snake algorithm that can find the target path autonomously has broad prospects in the industrial field, such as: UAV oil and gas field inspection, Use drones to search for and rescue injured people after a complex disaster. As we all know, post-disaster relief requires careful staffing and material dispatch. There are many factors that need to be considered in the artificial planning of disaster relief. Therefore, we want to design a drone that can search and rescue personnel and dispatch materials. Current drones are quite mature in terms of automation control, but current drones require manual control. Therefore, the Snake algorithm proposed here to be able to find the target path autonomously is an attempt and key technology in the design of autonomous search and rescue personnel and material dispatching drones.

Wireless sensor networks (WSNs) are widely used in military, traffic, medical and so on. The design of routing protocol for WSNs is limited by the single nature of the local topology information. Meanwhile, the power supply of sensor networks node, communication ability and storage capacity are limited, so how to improve the efficient energy of nodes and extend the networks life cycle is the focus of current research. This study proposes the improved algorithm for the LEACH (Low Energy Adaptive Clustering Hierarchy) clustering algorithm, considering the residual energy of the nodes and the factors of the long distance node, the T ( n ) is readjusted and the new method is proposed. Then the data fusion rate is introduced to allow the cluster‐heads to fuse data before sending the data, and send the data to the base station. Finally, the free‐space model and the multi‐path fading model are adopted to avoid the excessive consumption of energy caused by the node d ⁴ . The authors’ simulation results show that the improved algorithm can reduce the energy consumption of the networks and prolongs life cycle.

This paper proposes a new deep learning method, the greedy deep weighted dictionary learning for mobile multimedia for medical diseases analysis. Based on the traditional dictionary learning methods, which neglects the relationship between the sample and the dictionary atom, we propose the weighted mechanism to connect the sample with the dictionary atom in this paper. Meanwhile, the traditional dictionary learning method is prone to cause over-fitting for patient classification of the limited training data set. Therefore, this paper adopts l ₂ -norm regularization constraint, which realizes the limitation of the model space, and enhances the generalization ability of the model and avoids over-fitting to some extent. Compared with the previous shallow dictionary learning, this paper proposed the greedy deep dictionary learning. We adopt the thinking of layer by layer training to increase the hidden layer, so that the local information between the layer and the layer can be trained to maintain their own characteristics, reduce the risk of overfitting and make sure that each layer of the network is convergent, which improves the accuracy of training and learning. With the development of Internet of Things and the soundness of healthcare monitoring system, the method proposed have better reliability in the field of mobile multimedia for healthcare. The results show that the learning method has a good effect on the classification of mobile multimedia for medical diseases, and the accuracy, sensitivity, and specificity of the classification have good performance, which may provide guidance for the diagnosis of disease in wisdom medical.

Strip steel surface defects occur frequently during the manufacturing process, and these defects cause hidden risks in the use of subsequent strip products. Therefore, it is crucial to classify the strip steel’s surface defects accurately and efficiently. Most classification models of strip steel surface defects are generally based on convolutional neural networks (CNNs). However, CNNs, with local receptive fields, do not have admirable global representation ability, resulting in poor classification performance. To this end, we proposed a hybrid network architecture (CNN-T), which merges CNN and Transformer encoder. The CNN-T network has both strong inductive biases (e.g., translation invariance, locality) and global modeling capability. Specifically, CNN first extracts low-level and local features from the images. The Transformer encoder then globally models these features, extracting abstract and high-level semantic information and finally sending them to the multilayer perceptron classifier for classification. Extensive experiments show that the classification performance of CNN-T outperforms pure Transformer networks and CNNs (e.g., GoogLeNet, MobileNet v2, ResNet18) on the NEU-CLS dataset (training ratio is 80%) with a 0.28–2.23% improvement in classification accuracy, with fewer parameters (0.45 M) and floating-point operations (0.12 G).

The continuous growth of interconnected objects in the Internet of things (IoT) raises a challenge to the wireless communication technology. Cognitive radio could make full use of the dynamic spectrum access and spectrum diversity over wide spectrum to alleviate the spectrum scarcity problem and satisfy the enormous connectivity demands in IoT, which has garnered significant attention over the last few years. This paper addresses the spectrum allocation problem with respect to both spectrum utilization and network throughput in the cognitiveradio- based IoT. On the one side, each link in a transmission path intends to improve the transmission performance on the assigned spectrum channel to maximize the end-to-end throughput. On the other side, these links share the same spectrum channel to concurrently transmit as much as possible to achieve the maximum spectrum utilization. In order to solve the problem, we propose a concurrent transmission model in the network which reveals the constraints of mutual interference and resource competition in links concurrent transmissions. Based on this model, we formulate the spectrum allocation plan for links as the chromosome (solution) in genetic algorithms. Then, we apply the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) to solve the multi-objective spectrum allocation problem. Simulation results validate that the proposed strategy can search the optimal solutions efficiently and satisfy the requirements of spectrum allocation in various cases.

Color information plays an important role in the color image segmentation and real-time color sensor, which affects the result of video image segmentation and correct real-time temperature value. In this paper, a novel real-time color image segmentation method is proposed, which is based on color similarity in RGB color space. According to the color and luminance information in RGB color space, the dominant color is determined at first, and then color similarity can be calculated with the proposed calculation method of color component, which creates a color-class map. Next, the information of the corresponding color-class map is utilized to classify the pixels. Due to the characteristic that thermal inks feature color values that change in real time as the temperature changes, the segmentation results of thermal ink can be used as a real-time color sensor. Then, we also propose a method of color correction and light source compensation for the sake of potential inaccuracy of its measures. We discuss the proposed segmentation method application combining with color sensor (thermal ink) in real-time color image segmentation for Cyber physical system (CPS) by the application in fire detection and summarize a new method in identifying fire in a video based on these characteristics. The experiments showed that the proposed method in vision-based fire detection and identification in videos was effective; the results were accurate and can be used in real-time analysis.