By running applications and services closer to the user, edge processing provides many advantages, such as short response time and reduced network traffic. Deep-learning based algorithms provide significantly better performances than traditional algorithms in many fields but demand more resources, such as higher computational power and more memory. Hence, designing deep learning algorithms that are more suitable for resource-constrained mobile devices is vital. In this paper, we build a lightweight neural network, termed LiteNet which uses a deep learning algorithm design to diagnose arrhythmias, as an example to show how we design deep learning schemes for resource-constrained mobile devices. Compare to other deep learning models with an equivalent accuracy, LiteNet has several advantages. It requires less memory, incurs lower computational cost, and is more feasible for deployment on resource-constrained mobile devices. It can be trained faster than other neural network algorithms and requires less communication across different processing units during distributed training. It uses filters of heterogeneous size in a convolutional layer, which contributes to the generation of various feature maps. The algorithm was tested using the MIT-BIH electrocardiogram (ECG) arrhythmia database; the results showed that LiteNet outperforms comparable schemes in diagnosing arrhythmias, and in its feasibility for use at the mobile devices.

No abstract

The protection of intellectual property in the process of resource-based cities’ transformation and upgrading is a worldwide problem; China’s large-scale resource-based cities’ transformation process is facing this problem. The research object of this paper is to protect the intellectual property in the process of Daqing City’s transformation and upgrading. This paper inspects the protection of intellectual property in Daqing City in the process of transformation and upgrading and puts forward the specific measures to improve the protection of intellectual property based on the analysis of the present situation of intellectual property protection in Daqing City, as well as combined with corresponding research in the process of cities’ transformation and upgrading.

No abstract

This thesis presents an algorithm to automatically select the positions for friction stir spot welding (FSSW) in a laminated rapid tooling process. The work combines a twodimensional structural analysis with tool path planning to realize the overall process planning for the rapid tooling of a plastic injection mold. The work starts from a two-dimensional cantilever beam model, defining the effective distance of a single spot joint strength, and also considers the effect of a single layer thickness. Secondly, an efficient medial axis transformation algorithm, which is suitable for the general two-dimensional boundary curves, has been proposed to generate the adaptive equidistance offsetting curves. In addition, through different working conditions of the internal and external spot welds, an adaptive discretization method is presented. Then, a selection principle for choosing the initial spot weld location and processing order with optimization to avoid redundancy is presented. Finally, the authors compare the advantages of this novel algorithm and traditional path planning algorithms with respect to strength and processing efficiency while taking into account structural strength.