Search

Abstract There is a trend to deploy neural network on edge devices in recent years. While the mainstream of research often concerns with single edge device processing and edge-cloud two-layer neural network collaborative computing, in this paper, we propose partitioning multi-layer edge network for neural network collaborative computing. With the proposed method, sub-models of neural network are deployed on multi-layer edge devices along the communication path from end users to cloud. Firstly, we propose an optimal path selection method to form a neural network collaborative computing path with lowest communication overhead. Secondly, we establish a time-delay optimization mathematical model to evaluate the effects of different partitioning solutions. To find the optimal partition solution, an ordered elitist genetic algorithm (OEGA) is proposed. The experimental results show that, compared with traditional cloud computing, single-device edge computing and edge-cloud collaborative computing, the proposed multi-layer edge network collaborative computing has a smaller runtime delay with limited bandwidth resources, and because of the pipeline computing characteristics, the proposed method has a better response speed when processing large number of requests. Meanwhile, the OEGA algorithm has better performance than conventional methods, and the optimized partitioning method outperforms other methods like random and evenly partition.

A two-hop wireless regional area network (WRAN) providing monitoring services operating in Television White Space (TVWS), i.e., IEEE P802.22b, may employ a great number of subscriber customer-premises equipments (S-CPEs) possibly without mains power supply, leading to requirement of cost-effective and power-saving design. This paper proposes a framework of cooperative spectrum sensing (CSS) and an energy/bandwidth saving CSS scheme to P802.22b. In each round of sensing, S-CPEs with SNRs lower than a predefined threshold are excluded from reporting sensing results. Numerical results show that the fused missed-detection probability and false alarmprobability could remainmeeting sensing requirements, and the overall fused error probability changes very little. With 10 S-CPEs, it is possible to save more than 40% of the energy/bandwidth on a Rayleigh channel. The principle proposed can apply to other advanced sensing technologies capable of detecting primary signals with low average SNR.