생산 현장에서 가상화 리소스 할당을 반영한 효율적인 AIoT 리소스 관리 기법 = Efficient AIoT Resource Management Techniques Reflecting Virtualization Resource Allocation in Production Environments

다국어 초록 (Multilingual Abstract)

As the convergence of artificial intelligence and the Internet of Things accelerates, many companies today use AIoT extensively in their industrial sites to increase competitiveness and efficiency. In industrial areas, computing-intensive and resource-intensive network edges are efficient, but resource capacity and power budgets require management methods that can increase low latency and energy efficiency. This paper proposes a virtualized resource allocation-based AIoT resource management technique that can simultaneously optimize the model performance(accuracy and robustness) and resource cost of AIoT resource management. The proposed method uses Lyapunov optimization theory to minimize bottlenecks in edge and cloud resources to perform continuous learning at a low cost-efficiently. In addition, the proposed technique continuously updates weights in the constant learning model of AIoT resources to minimize processing delays and network overhead of resources through virtualization resource allocation.

참고문헌 (Reference)

1 이재선, "중소기업 간 협력철학이 협력활동과 협력성과에 미치는 영향에 관한 연구" 한국융합학회 8 (8): 301-309, 2017

2 F. Forooghifar, "Resource-aware distributed epilepsy monitoring using self-awareness from edge to cloud" 13 (13): 1338-1350, 2019

3 L. Zhao, "Optimal edge resource allocation in IoT-based smart cities" 33 (33): 30-35, 2019

4 B. Yang, "Offloading Optimization in Edge Computing for Deep-Learning-Enabled Target Tracking by Internet of UAVs" 8 (8): 9878-9893, 2021

5 Y. Kang, "Neurosurgeon : Collaborative intelligence between the cloud and mobile edge" 45 (45): 615-629, 2017

6 T. X. Tran, "Joint task offloading and resource allocation for multi-server mobile-edge computing networks" 68 (68): 856-868, 2019

7 C. Garrido-Hidalgo, "Iot heterogeneous mesh network deployment for human-in-the-loop challenges towards a social and sustainable industry 4. 0" 6 : 28417-28437, 2018

8 D. Liu, "FitCNN : A cloud-assisted and low-cost framework for updating CNNs on IoT devices" 91 : 277-289, 2019

9 T. D. Nguyen, "DÏoT : A federated self-learning anomaly detection system for IoT" 756-767, 2019

10 J. Mills, "Communication-Efficient Federated Learning for Wireless Edge Intelligence in IoT" 7 (7): 5986-5994, 2020

11 C. Gamanayake, "Cluster Pruning : An efficient filter pruning method for edge AI vision applications" 14 (14): 802-816, 2020

12 S. Wang, "Adaptive Federated Learning in Resource Constrained Edge Computing Systems" 37 (37): 1205-1221, 2019

13 Z. Zhao, "A Novel Framework of Three-Hierarchical Offloading Optimization for MEC in Industrial IoT Networks" 16 (16): 5424-5434, 2020

14 H. Gu, "A Collaborative and Sustainable Edge-Cloud Architecture for Object Tracking with Convolutional Siamese Networks" 6 (6): 144-154, 2021