동적 저궤도 위성 네트워크에서 실시간 라우팅을 위한 FPGA 기반 컨볼루션층 추론 병렬화 기술 = FPGA-based Inference Parallelization of Convolutional Layers for Real-Time Routing in Dynamic LEO Satellite Networks

다국어 초록 (Multilingual Abstract)

This paper addresses the real-time routing problem in Low Earth Orbit (LEO) satellite networks. Existing routing algorithms have been found to struggle to adapt effectively to dynamic satellite network environments. As such, this study proposes a reinforcement learning-based routing approach and implements it using a dueling deep Q-network model. However, the inference process on satellites faces challenges in meeting real-time requirements due to limited computational capabilities. To resolve this, we propose an approach that accelerates inference speed by parallelizing the convolutional layer's inference process. Experimental results show that our proposed method has reduced the computation time of the convolutional layer by 90.2% and the total algorithm execution time by 29.0% compared to the existing methods.

참고문헌 (Reference)

1 L. P. Kaelbling, "Reinforcement Learning : A Survey" 4 : 237-285, 1996

2 J. Wang, "Model Parallelism Optimization for CNN FPGA Accelerator" 16 (16): 1-13, 2023

3 P. Yuan, "Markov decision process-based routing algorithm in hybrid Satellites/UAVs disruption-tolerant sensing networks" 13 (13): 1415-1424, 2019

4 X. Wang, "LEO Satellite Network Routing Algorithm Based on Reinforcement Learning" 1105-1109, 2021

5 L. Gong, "Improving HW/SW Adaptability for Accelerating CNNs on FPGAs Through A Dynamic/Static Co-Reconfiguration Approach" 32 (32): 1854-1865, 2021

6 IEEE, "IEEE Standard for Floating-Point Arithmetic, IEEE Std 754-2019 (Revision of IEEE 754-2008)"

7 L. He, "Hierarchical scheduling for real-time agile satellite task scheduling in a dynamic environment" 63 (63): 897-912, 2019

8 S. Lee, "GPU-Accelerated PD-IPM for Real-Time Model Predictive Control in Integrated Missile Guidance and Control Systems" 22 (22): 1-14, 2022

9 M. Sewak, "Deep Reinforcement Learning" 95-108, 2019

10 Y. -E. Lee, "Cross-Point Based Routing Protocol in Low Earth Orbit Communication Networks" 72-74, 2021

11 M. Qasaimeh, "Comparing Energy Efficiency of CPU, GPU and FPGA Implementations for Vision Kernels" 1-8, 2019

12 Emma, "Characterization of Branch and Data Dependencies in Programs for Evaluating Pipeline Performance" C-36 (C-36): 859-875, 1987

13 AMBA, "Axi4-stream protocol specification, Vol. IHI B 51"

14 S. Lahti, "Are We There Yet? A Study on the State of High-Level Synthesis" 38 (38): 898-911, 2019

15 K. O'Shea, "An introduction to convolutional neural networks, arXiv:1511.08458"

16 P. Zuo, "An Intelligent Routing Algorithm for LEO Satellites Based on Deep Reinforcement Learning" 1-5, 2021

17 B. Plancher, "Accelerating Robot Dynamics Gradients on a CPU, GPU, and FPGA" 6 (6): 2335-2342, 2021

18 D. Kim, "Accelerated Particle Filter with GPU for Real-Time Ballistic Target Tracking" 11 : 12139-12149, 2023

19 N. Hou, "A survey on partitioning models, solution algorithms and algorithm parallelization for hardware/software co-design" 23 : 57-77, 2019

20 B. -S. Roh, "A Study on the Reinforcement Learning Routing for LEO Satellite Network" 537-538, 2022

21 H. Yan, "A Novel Routing Scheme for LEO Satellite Networks Based on Link State Routing" 876-880, 2014