1. Mobility-Aware Coded Distributed Computing

In an airborne network consisting of multiple mobile heterogeneous UAVs with diverse computing capabilities, to enable NAC, how should a UAV partition and allocate a computation task to other UAVs is one of the most critical problems to address. Considering the high UAV mobility and the presence of various uncertainties, a desired distributed computing scheme should not only address node heterogeneity, but also be robust to uncertainties. Furthermore, for computation tasks that are critical for the safety of UAV operations, the distributed computing scheme is also desired to generate results in a timely manner. This study aims to develop an innovative distributed computing scheme that possesses these desired characteristics. Specific results developed so far include 1) a batch-processing based coded computation (BPCC) scheme for matrix multiplications over static UAV networks, 2) a multi-agent reinforcement learning (MARL) based coded computation scheme for matrix multiplications over dynamic UAV networks, 3) a coded computation scheme for vector convolutions over dynamic UAV networks, 4) a coded federated learning framework for distributed machine learning over static UAV networks, and 5) a coded distributed learning framework for accelerating the training speed of MARL algorithms.


Reference

  1. B. Zhou, J. Xie, B. Wang, "Dynamic Coded Distributed Convolution for UAV-based Networked Airborne Computing", in Proceedings of International Conference on Unmanned Aircraft Systems (ICUAS), 2022.
  2. B. Wang, J. Xie, K. Lu, Y. Wan, S. Fu, “On Batch-Processing Based Coded Computing for Heterogeneous Distributed Computing Systems”, IEEE Transactions on Network Science and Engineering, Vol. 8, No. 3, pp. 2438-2454, 2021.
  3. D. Wang, B. Wang, J. Zhang, K. Lu, J. Xie, S. Fu, Y. Wan, "CFL-HC: A Coded Federated Learning Framework for Heterogeneous Computing Scenarios", accepted by IEEE GLOBECOM Symposium on Cloud Computing, Networking and Storage, Dec. 2021.
  4. B. Wang, J. Xie, K. Lu, Y. Wan, S. Fu, "Multi-Agent Reinforcement Learning based Coded Computation for Mobile Ad Hoc Computing", in Proceedings of IEEE International Conference on Communications (ICC), June. 2021.
  5. B. Wang, J. Xie, N. Atanasov, "Coding for Distributed Multi-Agent Reinforcement Learning", in Proceedings of IEEE International Conference on Robotics and Automation (ICRA), Xi'an, China, May. 2021.
  6. B. Wang, J. Xie, K. Lu, Y. Wan, S. Fu, “Coding for Heterogeneous UAV-based Networked Airborne Computing”, in Proceedings of the IEEE GLOBECOM Workshop on Computing-Centric Drone Networks, Waikoloa, HI, Dec. 2019.

2. Stochastic Mobility Control to Facilitate Robust Computing under Uncertainty

Uncertain aerial mobility and fast-changing network topology are major challenges for NAC. This study aims to address these challenges by leveraging the high maneuverability of UAVs to turn mobility into a facilitator for computing. So far, we have developed a realistic 3-D smooth-turn random mobility model for aerial vehicles to serve as the design and evaluation foundation, a scalable and efficient uncertainty evaluation method called M-PCM-OFFD to facilitate robust NAC, and efficient stochastic optimal control algorithms.


Reference

  1. M. Liu, Y. Wan, Z. Lin, F. L. Lewis, J. Xie, and Brian A. Jalaian, "Computational Intelligence in Uncertainty Quantification for Learning Control and Differential Games", Handbook of Reinforcement Learning and Control, edited by V. Kyriakos, Y. Wan, F. L. Lewis, and Derya H. Cansever, Vol. 325. Springer, Cham, 2021.
  2. J. Xie, Y. Wan, Y. Zhou, K. Mills, J. J. Filliben, and Y. Lei, Effective Uncertainty Evaluation in Large-Scale Systems (book chapter), Principles of Cyber-Physical Systems: An Interdisciplinary Approach, edited by Sandip Roy and Sajal K. Das, Cambridge University Press, pp. 25-50, 2020.
  3. J. Xie, Y. Wan, K. Mills, J. J. Filliben, Y. Lei, and Z. Lin, "M-PCM-OFFD: An Effective Output Statistics Estimation Method for Systems of High Dimensional Uncertainties Subject to Low-Order Parameter Interactions," Mathematics and Computers in Simulation, Vol. 159, pp. 93-118, May 2019.
  4. J. Xie, Y. Wan, K. Mills, J. J. Filliben, and F. Lewis, “A Scalable Sampling Method to High-dimensional Uncertainties for Optimal and Reinforcement Learning-based Controls,” IEEE Control Systems Letters, vol. 1, no. 1, pp. 98-103, July 2017.
  5. J. Xie, Y. Wan, Y. Zhou, K. Mills, J.J. Filliben, and Y. Lei, “Effective and Scalable Uncertainty Evaluation for Large-Scale Complex System Applications”, in Proceedings of Winter Simulation Conference (WSC), Savannah, GA, Dec. 2014.

3. Adaptive Distributed Computing, Mobility Control and Networking Co-Design

In traditional distributed computing schemes, computation loads are usually distributed in a centralized fashion, which may cause communication traffic jam on the master node and degrade computing performance if network bandwidth is low. Such centralized structure also limits the range of computing resources the master node can utilize. This study aims to first tackle these issues by exploring practical networking strategies. We then develop a co-design framework that integrates the distributed computing scheme with mobility control and networking.

4. Testbed Development and Applications

To evaluate and improve the proposed theoretical framework, we build both software and hardware testbeds. We also explore its real applications to, e.g., multi-UAV coordinated navigation, real-time surveillance, and mobile edge computing.


Reference

  1. J. Zhang, B. Wang*, D. Wang, S. Fu, K. Lu, Y. Wan, J. Xie, "An SDR-based LTE System for Unmanned Aerial System", in Proceedings of International Conference on Unmanned Aircraft Systems (ICUAS), 2022
  2. J. Chen, J. Xie, "Joint Task Scheduling, Routing, and Charging for Multi-UAV Based Mobile Edge Computing", in Proceedings of IEEE International Conference on Communications (ICC), 2022
  3. C. Douma, B. Wang, J. Xie, "Coded Distributed Path Planning for Unmanned Aerial Vehicles", in Proceedings of AIAA Aviation, Jun. 2021
  4. B. Wang, J. Xie, N. Atanasov, "Coding for Distributed Multi-Agent Reinforcement Learning", in Proceedings of IEEE International Conference on Robotics and Automation (ICRA), Jun. 2021
  5. B. Wang, J. Xie, S. Li, Y. Wan, Y. Gu, S. Fu, K. Lu, “Computing in the Air: An Open Airborne Computing Platform”, IET Communications, Vol. 14, No. 15, pp. 2410-2419, 2020.
  6. S. Li, C. He, M. Liu, Y. Wan, Y. Gu, J. Xie, S. Fu, and K. Lu, “The Design and Implementation of Aerial Communication Using Directional Antennas: Learning Control in Unknown Communication Environment,” IET Control Theory and Application, vol. 13, no. 17, pp. 2906-2916, Nov. 2019.
  7. J. Chen, J. Xie, Y. Gu, S. Li, S. Fu, Y. Wan, K. Lu, “Long-Range and Broadband Aerial Networking using Directional Antenna (ANDA): Design and Implementation”, IEEE Transactions on Vehicular Technology, Vol. 66, No. 12, 2017.
  8. J. Xie, F. AI-Enrani, Y. Gu, Y. Wan and S. Fu, “UAV-Carried Long-distance Wi-Fi Communication Infrastructure”, in Proceedings of AIAA SciTech Forum, San Diego, CA, Jan. 2016.