TY - GEN
T1 - A Middleware for Digital Twin-Enabled Flying Network Simulations Using UBSim and UB-ANC
AU - Moorthy, Sabarish Krishna
AU - Harindranath, Ankush
AU - McManus, Maxwell
AU - Guan, Zhangyu
AU - Mastronarde, Nicholas
AU - Bentley, Elizabeth Serena
AU - Medley, Michael
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Data-driven control based on AI/ML techniques has a great potential to enable zero-touch automated modeling, optimization and control of complex wireless systems. However, it is challenging to collect network traces in the real world because of high time and labor cost, weather limitations as well as safety concerns. In this work we attempt to tackle this challenge by designing a multi-fidelity simulator taking wireless Unmanned Aerial Vehicle (UAV) networks into consideration. We design the simulator by interfacing two Unmanned Aerial System (UAS) simulators we have developed in prior years: UBSim and UB-ANC. The former focuses on UAV network optimization and policy training by considering explicitly the network environments such as blockage dynamics, while the latter focuses more on high-fidelity UAV flight control. We first develop a coordination interface referred to as SimSocket for signaling exchanges between UBSim and UB-ANC in simulations, and then showcase coordinated simulations based on UBSim and UB-ANC. The new research that can be enabled by the integrated simulator is also discussed for digital twin-based UAS systems.
AB - Data-driven control based on AI/ML techniques has a great potential to enable zero-touch automated modeling, optimization and control of complex wireless systems. However, it is challenging to collect network traces in the real world because of high time and labor cost, weather limitations as well as safety concerns. In this work we attempt to tackle this challenge by designing a multi-fidelity simulator taking wireless Unmanned Aerial Vehicle (UAV) networks into consideration. We design the simulator by interfacing two Unmanned Aerial System (UAS) simulators we have developed in prior years: UBSim and UB-ANC. The former focuses on UAV network optimization and policy training by considering explicitly the network environments such as blockage dynamics, while the latter focuses more on high-fidelity UAV flight control. We first develop a coordination interface referred to as SimSocket for signaling exchanges between UBSim and UB-ANC in simulations, and then showcase coordinated simulations based on UBSim and UB-ANC. The new research that can be enabled by the integrated simulator is also discussed for digital twin-based UAS systems.
KW - Data-Driven Control
KW - Multi-Fidelity Simulation
KW - UB-ANC
KW - UBSim
KW - Unmanned Aerial Vehicle (UAV)
UR - https://www.scopus.com/pages/publications/85139438927
U2 - 10.1109/DCOSS54816.2022.00059
DO - 10.1109/DCOSS54816.2022.00059
M3 - Conference contribution
AN - SCOPUS:85139438927
T3 - Proceedings - 18th Annual International Conference on Distributed Computing in Sensor Systems, DCOSS 2022
SP - 322
EP - 327
BT - Proceedings - 18th Annual International Conference on Distributed Computing in Sensor Systems, DCOSS 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 18th Annual International Conference on Distributed Computing in Sensor Systems, DCOSS 2022
Y2 - 30 May 2022 through 1 June 2022
ER -