Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - Onboard Double Q-Learning for Airborne Data Capture in Wireless Powered IoT Networks
AU - Li, Kai
AU - Ni, Wei
AU - Wei, Bo
AU - Tovar, Eduardo
PY - 2020/6/30
Y1 - 2020/6/30
N2 - This letter studies the use of Unmanned Aerial Vehicles (UAVs) in Internet-of-Things (IoT) networks, where the UAV with microwave power transfer (MPT) capability is employed to hover over the area of interest, charging IoT nodes remotely and collecting their data. Scheduling MPT and data transmission is critical to reduce the data packet loss resulting from buffer overflows and channel fading. In practice, the prior knowledge of the battery level and data queue length of the IoT nodes is not available at the UAV. A new onboard double Q-learning scheduling algorithm is proposed to optimally select the IoT node to be interrogated for data collection and MPT along the flight trajectory of the UAV, thereby minimizing asymptotically the packet loss of the IoT networks. Simulations confirm the superiority of our algorithm to Q-learning based alternatives in terms of packet loss and learning efficiency/speed.
AB - This letter studies the use of Unmanned Aerial Vehicles (UAVs) in Internet-of-Things (IoT) networks, where the UAV with microwave power transfer (MPT) capability is employed to hover over the area of interest, charging IoT nodes remotely and collecting their data. Scheduling MPT and data transmission is critical to reduce the data packet loss resulting from buffer overflows and channel fading. In practice, the prior knowledge of the battery level and data queue length of the IoT nodes is not available at the UAV. A new onboard double Q-learning scheduling algorithm is proposed to optimally select the IoT node to be interrogated for data collection and MPT along the flight trajectory of the UAV, thereby minimizing asymptotically the packet loss of the IoT networks. Simulations confirm the superiority of our algorithm to Q-learning based alternatives in terms of packet loss and learning efficiency/speed.
KW - Intelligent transportation systems
KW - unmanned vehicles
KW - unmanned aerial vehicles
U2 - 10.1109/LNET.2020.2989130
DO - 10.1109/LNET.2020.2989130
M3 - Journal article
VL - 2
SP - 71
EP - 75
JO - IEEE Networking Letters
JF - IEEE Networking Letters
SN - 2576-3156
IS - 2
ER -