HomePublicationsInvited Talk

Towards Intelligent Cruise Control and Data Aggregation in UAVs-aided Sensor Networks
Ref: CISTER-TR-210902       Publication Date: 2021

Abstract:
I was invited to give this online presentation for Research Seminar Series in Department of Aeronautical and Aviation Engineering, The Hong Kong Polytechnic University, Hong Kong, China.
In UAVs-aided sensor networks, the flexibility of the UAV and favourable line-of-sight connection between the UAV and ground sensors are exploited to improve data reception. However, the ground nodes can suffer from buffer overflows, because of poor cruise control of the UAV and lossy airborne channels when the UAV is far away from the ground nodes. In this talk, we discuss a new deep reinforcement learning based onboard control framework, where the cruise control of the UAV and its schedule of data aggregation can be learned online. Moreover, deep deterministic policy gradient (DDPG) can be adopted onboard at the UAV to deliver effective continuous control operations and communication decisions, using its substantially outdated knowledge of the ground nodes, network states and channel conditions.

Kai Li