1. Adaptive and Kinodynamic UAV Motion Planning in Complex Dynamic Environments
-
Description: Development of hierarchical motion planning architectures for Unmanned Aerial Vehicles (UAVs) navigating cluttered, high-density, and GNSS-denied environments (e.g., autonomous drone racing and confined industrial spaces). The research focuses on $C^2$-continuous trajectory generation (B-spline optimization minimizing jerk and snap), risk-aware velocity modulation, and high-frequency reactive re-planning to ensure dynamic feasibility and flight safety at extreme aerodynamic limits.
2. UAV-Assisted Wireless Sensor Networks and Cyber-Physical Dynamic Perception (UAV-WSN)
-
Description: Investigation of UAV swarms functioning as aerial base stations and mobile data collectors within heterogeneous sensor networks. Key research topics include probabilistic 3D Line-of-Sight (LoS) channel modeling, Active Wake-up Radio (WUR) triggering mechanisms to drastically extend ground sensor lifetime, and Age of Information (AoI) minimization to guarantee the timeliness of dynamic sensing data.
3. Energy-Efficient Joint Trajectory and Communication Optimization via Multi-Agent Deep Reinforcement Learning (MADRL)
-
Description: Formulation of a unified optimization engine based on the Centralized Training with Decentralized Execution (CTDE) paradigm for multi-UAV cooperative systems. The research addresses the non-convex trade-off between aerodynamic propulsion energy consumption (the dominant power budget component) and wireless network utility, enabling autonomous agents to dynamically co-optimize 3D flight trajectories and transmission scheduling in dynamic, contested environments.
