MURM-A*: An Improved A* Within Comprehensive Path-Planning Scheme for Cellular-Connected Multi-UAVs Based on Radio...

Area of Science:

• Robotics and Autonomous Systems
• Network Engineering
• Aerospace Engineering

Background:

• Cellular-connected Unmanned Aerial Vehicles (UAVs) require persistent network connectivity and flight safety in dense urban airspace.
• Existing path-planning methods struggle with environmental data processing, obstacle avoidance, flight dynamics, and multi-UAV conflict resolution.
• Traditional A* algorithm limitations hinder simultaneous optimization of path efficiency and radio quality for UAVs.

Purpose of the Study:

• To develop a comprehensive path-planning scheme for multiple cellular-connected UAVs.
• To address limitations in existing research regarding environmental map data processing and multi-constraint path planning.
• To improve flight safety and network connectivity for UAVs in complex urban radio environments.

Main Methods:

• Constructed a path-planning model using complex-network theory, environmental data, and radio maps.
• Developed an improved A* algorithm (MURM-A*) for multi-UAV scenarios.
• Separated environmental representation from algorithmic search for enhanced processing.

Main Results:

• The MURM-A* algorithm effectively avoids obstacles and prevents spatial conflicts between UAV paths.
• Achieved joint optimization of path efficiency and radio quality, reducing radio-outage time compared to Deep Reinforcement Learning (DRL).
• The path-planning model improved environmental information identification and reduced modeling time compared to DRL.

Conclusions:

• The proposed systematic framework provides reliable path planning for multiple cellular-connected UAVs in complex radio environments.
• MURM-A* offers a valid and efficient alternative to traditional A* and DRL methods for UAV path planning.
• The study establishes a well-structured and extensible framework for future UAV path-planning research.