Position Checking-Based Sampling Approach Combined with Attraction Point Local Optimization for Safe Flight of UAVs

Area of Science:

• Robotics
• Artificial Intelligence
• Aerospace Engineering

Background:

• Efficient motion planning is crucial for unmanned aerial vehicle (UAV) operations.
• Balancing computational resources between path generation and trajectory optimization is a key challenge.
• Existing methods often struggle with complex environments and high computational costs.

Purpose of the Study:

• To propose a sampling-based kinodynamic planning method for UAVs.
• To reduce computational cost and flight risks associated with UAV motion planning.
• To enhance adaptability to complex environments and improve trajectory success rates.

Main Methods:

• An initial trajectory is generated without considering obstacles.
• Spherical spaces are constructed around environmental topological vertices based on trajectory-obstacle intersections.
• A position-checking strategy discards unnecessary sampling points and performs node rewiring.
• Environment-adaptive attraction point selection strategies are designed.

Main Results:

• The proposed method effectively minimizes computational cost.
• It demonstrably reduces the risks associated with UAV flight.
• Simulations and real-world experiments confirm improved performance on a vision-based platform.
• Enhanced success rate in front-end trajectory generation was observed.

Conclusions:

• The developed method offers an efficient and risk-mitigating solution for UAV motion planning.
• It proves adaptable to complex scenarios, ensuring safer UAV flights.
• The strategy optimizes resource allocation, enhancing overall planning efficiency.