HAARN: A Deep Neural Network-Based Intelligent Control Method for High-Altitude Adaptability of Heavy-Load UAV Power...

Area of Science:

• Aerospace Engineering
• Artificial Intelligence
• Control Systems

Background:

• High-altitude operation of heavy-load unmanned aerial vehicles (UAVs) for tasks like ultra-high voltage line construction faces challenges due to varying air density and temperature.
• Traditional control methods struggle with adaptability in dynamic high-altitude environments, leading to suboptimal performance.

Purpose of the Study:

• To develop an intelligent control method for heavy-load UAV power systems to enhance high-altitude adaptability.
• To improve the efficiency and stability of UAV power systems operating at high altitudes.

Main Methods:

• A deep neural network, the High-Altitude Adaptive Regulation Network (HAARN), was developed to learn complex nonlinear relationships.
• Real-time environmental data (altitude, temperature, air pressure) were collected using barometric altimeters and GPS receivers.
• The HAARN model was trained on a dataset of 12,000 samples from controlled experiments and plateau flight trials (0-4500 m).

Main Results:

• The proposed HAARN method reduced thrust attenuation by approximately 12.5% at 4000 m altitude.
• Energy efficiency was improved by 8.3% at 4000 m compared to traditional PID and lookup-table methods.
• Consistent performance improvements were observed across the tested altitude range (0-4500 m).

Conclusions:

• The deep neural network-based HAARN offers superior high-altitude adaptability for heavy-load UAV power systems.
• This intelligent method provides a robust solution for optimizing UAV performance in challenging environmental conditions.
• The findings demonstrate significant improvements in thrust and energy efficiency, paving the way for more reliable high-altitude UAV operations.