Differential GNSS and Vision-Based Tracking to Improve Navigation Performance in Cooperative Multi-UAV Systems

Area of Science:

• Robotics and Control Systems
• Navigation and Guidance
• Aerospace Engineering

Background:

• Micro-unmanned aerial vehicles (UAVs) commonly use Global Navigation Satellite System (GNSS) and Micro-Electro-Mechanical Systems (MEMS) for navigation.
• Standard navigation systems may lack the required accuracy for precision tasks like fine sensor pointing.
• Cooperative navigation strategies are needed to overcome individual vehicle limitations.

Purpose of the Study:

• To develop and assess a cooperative UAV navigation algorithm for enhanced real-time and post-processing performance.
• To improve the position and attitude accuracy of a chief UAV by leveraging formation-flying deputy vehicles.
• To create a virtual navigation sensor through the fusion of differential GPS and vision-based tracking data.

Main Methods:

• Implementation of a cooperative navigation algorithm for UAVs in outdoor environments.
• Utilization of differential GPS (DGPS) among formation-flying vehicles for precise positioning.
• Integration of vision-based tracking data with DGPS information to form a virtual sensor.
• Application of an Extended Kalman Filter for sensor fusion and navigation estimation.
• Focus on the DGPS/Vision attitude determination algorithm.

Main Results:

• Demonstrated potential for significantly improved navigation accuracy, particularly in attitude determination.
• Achieved accurate attitude information independent of magnetic and inertial sensors.
• Validated the approach through both numerical simulations and real-world flight tests.
• Compared DGPS/Vision estimates with onboard autopilot system data on a customized quadrotor.

Conclusions:

• The developed cooperative DGPS/Vision navigation approach offers a viable solution for enhancing UAV navigation performance.
• Exploiting formation flying and sensor fusion provides accurate attitude estimation, crucial for advanced missions.
• The method shows promise for applications requiring high-precision navigation beyond standard GNSS/MEMS capabilities.