Event-Driven Maximum Correntropy Filter Based on Cauchy Kernel for Spatial Orientation Using Gyros/Star Sensor...

Area of Science:

• Aerospace Engineering
• Control Systems
• Signal Processing

Background:

• Gyro/star sensor integration offers high-accuracy spatial orientation for turntable structures.
• Complex spatial environments introduce non-Gaussian measurement noise, degrading accuracy.
• Existing filtering methods struggle with non-Gaussian noise, impacting robustness.

Purpose of the Study:

• To develop a robust filtering method for gyro/star sensor integration under non-Gaussian noise conditions.
• To enhance the real-time performance and computational efficiency of spatial orientation systems.
• To improve the control stability and accuracy of turntable structures.

Main Methods:

• Established a direct installation mode for gyro/star sensor integration with a derived mathematical model.
• Developed a Cauchy kernel-based maximum correntropy filter to mitigate non-Gaussian noise effects.
• Constructed an event-driven mechanism using filter innovation information to reduce computational load.

Main Results:

• The proposed filter demonstrated strong robustness against non-Gaussian measurement noise.
• Simulations confirmed enhanced spatial orientation accuracy for turntable structures.
• The event-driven mechanism significantly optimized real-time performance with reduced computational cost.

Conclusions:

• The event-driven maximum correntropy filter effectively handles non-Gaussian noise in gyro/star sensor integration.
• The methodology provides a robust and computationally efficient solution for real-time spatial orientation applications.
• This approach enhances the overall performance of turntable spatial orientation systems.