Optimizing prediction accuracy in dynamic systems through neural network integration with Kalman and alpha-beta...

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Traditional Kalman and alpha-beta filters have limitations due to static parameters in dynamic environments.
• Adaptability and accuracy are crucial for effective tracking and prediction in changing conditions.

Purpose of the Study:

• To develop and evaluate novel neural network-based prediction models for enhancing Kalman and alpha-beta filters.
• To enable dynamic parameter tuning in filters using neural network feedback.

Main Methods:

• Integration of neural networks into Kalman and alpha-beta filter algorithms.
• Neural network-driven dynamic augmentation of filter parameters (α, β, R, F).
• Performance assessment using Root Mean Square Error (RMSE) metric.

Main Results:

• The neural network-based alpha-beta filter showed a 38.2% improvement in prediction accuracy.
• The neural network-based Kalman filter achieved a 53.4% enhancement in prediction accuracy.
• Both modified filters demonstrated superior adaptability and accuracy in dynamic environments.

Conclusions:

• Integrating neural networks into filtering algorithms is an effective strategy for enhancing performance in dynamic systems.
• The proposed neural network-based filters offer significant improvements over conventional static-parameter filters.
• This approach provides a robust solution for advanced tracking and prediction applications.