고속 목표물 실시간 추적을 위한 FPGA를 이용한 병렬적 Huber 기반 확장 칼만 필터링

초록

This paper addresses a Huber-based Extended Kalman Filter(EKF) that applies the Huber function to reduce the influence of outliers and enhance the robustness of the filter in the EKF algorithm. The Huber function improves filtering performance by suppressing outliers but imposes additional computational complexity, making real-time processing challenging. This study proposes an optimized hardware architecture utilizing FPGA-based parallel processing techniques to address these challenges, enabling computational efficiency and ensuring real-time processing capability. The proposed method processes complex matrix computations in the Huber initialization phase in parallel and minimizes computational latency through an optimized hardware architecture. Simulation results demonstrate that the proposed method reduces computation time by approximately 20.2% while maintaining the same level of noise reduction performance and estimation accuracy as conventional methods.

키워드