GPU-Based Parallel Kalman Filter

The Kalman Filter and its variants have been highly successful in numerous applications in technology. However, the Kalman filter is under heavy computational burden. When suffers from big data, it becomes pretty slow. On the other hand, the GPU, a processor unit with highly parallel structure, becomes more and more popular in generous purpose computing. This paper focuses on how to make Kalman filters faster on GPU while introducing flexibility between accuracy and speed. Our parallel Kalman filter can achieve nearly linear speedup in simple applications and can outperform CPU program by an order of magnitude in some real world applications.