No abstract

This paper presents an approach to implementing the parallel Kalman filter. The parallel Kalman filter is computationally intensive and hence complex due to the inherent matrix operations. Often, the filter cannot be computed in real time when the system has a large number of state variables. Discussed is a method for achieving almost real-time performance. In addition, a method for determining stability of the Kalman filter is presented. This method involves the use a paradigm that achieves high performance and close to real time results while maintaining accuracy for the Kalman filter. In addition, the paradigm shows an improvement of the processor utilization over other implementations and provides flexibility in terms of the hardware used for implementation.

No abstract

No abstract