Acceleration of Signal Prediction in Wireless Sensor Networks Based on Kalman Filter and Goldschmidt Algorithm

Abstract

The expansion of possibilities in the use of wireless sensor networks requires the continuous improvement of their technical characteristics. From this point of view, the development of promising methods and algorithms for the transmission and processing of digital data obtained from sensor sensors in wireless sensor networks in real time is one of the topical areas of research in digital signal processing. Known methods and algorithms for data processing are based on computational structures, including the execution of the arithmetic operation of division. Performing a division operation increases the computational complexity of the system, which leads to a significant decrease in the performance of the wireless sensor network. To solve this problem, this paper proposes a modified probabilistic Kalman filtering algorithm with a Goldschmidt divisor. Software simulation showed that the time-varying filter has a higher performance than the known Kalman filtering algorithm and the same measurement error caused by the presence of measurement noise. The resulting algorithm can be effectively applied in various monitoring systems.