Predictive estimation of low density parity check decoded data based on error pattern modelling in wireless body area networks

Abstract

Summary Wireless body area networks (WBANs) dynamically track a person's health data wirelessly. The implant‐to‐implant communication (CM1) inside the human body requires low power, highly reliable channels amidst the high losses and attenuation due to the various portions of the body. This attenuation and noise results in complete variation of the monitored health data, resulting in faulty medical diagnosis and alarms. Low density parity check (LDPC) algorithms are popular for error correction in this scenario. However, they still need to be improved for a more reasonable analysis of the health status. We suggest processing of the sensed data to increase the sparsity, thereby improving the bit error probability performance of the LDPC coding technique. Further, we propose the use of Kalman filter to estimate the actually transmitted bit from the LDPC decoded message output. The simulation results show that the proposed technique improve the error correction capability of the LDPC code. The increased energy, time and computations required for the same is found to be within a satisfactory level for the proper functioning of the WBANs.