BP Bit-Flipping Decoding Algorithm for Polar Codes in Satellite Internet of Things Communication Based on Information Geometry

With the emergence of the Satellite Internet of Things (S-IoT), mega-constellations are playing a pivotal role as the core communication infrastructure in large-scale connectivity scenarios. However, due to characteristics such as long link delays, Doppler frequency offsets, and highly dynamic handovers, satellite communication systems tailored for the IoT are prone to decoding errors under medium-to-low signal-to-noise ratio (SNR) conditions, severely compromising the reliable communication of end devices. Polar codes’ Belief Propagation (BP) decoding inherently possesses parallelism, making it suitable for resource-constrained edge IoT terminals, yet its decoding accuracy remains limited under severe noise conditions. To enhance the decoding performance of polar codes in satellite communication, this paper proposes a novel polar codes BP bit-flipping (BPF) decoding algorithm based on the information geometric framework. By introducing trajectory energy density and a joint sensitivity metric, the algorithm integrates the dynamic evolution of soft information during BP decoding with local Fisher information, enabling precise localization of unstable bits in the decoding process. Experimental results demonstrate that, under low SNR conditions, the proposed algorithm significantly reduces the block error rate (BLER). It enhances decoding accuracy through dynamic adjustment of the number of flipped bits and geometric perturbations, effectively meeting the stringent requirements of satellite communication systems for high reliability and low BLER.