8B-PGDM:An Improved Dead Reckoning Algorithm Based on Local Invariance of Navigation

This paper presents a comprehensive investigation into dead reckoning algorithms. The study begins with an in-depth analysis of the fundamental mathematical formulation for navigation position calculations, then introduces the concept of local invariance to refine traditional methods by examining state parameters with inherent stability. Building on this foundation, we propose an efficient iterative optimization algorithm designed specifically for real-time trajectory prediction under sparse sensor data conditions(Only three samples are required). This approach effectively mitigates the impact of data scarcity, enabling robust and accurate trajectory predictions in challenging environments. The proposed method is anticipated to play a pivotal role in following control systems, thereby significantly improving their operational reliability and performance.