PDRNN: Modular Data-driven Pedestrian Dead Reckoning on Loosely Coupled Radio- and Inertial-Signalstreams
The paper presents PDRNN, a modular data-driven pedestrian dead reckoning system that integrates radio and inertial signal streams. It addresses challenges in traditional methods by employing a recurrent neural network architecture to forecast asynchronous sensor data. Experimental results indicate that PDRNN outperforms classic methods in accuracy and precision during dynamic movements.
- ▪PDRNN is designed to fuse noisy and biased estimates from loosely coupled sensors for accurate localization.
- ▪The system uses separate machine learning models to estimate orientation, velocity, and distance from sensor data.
- ▪PDRNN's modular design allows for individual components to be updated without affecting the entire system.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Machine Learning arXiv:2605.15252 (cs) [Submitted on 14 May 2026] Title:PDRNN: Modular Data-driven Pedestrian Dead Reckoning on Loosely Coupled Radio- and Inertial-Signalstreams Authors:Peter Bauer, Andreas Porada, Felix Ott, Christopher Mutschler, Tobias Feigl View a PDF of the paper titled PDRNN: Modular Data-driven Pedestrian Dead Reckoning on Loosely Coupled Radio- and Inertial-Signalstreams, by Peter Bauer and Andreas Porada and Felix Ott and Christopher Mutschler and Tobias Feigl View PDF HTML (experimental) Abstract:Modern pedestrian dead reckoning (PDR) systems rely on fusing noisy and biased estimates of position, velocity, and calibrated orientation derived from loosely coupled sensors to determine the current pose of a localized object.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
