Comparative Analysis of Pedestrian Dead Reckoning Algorithms for Indoor and Outdoor Localization

Pedestrian Dead Reckoning (PDR) plays a crucial role in indoor and outdoor localization, particularly in environments where Global Navigation Satellite Systems (GNSS) signals are limited or unavailable. In this study, we compare various PDR algorithms using real-world data collected on both indoor and outdoor tracks. Our investigation focuses on identifying optimal methods for step-detection and step-length estimation, essential components of PDR systems. We find that simpler algorithms, such as the SciPy find peaks function and simple threshold techniques, perform better than more complex approaches, showing robustness across different individuals and environmental conditions. Additionally, we propose a new method for step detection and path reconstruction and we show the critical role of Orientation data (AHRS). We highlight the importance of proper preparation and calibration for reliable trajectory reconstruction. Our findings provide valuable insights for developing and implementing effective PDR systems in both indoor and outdoor settings.