Optimizing GNSS Indoor Outdoor Detection: Balancing Observation Window and Sampling for Accuracy and Responsiveness

The balance between sampling rate and response time is important for the efficiency of Global Navigation Satellite System (GNSS) based Indoor Outdoor Detection (IOD) systems. This paper explores the tradeoff between sampling rate and observation window, investigating how these factors influence the accuracy of IOD. For data collection, we adopt Raspberry Pi 4B+ with PmodGPS receivers, GlobalTop FGPMMOPA6H, allowing us to customize the data sampling. Our study utilizes Support Vector Machine (SVM) as the base model for the experiment, training the model with resampled datasets to simulate different sampling rates and observation windows. The results indicate that a higher sampling rate and shorter observation window lead to improved accuracy and response time. By adjusting the sampling rate from 1Hz to 10Hz, the accuracy can be improved by 2% and approximately 21 seconds entry delay improvement. Further, shortening the observation window will provide better response time, which in our case can achieve up to 10 minutes for exit delay. The study comparing with related works also indicate the importance of Assisted Global Positioning System (AGPS) and multiple GNSS for the IOD, with both AGPS and multiple GNSS, the responsiveness can be improved significantly.