Copyright @ 200 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.7 Study of Human Outdoor Walking with a Low-Cost GPS and Simple Spreadsheet Analysis

walking of healthy subjects, allowing the detection of walking and resting (nonwalking) periods and the accurate estimation of speed and distance of each walking periods. Methods: The same EGNOS-enabled GPS receiver was used for all experiments. In experiment 1, various signal-processing methodologies were tested for the detection of both walking and resting bouts from a prescribed walking protocol (PWP) that was performed 21 times by six healthy subjects on an outdoor athletic track. In experiment 2, the accuracies of these processing methodologies were then tested through a blinded analysis of different PWP for 10 other healthy subjects in a designated public park. In experiment 3, speed and distance calculated by the GPS receiver during series of 100–400 m on an outdoor athletic track were compared with actual speed and distance. Results: Raw data were inaccurate, but the combination of a low-pass filter, an adapted high-pass filter, and artifact processing enabled one to detect walking and resting bouts with an accuracy of 89.8 % (95 % CI, 84.4–93.4). A manual post-processing methodology, used to complete previous automatic processing results, provided the highest concordance with the PWP, reaching an accuracy of 97.1 % (95 % CI, 93.5–98.8). There was an excellent relationship both between actual and processed distances (R2 = 1.000) and between actual and processed speeds (R2 = 0.947). Conclusion: Low-cost, commercially available GPS may b