WSN-based Real-Time Indoor Location System at the Taipei World Trade Center: Implementation, Deployment, Measurement, and Experience

Given the popularity of location-based services today and the advances in indoor localization, we find it curious that the systems are rarely deployed for public use. Having implemented a RF-fingerprint-based location system at the Taipei World Trade Center (TWTC), we are able to examine quantitatively the system performance. In that, the 50-percentile localization error is approximately 6.2m. This is significantly larger than that of most testbed studies, indicating significant discrepancy between the results taken from fine-tuned testbeds vs. flash, in-situ deployments. We also find that dynamics in physical environment and RF signals impact the results more than the survey density, indicating a well-designed estimation method may alleviate the labor requirement for RF-fingerprint survey. Our measurement study is unique in that the area of deployment is large and the 'window of opportunity' for deployment is limited to a few hours. This we refer to as a 'flash deployment'.