Compensation scheme for PDR using sparse location and error model

One of the indoor localization methods utilizing accelerometer and gyroscope is called PDR (Pedestrian Dead Reckoning). Various schemes have been proposed in PDR, however, sufficient precision has not been achieved because of the error accumulation. In this research, we propose a PDR error compensation scheme based on an assumption that can obtain sparse locations. Sparse locations mean discontinuous locations obtained by using absolute localization method or passage detection devices (ex. BLE, Magnetic Field). We attempt to compensate the trajectory of PDR estimation and improve estimation precision by sparse locations. Then we tune parameters describing errors. In our scheme, we define error models that represent errors in PDR, including moving distance error and orientation changing error. Moreover, we define parameters to describe these errors. As a result, proposed scheme improved position error rate approximately 10% and route distance error rate approximately 7%.