Robust motion mode recognition for portable navigation independent on device usage

Abstract

Portable navigation has become increasingly prevalent in daily activities. The need for accurate user positioning information, including a person's location and velocity, when using a portable device (such as a cell phone, tablet, or even a smart watch) is growing in various fields. Knowing the user's mode of motion or conveyance allows appropriate algorithms or constraints, related to each mode, to be used to estimate a more accurate position and velocity. The modes covered in this paper are walking, running, cycling, and land-based vessels (including vehicles, truck, buses, and trains which include light rail trains and subways). The work discussed in this paper involves the use of sensors - with and without Global Navigation Satellite Systems (GNSS) signal availability - in portable devices to help recognize the mode of motion for an arbitrary user, an arbitrary use case - whether the device is held in the hand, in the pocket, or at the ear, etc. - and an arbitrary orientation of the device.