Real-time pedestrian tracking in indoor environments

Abstract

Tracking pedestrians in indoor environments has increasingly become a needed topic to optimize human activities in case of emergency interventions, guidance, entertainment, healthcare, advertisement, etc. Additionally, portable and wearable devices are overtaking traditional computers. Their ubiquitous/pervasive features added to their increasing technical specifications and sensing capabilities have made them an attractive option for indoor pedestrian tracking based on dead-reckoning (DR). This paper presents a solution for real-time tracking of pedestrians in indoor environments using sensing raw data from low cost sensors (such as accelerometer or gyroscope) embedded in modern smartphones. The proposed application combines a simplified DR approach, based on the identification of human gait for step detection and distance estimation, along with the detection of identifiable indoor signatures such as corners. The application was implemented and tested on Android-based mobile phones obtaining an accuracy of 89.93% in common indoor human motion displacements.