Evaluation of a New Energy-Based Human Tracking Method in a Smart Building Using Floor Vibration Measurements

Tracking occupants in an indoor environment has applications in intruder detection, emergency response and evacuation (e.g., locating an occupant in a burning building), and energy saving (through activity-based control of building lighting and HVAC system). In this document, we show that tracking occupants in an indoor environment can be done using the floor vibration caused by occupant footstep impacts. In order to track an occupant, each footstep impact location must first be estimated. For that purpose, we evaluate the performance of a newly developed energy-based localization (multilateration) method for the case of localizing occupant footsteps in a real-life operational smart building. The new method is based on the fact that the energy of the impact-generated wave will be attenuated as the wave travels away from the impact location. Localization is achieved using a network of vibration sensors (accelerometers) placed underneath the walking floor, which provides a non-intrusive and tamper-proof localization system. The new method has small computational time and requires a relatively small sensor data sampling rate. It is anticipated that the new method will have a smaller footstep localization error compared to conventional time of flight/arrival methods. Occupant tracking experiments show that the new method has a promisingly small localization error.