Most modern cellular phones and tablets come with numerous sensors. You probably already know about WiFi and Bluetooth and use them everyday, but there are others as well.
For example, current smart devices come with built-in accelerometer, barometer, magnetic sensor and directional sensor. While these devices help fitness apps and the compass do their jobs, the patent-pending SPREO indoor positioning algorithm incorporates these signals for accurate and dynamic indoor position definition. We call it sensor fusion, and we use these additional sensors in addition to Bluetooth Low Energy beacons (“iBeacons”) and Wi-Fi to get accuracy to better than a meter.
The sensor fusion approach does create a unique set of challenges. For example, there is a chance for “accumulated error.” Most people vary their gait (“average step size”) and meander slightly instead of walking a true straight path, especially when talking through crowded hallways and stairs.
We have created a set of discernment rules to interpret the sensors input values and determine whether a step was actually taken in a certain direction or it was just “noise” (irrelevant data inputs). Consider for example how sensor data points would fluctuate if a person was talking with their hands while holding their phone. Another practical example is more subtle and technical: many environments have electrical magnetic resonance which can generate environmental ‘noise’ (data fluctuations).
The correct approach is ultimately a sophisticated algorithm that strategically integrates all available and applicable sensors on smart devices. The key to sensor fusion is to correct and minimize errors in order to achieve accurate position prediction.
We’ve taken indoor location with sensor fusion a step further by developing machine learning algorithms that analyze user behavior and environmental conditions. As a result, our proprietary methodology today provides positioning accuracy under a meter.