OnQ Blog

Sensor fusion: merging your data into one story

20 Jan 2016

Qualcomm products mentioned within this post are offered by Qualcomm Technologies, Inc. and/or its subsidiaries.

We’re at the dawn of a new, connected age that will abound with devices that make up the Internet of Things. Some of these devices rely on data collected by sensors—and when a device (such as a fitness tracker or a car) is collecting data from multiple sensors, it needs a way to integrate the data to create a cohesive whole. That’s where sensor fusion comes in.

The term “sensor fusion” means what it sounds like—it’s the combining of data from numerous sensors that provides a larger, broader picture than individual sensors can supply alone. Because each sensor has defined functions and inherent limits, collecting information from each sensor separately wouldn’t give a complete report. Instead, the collected data is merged for analysis, so your device receives a clearer indication of what you’re doing and the environment around it.

For example, take wearable devices, like smartwatches and activity trackers. You may be familiar with some of the types of sensors these devices include, such as accelerometers, gyroscopes, barometers, thermometers, and altimeters. But additional sensors, such as the optical sensor (which accurately measures your heart rate by shining an LED light through your skin), the galvanic skin response sensor (which can help indicate your stress response by measuring the electrical conductivity of your skin due to external forces like sweat), and the ambient light sensor (which can help in sleep tracking by determining how bright the surrounding environment is) are capturing a more well-rounded picture of your current physical or emotional state.

But for a wearable device to understand that you’re hiking uphill, that you’re burning enough calories, and that you’re doing all of this at a national park, a digital signal processor (DSP)—such as the Qualcomm Hexagon DSP—is needed to make sense of all of the data collected. The Hexagon DSP can juggle multiple sensor inputs, translating the information into context-aware actions. And a bonus: It does all of this without consuming precious battery life. Qualcomm Snapdragon processors smartly integrate the Hexagon DSP into the SoC for optimal sensor processing efficiency. Taking power savings to the next level, the Hexagon 680 DSP in the Snapdragon 820 processor has a completely separate DSP for sensor processing built on a low power island.

Sensor fusion contributes to a device’s contextual awareness; as do scene recognition, proximal awareness, and learned preferences. With contextual awareness, your device—a smartphone, a car or a VR headset with sensors and a powerful DSP—can learn about you and your surroundings, providing a personalized and thus more immersive experience.

Find out more about how Qualcomm® technologies help to provide more immersive experiences on your devices.


Qualcomm Snapdragon and Qualcomm Hexagon are products of Qualcomm Technologies, Inc.