Mimicking the brain: The emerging science of sensor fusion

A single sensor is hardly ever enough. An accelerometer can detect orientation and tilt, and can track pitch and roll — if the device is not moving. Coupling an accelerometer with a gyroscope allows the measurement of more complex motion, such as pitch and roll while the device is moving, or high-speed rotation.

Sensor fusion technology combines the outputs of multiple sensors to monitor complex or rapid movements accurately. “Fusing” many sensors together enables such diverse capabilities as gesture control of devices and the detection of falling human bodies, for example.

The science behind sensor fusion technology is neuroscience: The brain integrates perceptual information from a variety of sources into a coherent description of the world. The peripheral nervous system brings in sensory information and sends out motor information, but it is the brain that makes the decision to walk, or talk, or whatever else. That model of the brain is the template for the emerging science of sensor fusion.

But how far have we gotten in the quest to mimic the brain? We’re getting there quickly. Advances in sensor technology and processing techniques, combined with improved hardware are making real-time fusion of data possible. There’s even a new generation of sensors that can perform sensor fusion with no sensor hub connected.