Sensor Fusion
Sensor Fusion — The combination of measurements from multiple sensors with different noise, timing and observability characteristics into a shared state estimate or decision.
Why Multiple Sensors?
Different sensors observe the same physical system with different failure and noise characteristics. One sensor may respond quickly but drift, while another is slower but more stable over time.
Fusion attempts to use that complementary information rather than treating every measurement as equally reliable.
Time Alignment
Mathematical combination is not enough. Sensors can have different sample rates, timestamp sources and transport delays. Two individually correct measurements can produce a wrong estimate when aligned to the wrong time.
Model and Trust
Kalman-family filters model process and measurement uncertainty explicitly, while simpler complementary filters or domain-specific logic may also be appropriate.
Adding sensors does not automatically improve accuracy; correlated errors and shared failure modes still matter.