An Inertial Measurement Unit (IMU) is an electronic sensor assembly that measures linear acceleration and angular rotation rates along three orthogonal axes, providing the fundamental data needed for inertial navigation and motion sensing applications. Modern IMUs integrate multiple microelectromechanical systems (MEMS) sensors, typically three accelerometers measuring acceleration along the X, Y, and Z axes, and three gyroscopes measuring rotation rates around those same axes, into compact packages suitable for integration into vehicles, robots, drones, and mobile devices.

IMUs serve as critical components in GNSS/INS (Inertial Navigation System) integration, providing continuous motion measurements that bridge GNSS gaps and improve positioning robustness. While GNSS provides absolute position but may suffer outages in tunnels, urban canyons, or under dense foliage, IMUs operate independently of external signals and provide high-rate measurements (typically 100-400 Hz or higher) regardless of environment. By mathematically integrating acceleration and rotation measurements, navigation systems can track position, velocity, and orientation changes relative to a known starting point, a technique called dead reckoning.

The accuracy and performance of IMUs vary enormously across different grades and price points. Consumer-grade MEMS IMUs found in smartphones exhibit significant noise and bias instability, limiting dead reckoning accuracy to tens of meters over just a few minutes. Tactical-grade IMUs used in automotive and robotics applications offer improved stability, supporting meter-level dead reckoning over similar periods. Navigation-grade and strategic-grade IMUs used in aerospace applications achieve exceptional stability but at substantially higher cost. Key performance specifications include bias stability, noise density, scale factor accuracy, and bandwidth.

For safety-critical applications like autonomous vehicles, IMUs must meet specific functional safety requirements defined by standards such as ISO 26262. This may include self-diagnostic capabilities to detect sensor failures, redundant sensor configurations for fault tolerance, and documented safety integrity levels. The tight integration of IMU data with GNSS observations through Extended Kalman Filter or similar fusion algorithms has become standard practice in modern positioning systems, combining the absolute accuracy of satellite positioning with the continuity and high update rates of inertial sensing.