IMU Drop Detection With Adaptive Sampling and Rotation Correction

The patent implements dynamic sampling rate adjustment based on device state. The system transitions between low sampling rate (normal state) and high sampling rate (drop detection state), allowing high signal resolution only when needed for drop detection while maintaining low power consumption during normal operation. This is achieved through state-machine based control that activates high-rate sampling only when drop indicators are detected.

The system changes the sampling rate parameter dynamically based on detected conditions. When the device is in a normal state, a low sampling rate is used to conserve energy. When indicators such as angular velocity thresholds or acceleration patterns suggest a drop event, the system switches to a high sampling rate to capture detailed motion data for accurate drop verification.

The patent performs preliminary correction of device parameters during manufacturing or initial setup. Intrinsic parameters such as center of mass position, moment of inertia, and sensor orientation are calibrated beforehand and stored in the device. During runtime, these pre-corrected parameters are used directly without requiring complex real-time calculations, thus improving detection accuracy while avoiding increased operational complexity.

The system uses pre-measured or pre-calculated device parameters (copies of the actual physical properties) to simplify runtime operations. Instead of measuring physical properties like moment of inertia or center of mass during operation, the system uses stored parameter values that represent these properties, enabling accurate drop detection without complex real-time physics calculations.