Model Expansion
Mathematical enrichment of a system model includes temperature-dependent variables to improve estimation accuracy under varying environmental conditions. Implementing dynamic thermal state augmentation allows a filter to track internal sensor temperatures as part of the state vector. This approach treats thermal influences as active states rather than static offsets.
State Integration
The filter appends temperature and its derivatives to the existing motion states. Through dynamic thermal state augmentation, the system estimates the bias of an accelerometer as a function of its current thermal profile. It uses a state transition matrix that includes thermal time constants.
Accuracy Enhancement
Standard calibration often fails when temperatures shift rapidly because the sensor bias does not track the external thermometer perfectly. By using dynamic thermal state augmentation, the processor accounts for the lag between the heat source and the sensing element. The model uses heat transfer equations to estimate the internal temperature of the silicon die.
This leads to a more stable output during rapid warm-up periods or when the device is exposed to sunlight. Precise tracking of these internal states reduces the overall error budget of the navigation system.
Computational Demand
Adding these states increases the matrix dimensions and the required floating-point operations. Despite this cost, dynamic thermal state augmentation is necessary for high-performance inertial units.
