Vehicle measurements drive a neural ODE tire model that captures nonlinear coupling effects and improves tire force prediction accuracy.
Pile-up event classification separates multi-photon pulses from single-photon counts, improving energy bin accuracy and resolution.
An embedded physical layer integrates high-speed and low-speed capabilities within an FPGA.
Coordinate descent minimizes univariate quartic polynomials via closed-form cubic roots, resolving high computational complexity in nonconvex phase retrieval.
A measurement predictor synchronizes asynchronous sensor data with controller sampling rates using process models.
Differential recursive evaluation segments tuples into new, old, and removed sets to eliminate redundant computations while maintaining completeness.
A deflector with a backstep structure deflects the moving web to enable precise cross-direction tension profile measurements.
Directional filtering applies asymmetric biasing factors to prediction errors, reducing rejected products while minimizing production costs.
Reverse bump test identifies cross-directional controller alignment by inserting step signals into scanner measurements, maintaining closed-loop stability.