Lagrangian Point State Estimation with Optimized, Redundant Induction Coil Gauges

摘要

This research effort studies the use of redundant induction coil gauges to reduce state estimation uncertainties for moving Lagrangian points (LPs); e.g. discrete points, moving interfaces, projectiles, etc. The technique embeds a small, high-strength magnet at the LP and simultaneously tracks the magnet continuously with five (5) or more induction coils along a single axis of motion. A calibrated coil gauge model is presented as a function of LP position and velocity. The optimized LP state (position and velocity) estimate based upon redundant LP observations allows direct solution for LP velocity; requiring only one differentiation step to obtain acceleration. A specific experimental implementation (Particulate Materials Mesoscale Diagnostics system) is simulated to evaluate and minimize the expected state estimation errors. Induction coil signals with various levels of noise are simulated based upon a prescribed LP state variation with time. The state optimization algorithm attempts to recover the truth state values. Worst-case position estimation errors of ±0.3 mm and velocity estimation errors of ±0.46 m/s are determined for LPs travelling 0-1,000 m/s at realistic in-lab data noise levels.