Strapdown Seeker Line-of-Sight Rate Estimation via Pixel-Plane Kalman Filtering and Gyro Fusion

Abstract

Proportional-navigation (PN) guidance requires an accurate estimate of the inertial line-of-sight (LOS) rotation rate between an interceptor and its target. In a strapdown (body-fixed, non-gimbaled or partially stabilized) electro-optical seeker, this rate cannot be measured directly: the seeker only observes the target’s position in the image plane, while the vehicle’s own rotation is sensed separately by a gyro. We present a complete estimation pipeline that (i) tracks the target’s normalized image-plane coordinates with a constant-velocity Kalman filter, (ii) reconstructs the camera-frame LOS rate from the filtered pixel state, and (iii) fuses it with the measured body angular rate through the rigid-body transport theorem to recover the inertial LOS rate. The pipeline is validated across six progressively realistic simulation scenarios: a deterministic zero-noise baseline, a maneuvering target combined with a finite-bandwidth gimbal, a 1000-run Monte Carlo sweep under pixel and gyro noise/bias, a closed-loop three-dimensional true-PN intercept guided entirely by the estimated LOS rate, a Kalman process-noise sensitivity sweep exposing the classic tracking-lag/noise-rejection trade-off, and a pixel-channel latency sweep quantifying the effect of uncompensated tracker delay. The estimator reproduces the groundtruth LOS rate to within 3.13×10−4 deg/s RMS in the noise-free baseline, remains accurate under target maneuvers and gimbal lag, achieves a 0.899±0.028 deg/s RMS error under realistic sensor noise, and drives a closed-loop intercept to a 1.71m miss distance.