Roll Orientation Estimator for Smart Projectiles Using Thermopile Sensors

Jonathan Rogers, Mark Costello, David J. Hepner

Journal of Guidance Control and Dynamics · 2011 · 35 citations · 6 references

Concepts

Abstract

The use of inexpensive, commercially available thermopiles sensors for roll orientation estimation of spinning bodies is explored. The sensors convert observed thermal gradients into an electrical signal well suited for onboard data acquisition and real-time signal processing. An environmental model emulating sensor stimulus for a six-degree-of-freedom body is generated given standard atmospheric and typical ground conditions. When sensor characteristics are included, the fully developedmodel can be used to generate accurate sensor output as a function of Euler angles and altitude. Outputs from the model are then shown to compare favorably with experimental flight data, capturing the predominant and nearly sinusoidal signal variation as the projectile rolls. An extended Kalman filter algorithm is offered, which enables real-time roll angle and roll rate estimation using solely thermopiles as feedback. Example results demonstrate that the algorithm yields reasonably accurate roll information. Finally, a trade study demonstrates that roll error is further mitigated as the number of thermopile sensors is increased. This research shows that thermopiles could be useful in a diverse multisensor constellation as a convenient absolute inertial roll reference. Nomenclature II, JI,KI = inertial reference frame unit vectors IP, JP,KP = projectile reference frame unit vectors kv = sensor scale factor P = roll angle error covariance estimate Q = model error covariance for Kalman filter R = measurement error covariance for Kalman filter T = infrared transmission coefficient V = thermopile output signal xe, ye, ze = position coordinates with respect to the inertial frame of the projectile mass center y = infrared emissivity = azimuth angle with respect to inertial frame = angle with respect to horizon i = thermopile field-of-view roll angle offset from JP axis , = projectile Euler roll and pitch yaw angles I.

References

6