A Kalman filter is a mathematical method used to estimate the true position and motion of a system when the measurements are noisy. For a radio dish, it can make the pointing system smoother and more accurate by combining what the motors think they are doing with what the sensors actually measure.
Why use one on a radio dish?
Suppose your dish is tracking the Hydrogen line or a spacecraft. You may have:
Shaft encoders measuring azimuth and elevation.
Motor current sensors.
GPS time information.
An inclinometer.
Optical or radio pointing corrections.
Each measurement contains errors:
Encoder noise.
Wind gusts causing slight deflection.
Mechanical backlash.
Quantisation errors.
Timing jitter.
The Kalman filter estimates the dish’s true state despite these imperfections.
What does it estimate?
For an altitude-azimuth dish, the state might be:
Azimuth angle
Azimuth angular velocity
Elevation angle
Elevation angular velocity
Written as:
x = [Az, Az_rate, El, El_rate]^T
Every few milliseconds it performs two steps.
1. Prediction
Using the previous estimate and a model of how the dish moves:
New position
≈ Old position + velocity × time
For example:
Az(k+1) = Az(k) + Az_rate(k) × dt
It predicts where the dish should now be.
2. Update
It compares the prediction with the actual encoder readings.
If the encoder suddenly jumps slightly because of noise, the filter only partly trusts it.
The weighting depends on:
How noisy the sensors are.
How much confidence you have in the motion model.
The result is a statistically optimal estimate (assuming Gaussian noise and a linear system).
Benefits
A Kalman filter can:
Smooth noisy encoder readings.
Reduce pointing jitter.
Estimate angular velocity without numerically differentiating noisy data.
Continue producing sensible estimates if a measurement is briefly lost.
Improve tracking of fast-moving objects such as satellites.
Help compensate for wind disturbances.
Example
Suppose:
Encoder says: 120.015°
Prediction says: 120.002°
If the encoder is known to be noisy, the filter might estimate:
120.005°
instead of accepting either value completely.
Radio astronomy applications
Kalman filters have been used in systems such as:
NASA deep-space antenna tracking.
Satellite ground stations.
Large steerable radio telescopes.
Adaptive optics systems in optical astronomy.
Interferometer antenna control.
For an amateur hydrogen-line telescope with a relatively slow-moving dish, a simple Kalman filter running on a Raspberry Pi or microcontroller can improve the smoothness of encoder feedback and provide better estimates of pointing rates. The improvement is often most noticeable in windy conditions or when using low-cost encoders.
In essence, a Kalman filter acts like an intelligent averaging system: it predicts where the dish should be, checks what the sensors report, and continuously combines the two to give the best estimate of where the dish is actually pointing.