Ambiguity Function
Understanding the Ambiguity Function
A radar transmits a waveform and then applies a matched filter (the time-reversed conjugate of the transmitted waveform) to the received echo to maximize SNR. The matched filter output is a function of both time delay (range) and frequency shift (Doppler velocity). The ambiguity function maps this output power across all possible delay-Doppler combinations, showing exactly where the radar can and cannot distinguish between targets.
At the origin (τ = 0, ν = 0), the ambiguity function always peaks, because the filter is perfectly matched to the true target. Away from the origin, the function shows how much the filter responds to targets at different ranges or velocities. A narrow central peak means fine resolution; sidelobes away from the peak represent range-Doppler combinations that could produce false detections or mask weaker targets near a strong one.
χ(τ, ν) = ∫ u(t) × u*(t − τ) × ej2πνt dt
Ambiguity Surface (power):
|χ(τ, ν)|²
Volume Constraint (uncertainty principle):
∫∫ |χ(τ, ν)|² dτ dν = |E|² (constant for all waveforms)
Zero-Doppler Cut (autocorrelation):
|χ(τ, 0)|² = |R(τ)|², the standard range resolution profile
The volume constraint means you can reshape ambiguity but never eliminate it. Compressing the peak in one dimension spreads energy to sidelobes or the other dimension.
Waveform Ambiguity Shapes
| Waveform | Ambiguity Shape | Range Resolution | Doppler Resolution | Application |
|---|---|---|---|---|
| Unmodulated Pulse | Diagonal ridge | c×T/2 (poor for long T) | 1/T (good for long T) | Simple pulse radar |
| Linear FM (Chirp) | Rotated ridge (sheared) | c/(2B) (set by bandwidth) | 1/T (set by duration) | Pulse compression radar, FMCW |
| Barker Code | Thumbtack (narrow main, low SL) | c/(2B) | 1/T | Low-sidelobe search radar |
| Wideband Noise | Ideal thumbtack | c/(2B) | 1/T | Noise radar, LPI |
| CW (continuous) | Knife-edge (range axis) | None (no range info) | Excellent | CW Doppler, speed gun |
Frequently Asked Questions
What does the Ambiguity Function physically represent?
It shows the matched filter output power versus target range error (τ) and Doppler error (ν). At the origin, it peaks (perfectly matched). Any other point shows how much the filter responds to a target at a different range or velocity. A narrow peak means good resolution; sidelobes mean potential false targets. The ambiguity function is a property of the waveform alone, independent of target, noise, or channel.
Why can't a waveform have perfect resolution in both range and Doppler?
The total volume under the ambiguity surface is constrained to be constant (a consequence of the uncertainty principle). You cannot reduce ambiguity in one dimension without spreading it to another or to sidelobes. Short pulses have great range but poor Doppler resolution. Long CW tones have great Doppler but no range. Linear FM (chirp) achieves good resolution in both by trading for a sheared ridge with sidelobes.
How do you use the Ambiguity Function to choose a waveform?
Plot the ambiguity function for each candidate and compare mainlobe width (resolution), sidelobe levels (masking risk), and shape. Air surveillance needs a thumbtack shape (wideband coded waveforms). Weather radar prefers a knife-edge along range (CW pulse Doppler). Automotive FMCW uses the diagonal ridge of a linear chirp. The ambiguity function is the single most informative tool for waveform selection.