The range-resolution trade-off
A short pulse gives fine detail — you can tell two aircraft 30 metres apart are separate. But it contains so little energy that it vanishes into noise before it reaches a distant target. A long pulse travels farther but smears everything into a blur. You want long-pulse energy with short-pulse sharpness.
Chirp and linear FM
Instead of a single-frequency pulse, sweep the frequency linearly across the pulse duration — a chirp. On receive, pass the echo through a matched filter that does the inverse sweep. The result compresses the long pulse into a spike as short as the inverse of the bandwidth. A 10 µs pulse with 100 MHz bandwidth compresses to 10 ns — a 1,000-fold improvement.
Phase codes: Barker and beyond
Another way: divide the pulse into sub-pulses and flip their phase according to a code. Barker codes have ideal autocorrelation — sidelobes barely above noise. Longer Frank and Pn codes stretch to thousands of chips, giving enormous compression ratios. Military radars love them because they are harder to detect and jam.
Everywhere now
Pulse compression is in your car's adaptive cruise control (FMCW + chirp), in weather NEXRAD (clear-air mode), in every military phased array, and in SAR satellites imaging the ground from orbit. It is the single most important signal-processing innovation in radar since WWII.
The side-lobe suppression challenge
Pulse compression is not a free lunch; it introduces range side-lobes that can mask smaller targets near larger ones. When a compressed pulse creates artificial echoes in adjacent range bins, a nearby small drone might disappear in the 'pedestal' of a large cargo plane's return. To solve this, engineers apply weighting functions—such as Taylor or Hamming windows—during the reception phase. While these windows effectively suppress the side-lobes to as low as -40 dB, they come at the cost of slightly widening the main pulse and reducing the signal-to-noise ratio by several decibels.
The mathematical precision required is immense. Modern digital signal processing (DSP) handles this via Fast Fourier Transforms (FFT), but early implementations relied on surface acoustic wave (SAW) filters. These analog devices used precisely etched lithium niobate crystals to physically slow down specific frequencies, effectively 'stacking' the chirped pulse in time. This hardware-level processing was the backbone of Cold War-era early warning systems before the advent of high-speed analog-to-digital converters made the transition to pure software-defined radar possible in the late 1990s.
Synthetic Aperture Radar (SAR) and high-speed imagery
The most dramatic application of pulse compression is Synthetic Aperture Radar (SAR). By combining high-bandwidth pulse compression with the movement of the radar platform itself, SAR can achieve sub-meter resolution from orbital altitudes. This technique effectively simulates an antenna miles long. Without compression, the transmitter power required to get that level of detail from 400 kilometers away would melt the satellite's electronics. Pulse compression allows the system to spread that energy over time, maintaining a manageable peak power while capturing photographic-quality maps in all weather conditions.
A common misconception is that pulse compression only benefits long-range surveillance. In reality, it is critical for modern automotive radar used in self-driving suites. These 77 GHz sensors use Frequency Modulated Continuous Wave (FMCW), a form of infinite pulse compression. By constantly sweeping the carrier frequency, they can distinguish between a stationary parked car and a pedestrian just centimeters away. This high resolution at short ranges is what enables automated emergency braking systems to operate with high confidence in cluttered urban environments where traditional pulse radar would simply see a wall of noise.
The Pulse-Doppler interaction
A common misconception is that pulse compression only improves range resolution. In reality, it dictates how the radar handles moving targets through the ambiguity function. When using Linear Frequency Modulation (LFM), a Doppler shift in the returning echo causes a slight temporal shift in the compressed output. This 'range-Doppler coupling' means the radar might report a target slightly closer or further than its true position depending on its velocity. While this sounds like an error, engineers exploit this predictable shift to simplify the hardware, as a single matched filter can often detect moving targets across a wide frequency range without retuning.
Conversely, phase-coded waveforms like Barker or polyphase sequences are highly sensitive to Doppler shifts. If a target moves fast enough, the phase alignment required for compression breaks down, causing the signal to decorrelate. This 'Doppler intolerance' is why modern multi-mode radars dynamically switch between LFM for long-range volume search and phase-coding for precise tracking of stationary or slow-moving objects. By managing these trade-offs, radar systems achieve high sensitivity while maintaining the ability to distinguish a stealthy cruise missile from background clutter and weather interference.
Hardware evolution: From SAW filters to DSP
In the mid-20th century, pulse compression was a grueling hardware challenge. Before high-speed digital processors, engineers used Surface Acoustic Wave (SAW) filters to achieve the necessary dispersion. These analog components literally slowed down parts of the signal using physical quartz or lithium niobate crystals, creating the 'chirp' effect through acoustic propagation. The 1970s marked a turning point as these rigid, physical templates were replaced by Charge-Coupled Devices (CCDs) and eventually the first high-speed digital-to-analog converters, allowing the radar to change its compression ratio on-the-fly to adapt to jamming environments.
Today, the Heavy lifting is done entirely in the digital domain. Modern Field Programmable Gate Arrays (FPGAs) use Fast Fourier Transforms (FFTs) to perform convolution in milliseconds. This transition from fixed crystals to fluid software means a single radar can now emit a broad spectrum of 'LPI' (Low Probability of Intercept) waveforms. These involve pseudo-random sequences that resemble white noise, making it nearly impossible for an adversary's electronic support measures to recognize the signal as a radar pulse, effectively hiding the hunter while the hunter sees everything.