In the last two and a half decades there has been an increasing body of work on the micro- Doppler effect for various applications. Researchers used micro-Doppler signatures to analyze, classify and detect human gait, hovering helicopters and wind-turbines, as well as jet engine modulation (JEM) to detect jet aircrafts. In recent years, the use of the micro-Doppler effect has expanded and taken also to the monitoring of biological signals. Researchers started investigating the use of the effect to extract vital signs such as breathing and heartbeat. The employment of various algorithms, such as the Chirp Z Transform and Fourier analysis, has been advocated. Multiple radars have been investigated in this context, from UWB and X-Band, to 24, 60, and 77GHz radar bands. The methods developed suffer from insufficient spectral resolution, as Fourier analysis type algorithms need a large time- window of data to support a certain resolution. Another caveat is the fact that the respiration frequency is lower than the heart-beat, while its amplitude is much larger coupled with inherent nonlinearities in the radar hardware, rendering the spectrum of the signal densely populated with harmonics of both the respiration and heartbeat and their inter-modulations. This tutorial will give an introduction to micro-Doppler with the derivation of the micro-Doppler effect of a vibrating target with a live-demo of radar based vital signs extraction from afar.

This guide was created with the help of AI, based on the presentation's transcript. Its goal is to give you useful context and background so you can get the most out of the session.

What this presentation is about and why it matters

This talk introduces the micro-Doppler effect in radar and shows how micro-motions — such as breathing and heartbeats — appear in radar returns. The speaker explains the physical origin of these extra spectral components, sketches the key mathematics, and demonstrates real radar captures that extract respiration and heart rate from several meters away. For engineers working in DSP, sensors, or remote health monitoring, understanding micro-Doppler is useful because it connects signal models, spectral analysis, and practical system design: how waveform choice, sensing band, time–frequency resolution, array geometry and post-processing algorithms determine whether physiological signals can be detected and measured reliably.

Who will benefit the most from this presentation

• Radar and RF engineers who want to add physiological sensing or gesture recognition capabilities to existing sensors.
• Signal processing engineers and students learning applied spectral estimation, time–frequency tradeoffs, and model-based detection.
• Product engineers exploring non-contact vital-sign monitoring (baby monitors, elderly care, automotive occupant sensing).
• Researchers interested in micro-Doppler signatures for classification problems (drones, gait, gestures).

What you need to know

The talk assumes familiarity with basic radar and DSP concepts. Below are the compact items to review so you get the most out of the talk.

• Doppler basics: For a monostatic radar the Doppler frequency shift produced by a target moving with radial velocity $v_r$ is

  $f_D = \dfrac{2\,v_r}{\lambda}$

  where $\lambda$ is the carrier wavelength. This is the starting point for thinking about motion-induced phase changes.
• Baseband I/Q and phase: After quadrature demodulation the return is a complex envelope $z(t)$ whose instantaneous phase is proportional to range. Small oscillatory changes in range map to phase modulation and produce spectral sidebands.
• Vibrating-target model: A simple model of a vibrating point scatterer is

  $r(t)=R_0 + A_v\sin(\Omega_v t)$

  where $A_v$ is vibration amplitude and $\Omega_v$ is the angular vibration frequency. The phase term in the baseband becomes an FM-like term and can be expanded into harmonics using Bessel functions; that expansion explains the spectral lines spaced at integer multiples of the vibration frequency, i.e. the micro-Doppler.
• Harmonics and intermodulation: When two periodic motions coexist (e.g. respiration and heartbeat), the spectrum contains each fundamental and their harmonics, plus intermodulation products (sums and differences). Respiratory motion is low-frequency but large amplitude, heartbeats are higher-frequency but smaller amplitude — this amplitude separation plus hardware nonlinearity often creates dense spectral structure.
• Time–frequency resolution: Fourier-based estimators need sufficiently long time windows to resolve close frequencies. That creates the familiar tradeoff: higher spectral resolution vs. the ability to track time variations. The talk references algorithms (FFT, Chirp-Z, subspace methods like MUSIC) and shows how window length, sampling and aliasing affect what you can see.
• Practical considerations: Carrier band (24, 60, 77 GHz, etc.), CW vs pulsed waveforms, SNR, multipath, hardware nonlinearities and aliasing are all important. Array receivers add angular and spatial diversity, improving robustness and allowing multiple-person monitoring.

Glossary

• Doppler shift: Frequency change of the returned signal proportional to radial velocity.
• Micro-Doppler: Additional spectral components caused by small-scale vibrations or rotations superimposed on bulk motion.
• Radial velocity: Velocity component along the line-of-sight to the radar; it determines Doppler shift.
• Harmonic: Integer multiples of a fundamental frequency appearing in the spectrum due to nonlinear phase modulation.
• Intermodulation: Sum/difference frequencies produced when two or more periodic motions combine.
• CW radar: Continuous-wave radar that transmits a single continuous carrier; commonly used for micro-Doppler sensing.
• I/Q baseband: Complex representation of the demodulated radar return: in-phase (I) and quadrature (Q) components.
• Range–Doppler map: A 2D display where one axis is range (from time/gating) and the other is Doppler frequency (from spectral analysis).
• MUSIC: A subspace-based spectral estimator that can resolve closely spaced sinusoids better than classical periodograms under certain conditions.
• Bessel expansion: Mathematical series used to express an FM-like exponential of a sinusoid; it explains why phase-modulated motion produces discrete spectral sidebands.

A few words about the presentation

Nir Regev’s talk balances theory and practice: he presents a clear physical model (vibrating-range → phase modulation → Bessel harmonics) and then validates it with straightforward, relatable demos (phone vibration, seated breathing and heart-rate capture). If you appreciate talks that tie mathematical intuition to real captures and system-level constraints (bands, resolution, clutter), you’ll find this one both useful and motivating. Expect concrete examples you can build on for research or prototyping of non-contact vital-sign monitoring.