Skip to main content
Version: ACS CC

Phase-Correlated Doppler Velocity Sensor

A custom 24GHz continuous-wave radar system for velocity tracking (speed and orientation) using STM32, Embassy Rust, and Digital Signal Processing.

info

Author: Banila Rares-Sebastian
GitHub Project Link: https://github.com/UPB-PMRust-Students/acs-project-2026-rares567

Description

The system detects and tracks the movement of walking targets (humans) in its field of view, determining both their speed and their spatial orientation (Angle of Arrival). It achieves this by emitting a 24GHz microwave signal and capturing the reflections using a dual-antenna receiver. The microvolt-level phase-shifted signals are filtered and amplified by a custom 4-channel analog front-end (~1000x gain) built to strict low-noise tolerances. An STM32 microcontroller samples these 4 channels simultaneously via a hardware-triggered ADC and DMA to preserve phase correlation. Finally, the MCU runs a Fast Fourier Transform (FFT) algorithm to calculate the velocity vectors and displays the target's speed and angle on a digital screen.

Motivation

I chose this project as it provides a challenging combination of analog electronics (through working on a protoboard and soldering components manually) and embedded programming (through working on a STM32 MCU for digital signal processing). This makes it a great learning experience and a wonderful and precise sensor with utility in any human-related tracking or statistics.

Architecture

Diagram

Log

Week 7 - 13 April

Ordered components for the hardware from Sigmanortec and DigiKey.

Week 14 - 20 April

Started work on soldering components on the protoboard (op-amps, voltage regulator and adjacent passive components).

Week 21 - 27 April

Completed work on the first stage signal amplifier and successfully tested one channel.

Week 5 - 11 May

Week 12 - 18 May

Week 19 - 25 May

Hardware

The hardware features a custom Analog Front-End (AFE) bridging the 24GHz K-LC7 sensor and the STM32. Because the raw radar signals are in the microvolt range, the AFE utilizes LMV772 low-noise op-amps in a two-stage active amplifier configuration. By placing precision passives (1% resistors and C0G capacitors) in the negative feedback loops, the circuit forms an active low-pass filter. This actively suppresses high-frequency environmental noise while applying a massive 73 dB (~4,467x) voltage gain, boosting microscopic reflections to a readable 0–3.3V range. To guarantee signal integrity, an LT1763 ultra-low-noise LDO provides clean 3.3V power, driving a voltage divider that establishes a 1.65V virtual ground. This biases the op-amps so the AC radar waves can swing symmetrically without clipping above 3.3V and below GND.

Schematics

Schematics

Bill of Materials

DeviceUsagePrice
Nucleo-U545-RE-QMicrocontrollerLab provided
RFbeam-K-LC724GHz dual-receiver radar sensor233.6 RON
4x LMV772MA/NOPBLow-noise operational amplifiers40.76 RON
LT1763CS8-3.3Low-noise LDO linear voltage regulator33.08 RON
Display OLED 0.96" I2CI2C OLED screen for displaying output16.96 RON
Protoboard 7x9cm FR4Protoboard for creating the custom amplifier5.76 RON
5x SOIC-8 to DIP-8 adapterAdapter for converting SOIC-8 to protoboard-compatible DIP-84.95 RON
8x MFR-25FTE52-1M1M OHM 1% resistors1.32 RON
4x MFR-25FTF52-100R100 OHM 1% resistors0.66 RON
4x MFR-25FRF52-10K10K OHM 1% resistors0.69 RON
2x MFR-25FTE52-470K470K OHM 1% resistors0.33 RON
5x C320C104J5R5TA7301100nF X7R capacitors6.04 RON
9x ECE-A1EN4R7U4.7uF 20% capacitors10.18 RON
8x C317C101J1G5TA7301100pF C0G capacitors8.84 RON
4x MFR-25FRF52-22K22K OHM 1% resistors0.69 RON
105BPS100M1uF 20% capacitor0.99 RON
C322C103J1R5TA730110nF X7R capacitor1.86 RON
ECE-A1VN100UB10uF 20% capacitor1.86 RON

Software

WIP

LibraryDescriptionUsage
st7789Display driver for ST7789Used for the display for the Pico Explorer Base
embedded-graphics2D graphics libraryUsed for drawing to the display
  1. RFbeam K-LC7 datasheet
  2. RFbeam K-LC7 typical signal amplifier
  3. RFbeam K-LC7 radome design
  4. Embassy Framework Documentation