Created by Dr. Ing. Evans Baidoo.
Copyright © 2026.
- Signal Processing
- Automotive Radar Perception
- Multi-sensor Fusion
- Real-time algorithm deployment
This project implements an end-to-end FMCW radar signal processing pipeline, closely aligned with industrial automotive radar systems.
It simulates radar data, processes it through the full DSP chain, and performs target detection and spatial estimation, similar to production radar perception stacks used in ADAS and autonomous driving.
- FMCW radar signal simulation (multi-target, multi-antenna)
- Range estimation via Fast-Time FFT
- Velocity estimation via Slow-Time Doppler FFT
- Clutter suppression using:
- Mean cancellation (MTI)
- 2-pulse canceller (classic MTI filter)
- Adaptive detection using 2D CA-CFAR
- Angle estimation:
- FFT-based beamforming
- Conventional beamforming (switchable)
- Multi-dimensional outputs:
- Range Profile
- Range–Doppler Map
- Range–Angle Map
- Modular plotting system
Raw ADC Data (Simulated Radar Cube) ↓ Range FFT → Distance Estimation ↓ MTI / Clutter Suppression ↓ Doppler FFT → Velocity Estimation ↓ Range-Doppler Map ↓ CFAR Detection → Target Extraction ↓ Angle Estimation (FFT / Beamforming) ↓ Range-Angle Map
- Power vs Range (from beat frequencies)
- Range–Doppler Map with CFAR detections
- Range–Angle Map
- Velocity spectrum per target
- Multi-target tracking (Kalman / JPDA)
- Sensor fusion (Radar + Camera)
- Real radar data integration
- GPU acceleration
- C++ / pybind11 port for real-time deployment