Completed

Radar Detector based on Transformer

TransRAD: Retentive Vision Transformer for enhanced 3D radar object detection

Radar Detector based on Transformer

Overview

Despite significant advancements in environment perception capabilities for autonomous driving and intelligent robotics, cameras and LiDARs remain notoriously unreliable in low-light conditions and adverse weather. Radar serves as a reliable and low-cost sensor that can effectively complement these limitations. TransRAD is a 3D radar object detection model designed to address these challenges by leveraging the Retentive Vision Transformer (RMT) to learn features from information-dense radar Range-Azimuth-Doppler (RAD) data. The approach leverages the Retentive Manhattan Self-Attention (MaSA) mechanism to achieve precise 3D radar detection across Range-Azimuth-Doppler dimensions.

Key Highlights

  • Autonomous Driving
  • Intelligent Robotics
  • Object Detection
  • Transformer Architecture
  • Low-Light Conditions

Methodology

Retentive Vision Transformer architecture with Manhattan Self-Attention mechanism for radar RAD feature learning. Location-Aware Non-Maximum Suppression for improved detection accuracy.

Technologies

mmWave Radar Retentive Vision Transformer Deep Learning Range-Azimuth-Doppler Location-Aware NMS

Project Details

Start Date
January 2023
End Date
August 2025
Status
Completed

Resources

Related Publications