Autonomous Vehicle Multi-Sensor Fusion Workbench

Scenarios & Modalities

Preset Edge Case

Active Sensors

Fusion Architecture

Noise & Occlusion Stress

Fog Density 35%

Camera Glare / Blinding 15%

Radar Clutter DB 12 dB

LiDAR Point Drop 20%

Pipeline Active | Fusing FMCW + LiDAR + Camera

Ego Velocity: 45 km/h

Real-Time Diagnostics

Euclidean Error

0.18 m

Tracking mAP

94.2%

Covariance Ellipse Area

0.04 m²

Fusion Latency

14 ms

Fusion Status: EKF resolving camera opacity and LiDAR scatter. Total spatial position error reduced from 1.45m (camera-only) down to 0.18m using fused radar velocity vectors.

Sensor Modality Legend

Camera Bounding Box (RGB Visual)

LiDAR 3D Point Cloud Ground Hits

FMCW Radar Radial Velocity Vectors

Fused Track & Covariance Ellipse

Spatial Error History (m)