Engineer's 3-layer MLP classifies automotive radar targets, Macro F1 hits 0.764

bruno_pinto90 · reddit · 2026-09-07

A radar signal processing engineer shared a full writeup of training a 5-class classifier (car/largevehicle/twowheeler/pedestrian/pedestriangroup) on RadarScenes point clouds.

• Method: per-scan 16-bin histogram input into a 3-layer MLP with class-weighted cross-entropy, based on the "Histogram-based Deep Learning for Automotive Radar" paper, single scan only.
• Data challenges: class imbalance for twowheelers and largevehicles, aggregated class semantics, and sequence bias from long slow tracks causing high F1 variance across folds.
• Ablations: bigger MLPs, alternative encodings, and different binning all mattered less than the train/val/test split itself (measured over 6 folds); per-instance statistics slightly degraded performance.
• Key finding: Macro F1 rises from 0.381 to 0.764 as detections per instance grow from 1 to 5 — sparsity is the bottleneck. Cars and pedestrians work best; twowheelers are worst, since velocity-compensated distributions overlap with pedestrians and a stationary two-wheeler is indistinguishable from one.
• Next steps: PointNet-style spatial encodings, multi-scan accumulation, and micro-Doppler.

Full report on GitHub.

Original post →

More from Embodied

• FocusVLA routes robot camera views by predicted interaction intent across task stages — iscas-cloud · 2026-10-26
• Pandora M1: first prototype of a physical device for distributed GPU monitoring — tarantulae · 2026-10-26
• $150 modular Linux handheld Mecha Comet opens preorders — pocket SSH and Claude Code included — Paimaamu · 2026-10-26
• Apple Unveils All-New iPad mini: OLED Display, A20 Pro, Biggest Update Ever — CurieuxExplorer · 2026-10-26
• Purdue's GPS-free autonomous drone tracks runners through dense forests using on-chip LiDAR AI — KordingLab · 2026-10-26
• Apple lists OLED iPad Mini for sale on its online store — etothet · 2026-10-26

Embodied channel →