The easiest way to count pedestrians, cyclists, and vehicles on edge computing devices or live video feeds.

Application of a convolutional neural network (CNN) to accurately classify urban sounds in a bid to increase pedestrian safety using the UrbanSound8k dataset.

Our open-source academic research files, codes, data, and so on.

Official Dataset and Models for "PedestrianQA: A Benchmark for Vision-Language Models on Pedestrian Intention and Trajectory Prediction"

Vision Zero Report Card — 7 years of NHTSA FARS truck-involved fatalities across 19 US cities. Reproducible Python pipeline + filtered dataset under CC BY 4.0.

Privacy-preserving pedestrian crossing safety analytics (computer vision, aggregate-only, GDPR/AI-Act by design). Anonymous road-safety dashboard demonstrator - Poland.

LiDAR-based obstacle detection system providing directional audio warnings for visually impaired pedestrians

Satellite imagery + deep learning for crowd density analysis; upload imagery, mark obstructions, and get ML-powered pedestrian flow heatmaps for event safety planning

Arduino UNO R4 WiFi 기반 스몸비(보행자) 안전 경고 장치 — 버스/지하철/C-ITS API + RCWL/SR505 센서 4모드

This Unity-based VR simulation explores how distance and deceleration rates affect pedestrians' perception of vehicle slowing. Participants observe a virtual vehicle decelerating at various speeds and distances, with trials conducted in three environments: Air, Ground, and Road.

Open road-hazard and near-miss data pipeline with exposure-normalized rates, uncertainty intervals, spatial hotspot analysis, GeoJSON exports, and a nationwide NHTSA FARS evidence atlas for safe-streets advocacy.

Advanced multi-modal system for detecting falling humans using LWIR, NIR, and Ultrasonic fusion.

Open methodology and datasets behind the SafeStreets walkability score.

NoBonk. Your phone watches where you're going, because you won't. On-device AI (YOLO) warns distracted walkers before they hit people, walls, or potholes. Nothing recorded, ever. Built by an 8th grader using claude/codex/android studio; ACSEF 3rd place, STEM4All 1st place + IEEE award.

A modern web application for visualizing and analyzing pre-collision pedestrian data.

Automatically analyzes sidewalk images and generates a pedestrian safety score aligned with civil engineering walkability standards.

Circulate San Diego: Infrastructure Data Explorer (2023)

Claude Agent Skill: a trauma-informed guide for cyclists and pedestrians hit by a car. Based on the White Line guide.

Download, filter, and aggregate NHTSA FARS fatal-crash data with zero non-stdlib dependencies. Companion to the Vision Zero Report Card.