Hardware-Accelerated 4-Way Traffic Management: An Adaptive Learning Control System using NeuralONE-Optimized YOLO26 and 4K Vision on Orange Pi 6 Plus NPU

Roger Cristian Ripa Arias; Gonzalo Rafael Carpio Ramos; Franco Alessandro Arenas Mamani; German Alberto Echaiz Espinoza · Zenodo (CERN European Organization for Nuclear Research) (2026) · DOI: 10.5281/zenodo.19464141

Resumen

This replication package contains the source code, simulation environments, and validation datasets for the paper "Hardware-Accelerated 4-Way Traffic Management: An Adaptive Learning Control System using NeuralONE-Optimized YOLO26 and 4K Vision on Orange Pi 6 Plus NPU". The repository is structured into four main components: Simulation Engine: Pygame-based kinematic simulator with IDM and Webster traffic models. Analysis and Validation: Python scripts for numerical validation and simulation logs. Physical Hardware Inference: Documentation and pipelines for compiling YOLOv26 into the CIX CD8180 NPU (.cix format) using the NeuralONE SDK. Integrated System: The final Edge IoT deployment code featuring the NEMA TS-2 18-state controller.

Instituciones: Universidad Nacional de San Agustín de Arequipa