AI models predict risky driving hotspots in Sydney using vehicle data

A new research paper explores the use of connected vehicle data to proactively identify and forecast risky driving hotspots in Greater Sydney, Australia. The study quantifies risky driving through thresholds for hard braking, harsh acceleration, and cornering, creating spatio-temporal heatmaps to pinpoint high-risk zones. Researchers benchmarked eight predictive models, finding that traditional time-series methods like ARIMA performed comparably to deep learning models such as LSTM, especially with limited data volumes. The findings suggest that IoT-based vehicle data can support targeted road safety interventions, with specific areas like Sydney's CBD, Parramatta, and Bankstown identified as persistent high-risk zones. AI

IMPACT Demonstrates the application of time-series and deep learning models for proactive safety interventions using IoT data.

RANK_REASON Academic paper detailing a new methodology for predicting risky driving using ML models. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• Adriana-Simona Mihaita
• Arima
• Australia
• Bankstown
• CBD
• exponential smoothing
• LightGBM
• long short-term memory
• N Beats Neural Basis Expansion Analysis For Time Series Forecasting
• Parramatta
• Prophet
• random forest
• Sydney
• XGBoost

• paper
• product
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →