Decentralised Traffic Incident Detection via Network Lasso

TL;DR

This paper tackles traffic incident detection in data-distributed ITS environments by reframing the problem as a decentralised one-class SVM task and solving it with Network Lasso (NL). By constructing a fused graph that encodes geo-spatial and road-network similarities and optimising a population of OC-SVMs via ADMM, the approach achieves guaranteed global convergence while leveraging information from similar regions. Empirical results on the California PeMS dataset show that NL outperforms centralised and local baselines and is competitive with federated learning methods, particularly in detection rate and AUC, with robust performance across network scales. The work demonstrates that classical ML techniques, when distibuted through NL, can offer strong, privacy-preserving, edge-friendly AID with practical impact for scalable ITS deployments.

Abstract

Traffic incident detection plays a key role in intelligent transportation systems, which has gained great attention in transport engineering. In the past, traditional machine learning (ML) based detection methods achieved good performance under a centralised computing paradigm, where all data are transmitted to a central server for building ML models therein. Nowadays, deep neural networks based federated learning (FL) has become a mainstream detection approach to enable the model training in a decentralised manner while warranting local data governance. Such neural networks-centred techniques, however, have overshadowed the utility of well-established ML-based detection methods. In this work, we aim to explore the potential of potent conventional ML-based detection models in modern traffic scenarios featured by distributed data. We leverage an elegant but less explored distributed optimisation framework named Network Lasso, with guaranteed global convergence for convex problem formulations, integrate the potent convex ML model with it, and compare it with centralised learning, local learning, and federated learning methods atop a well-known traffic incident detection dataset. Experimental results show that the proposed network lasso-based approach provides a promising alternative to the FL-based approach in data-decentralised traffic scenarios, with a strong convergence guarantee while rekindling the significance of conventional ML-based detection methods.

Figures (6)

• Figure 1: The NL-based framework for detecting traffic incidents on modern traffic scenarios featured by data-decentralisation.
• Figure 2: The graph design
• Figure 3: The selected region from California I-80 freeway, blue dots represent local nodes, created via open-street-map library
• Figure 4: The accuracy metric for centralised OC-SVM, Fedavg AE and the proposed framework in each local node ordered by node id
• Figure 5: The performance metrics and computation time for different network scale using fused traffic graph in proposed framework