An Efficient Traffic Incident Detection and Classification Framework by Leveraging the Efficacy of Model Stacking

Abstract

Automatic incident detection (AID) plays a vital role among all the safety-critical applications under the parasol of Intelligent Transportation Systems (ITSs) to provide timely information to passengers and other stakeholders (hospitals and rescue, police, and insurance departments) in smart cities. Moreover, accurate classification of these incidents with respect to type and severity assists the Traffic Incident Management Systems (TIMSs) and stakeholders in devising better plans for incident site management and avoiding secondary incidents. Most of the AID systems presented in the literature are incident type-specific, i.e., either they are designed for the detection of accident or congestion. While traveling along the road, one may come across different types of traffic incidents, such as accidents, congestion, and reckless driving. This necessitates that the AID system detects and classifies not only all the popular traffic incident types, but severity as well that is associated with these incidents. Therefore, this study aims to propose an efficient incident detection and classification (E-IDC) framework for smart cities, by incorporating the efficacy of model stacking, to classify the incidents with respect to their types and severity levels. The experimental results showed that the proposed E-IDC framework achieved performance gains of 5%–56% in terms of incident severity classification and 1%–14% in terms of incident type classification when applied with different classifiers. We have also applied the Wilcoxon test to benchmark the performance of our proposed framework that reflects the significance of our approach over existing individual incident predictors in terms of severity and type classification. Moreover, it has been observed that the proposed E-IDC framework outperforms the existing ensemble technique, such as XGBoost used for the classification of incidents.

An Incident Detection Model Using Random Forest Classifier

“…The incident detection systems have evolved from non-automatic detection systems that relied on receiving phone calls from eyewitnesses or reports from traffic operators upon observing incidents [16] to Automatic Incident Detection (AID) systems. Such systems consider traffic data collected from the road continuously or simulated by traffic simulator software and analyze them to extract traffic patterns as either normal or abnormal [17]. The development of these AID systems has been ongoing since 1970's and continues to the present day [18][19][20][21][22][23].…”

“…To achieve this, Machine Learning (ML), a subset of AI, is used, which models the optimal mapping between a set of inputs and a set of outputs [69]. The ML algorithms used in AID systems include Random Forest (RF) [40,70], Artificial Neural Networks (ANN) [44,[71][72][73][74][75][76][77][78][79], Fuzzy Logic (FL) [80][81][82][83][84][85], Support Vector Machine (SVM) [28,86,87], or a combination of multiple models [17,29,30,35]. These algorithms have shown promising and superior performance in AID systems by learning from the input data and deducing the pattern without relying on a predefined mathematical models [65][66][67][68].…”

Real-Time Incident-Responsive Signal Control Strategy under Partially Connected Vehicle Environment

“…Most researchers obtain incident data from various traffic incident record systems from both freeways and arterial roads, where different types of incidents and their durations were recorded. Studies on freeway incidents include various incident types, such as accident, breakdown, and debris [33]; accident and disabled vehicles [34]; accident, congestion, and reckless driving [35]; collision, debris, disabled vehicle, and abandoned vehicle [36,37]; and crashes, hazard, and stationary vehicles [38,39].…”

Logical Structure of Actions to Reduce the Impacts of Traffic Incidents

“…It is also noted that the actions proposed through literature review seek to reduce the amount or duration (response time) of traffic incidents; however, they all need to involve all the interesting parts in order for really effective solutions. An example brought by Iqbal et al (2021) is that faster incidents detection provides timely information to passengers, generating better traffic management, in addition to helping other stakeholders (hospitals and rescue, police and insurance departments) in their respective responses. It is also noteworthy that the accurate classification of these incidents in relation to type and severity helps interested parties to develop better plans for managing the incident site and preventing secondary incidents.…”