Towards Real-Time Heavy Goods Vehicle Driving Behaviour Classification in the United Kingdom

Abstract

Determining the driving styles and the factors causing incidents in real time could assist stakeholders to promote actions and develop feedback systems to reduce risks, costs and to increase safety in roads. This paper presents a classification system for Heavy Goods Vehicles (HGVs) drivers, using a core set of driving pattern stereotypes which were uncovered from driving incidents across three years i.e. 2014, 2015 and 2016. To achieve that, the driving stereotypes are established by employing a 2-stage ensemble classification framework followed by a profile labelling algorithm to define the set of driving stereotypes. Very similar stereotypes are later merged to form the core driving stereotypes for UK HGV drivers. Upon establishing core driving stereotypes across these three years, a decision tree classifier learns the classification rules to identify the driving stereotypes for the HGV drivers in a new dataset. High accuracy is achieved, indicating that the core driving patterns uncovered in this work could potentially be employed to identify UK HGV driving patterns in real-time.

Contextual Intelligent Decisions: Expert Moderation of Machine Outputs for Fair Assessment of Commercial Driving

“…Contextual factors identified from the literature [3], [11], [24]- [27], [41] (see Fig. 7) were presented to the experts in the workshops for validation.…”

“…Deep learning approaches applied to driver-facing footage, for instance, have shown promising results in automatically identifying some of the affective states, such as, distracted or attentive driving [43], [44], different types of human emotions [45], [46], and tired or energetic [47], [48]. Alternatively, calm or aggressive driving is accurately detected using telematics incident data [2], [3], while more complex affective states such as confidence and insecurity are still difficult to detect.…”

HSDDD: A Hybrid Scheme for the Detection of Distracted Driving through Fusion of Deep Learning and Handcrafted Features

“…To avoid potential safety risks, there is a need to monitor, detect and predict early the human driver inattentive behavior, and necessary warnings should be provided to the driver by accurately classifying a distracted driving state and nondistracted driving state. Researchers have attempted to mitigate this problem by leveraging AI for developing driver assistance and alert systems by understanding risky driving behaviors [5,6]. Detection of a distracted human driver can be categorized into two types of measurement strategies, measuring visual behavior of human driver and measuring vehicle related features or vehicle dynamics [7].…”

Contextually Intelligent Decisions: A Co-designed Approach for Commercial Driving Assessment

“…The current literature on intelligence-supported road safety assessment of commercial driving are limited to the manner by which drivers operate vehicle controls (Figueredo et al [2015(Figueredo et al [ , 2018, Mase et al [2020a], Agrawal et al [2019], Mehdizadeh et al [2021], , Hébert et al [2021], Satrawala et al [2022]) and do not consider the impact of inevitable contextual factors on driving performance, such as individual drivers' physical and mental states, weather conditions, traffic conditions, road geometry, road types, and work schedules. These factors influence drivers' responses, and therefore need to be considered to better understand the circumstances that led to a driver's performance or/and produce context-specific driving risk assessments.…”