AI-Based Vehicular Network toward 6G and IoT: Deep Learning Approaches

In recent years, vehicular networks have become increasingly large, heterogeneous, and dynamic, making it difficult to meet strict requirements of ultralow latency, high reliability, high security, and massive connections for next generation (6G) networks. Recently, deep learning (DL ) has emerged as a powerful artificial intelligence (AI ) technique to optimize the efficiency and adaptability of vehicle and wireless communication. However, rapidly increasing absolute numbers of vehicles on the roads are leading to increased automobile accidents, many of which are attributable to drivers interacting with their mobile phones. To address potentially dangerous driver behavior, this study applies deep learning approaches to image recognition to develop an AI-based detection system that can detect potentially dangerous driving behavior. Multiple convolutional neural network (CNN )-based techniques including VGG16, VGG19, Densenet, and Openpose were compared in terms of their ability to detect and identify problematic driving.

Traffic Crimes and risky driving: The role of personality and driving anger

The current study aims to examine the influence of personality traits (alternative Zuckerman model) and driving anger in the explanation of risky driving style in individuals convicted for road safety offences (N = 245), using as a basis an adaptation of the context-mediated model. This is a transversal, descriptive study designed to be implemented by means of surveys, in which took part 245 men convicted of road safety offences from five prisons in Galicia (a region in northwestern Spain) took part. The average age of the participants was 38.73 years (Sx-9.61), with a range between 18 and 64 years. All participants had three or more years of driving experience. Our data shows that the Impulsive-Sensation Seeking (Imp-SS) personality trait had a direct and positive effect on dangerous driving, while the Activity (Act) trait had a direct but negative effect. The Aggression-Hostility (Agg-Host) trait, in turn, influenced the risky driving style, but not directly, but by raising driving anger levels, so it acted as a powerful mediator between the Aggression-Hostility (Agg-Hos) trait and the risky driving style. In general, our research partially replicates and expands previous findings regarding the model used, the aggression-hostility personality trait (Agg-Host) was placed in the distal context, driving anger in the proximal context, while age and personality traits Activity (Act) and Impulsive-Sensation Seeking (Imp-SS) were direct predictors. The results of this study may have practical implications for the detection and rehabilitation of offenders and penalties for road safety offences.

Left-Turn Driving Behavior Based On Situational Anger: Day Or Night

Anger is a key factor affecting drivers' subjective judgment and driving skills. The influence of anger on driving behavior has been widely studied, but there is a lack of comparative research under different lighting conditions. Through driving simulation experiment, this paper studies the influence of anger on left-turn driving behavior under two light conditions of day and night. In the experiment, 32 licensed participants were divided into two groups, one in emotional neutrality and the other in anger. Among them, the emotional state of anger is induced by a traffic related video. The results showed that compared with daytime participants, participants at night had higher anger intensity, shorter gap acceptance and post encroachment time (PET) when left-turn driving. In addition, compared with emotion neutral participants, angry participants tended to accept shorter gap acceptance and post encroachment time (PET) when turning left. This indicates that participants failed to respond correctly to left-turn driving behavior in a state of anger. However, the response of gender differences to situational driving anger was not affected by light conditions. The anger intensity of male participants during the day and night was higher than that of female participants, and the gap acceptance and post encroachment time (PET) during left-turn were shorter than that of female participants. This shows that male participants are more likely to produce high-intensity anger and are more likely to make dangerous driving decisions in a state of anger. This paper puts forward some suggestions on the identification of anger and the prevention of angry driving.

Early warning system for drivers’ phone usage with deep learning network

Dangerous driving, e.g., using mobile phone while driving, can result in serious traffic problem and threat to safely. To efficiently alleviate such problem, in this paper, we design a intelligent monitoring system to detect the dangerous behavior in driving. The monitoring system is combined by camera, terminal server, target detection algorithm and voice reminder. Furthermore, we applied an efficiently deep learning model, namely mobilenet combined with single shot multi-box detector (mobilenet-SSD), to identify the behavior of driver. To evaluate the performance of proposed system, we construct a dangerous driving dataset which consists of 6796 images. The experimental results show that the proposed system can achieve accuracy of 99% in 100 testing images. It can be used for real-time monitoring of the driver’s status.

DriverSonar

Dangerous driving due to drowsiness and distraction is the main cause of traffic accidents, resulting in casualties and economic loss. There is an urgent need to address this problem by accurately detecting dangerous driving behaviors and generating real-time alerts. Inspired by the observation that dangerous driving actions induce unique acoustic features that respond to the signal of an acoustic source, we present the DriverSonar system in this paper. The proposed system detects dangerous driving actions and generates real-time alarms using off-the-shelf smartphones. Compared with the state-of-the-arts, the DriverSonar system does not require dedicated sensors but just uses the built-in speaker and microphone in a smartphone. Specifically, DriverSonar is able to recognize head/hand motions such as nodding, yawning, and abrupt adjustment of the steering wheel. We design, implement and evaluate DriverSonar with extensive experiments. We conduct both simulator-based and and real driving-based experiments (IRB-approved) with 30 volunteers for a period over 12 months. Experiment results show that the proposed system can detect drowsy and distraction related dangerous driving actions at an precision up to 93.2% and a low false acceptance rate of 3.6%.

PSYCHOLOGICAL AND PEDAGOGICAL ASPECTS OF DIAGNOSTICS TA SHLYAKHIV FORMUVANNYA VIDPOVIDALNOSTI SPECIFICATIONS MAYBUTNIKH FAKHIVTSIV

The article studies the psychophysiological features of drivers and their impact on road safety. The levels of psychophysiological “danger” and psychological characteristics of drivers are studied based on the analysis of their professional activity. The individual psychological features of drivers caused by disorders of mental regulation of their behavior are determined. The authors propose to improve the system of psychological training of drivers, taking into account the factors of “dangerous driving”.

Pilot study to evaluate school safety zone built environment interventions

BackgroundSchool safety zones were created in 2017 under the City of Toronto’s Vision Zero Road Safety Plan. This pilot study examined the effect of built environment interventions on driver speeds, active school transportation (AST) and dangerous driving.MethodsInterventions were implemented at 34 schools and 45 matched controls (2017–2019). Drivers travelling over the speed limit of >30 km/hour and 85th percentile speeds were measured using pneumatic speed tubes at school frontages. Observers examined AST and dangerous driving at school arrival times. Repeated measures beta and multiple regression analyses were used to study the intervention effects.ResultsMost schools had posted speed limits of 40 km/hour (58%) or ≥50 km/hour (23%). A decrease in drivers travelling over the speed limit was observed at intervention schools post-intervention (from 44% to 40%; OR 0.79, 95% CI 0.66 to 0.96). Seventy-one per cent of drivers travelled >30 km/hour and the 85th percentile speed was 47 km/hour at intervention schools, with no change in either postintervention. There were no changes in speed metrics in the controls. AST increased by 5% (OR 1.22, 95% CI 0.97 to 1.54) at intervention schools. Reductions in dangerous driving were observed at all schools.ConclusionsPosted speed limits were >30 km/hour at most schools and high proportions of drivers were travelling above the speed limits. There were reductions in drivers exceeding the speed limit and in dangerous driving, and modest increased AST post intervention. Bolder interventions to slow traffic are required to effectively reduce speeding around schools, which may increase safe AST.