Road departure is one of the main causes of single vehicle and frontal crashes. By implementing lateral support systems, a significant amount of these accidents can be avoided. Typical accidents are normally occurring due to unintentional lane departure where the driver drifts towards and across the line identifying the edge of the lane. The Lane Support Systems (LSS) uses cameras to “read” the lines on the road and alert the driver if the car is approaching the lines. Anyway, despite the assumed technology readiness, there is still much uncertainty regarding the needs of vision systems for “reading” the road and limited results are still available from in field testing. In such framework the paper presents an experimental test of LSS performance carried out in two lane rural roads with different geometric alignments and road marking conditions. LSS faults, in day light and dry pavement conditions, were detected on average in 2% of the road sections. A decision tree method was used to analyze the cause of the faults and the importance of the variable involved in the process. The fault probability increased in road sections with radius less than 200 m and in poor conditions of road marking.

Internet of vehicles (IoVs) allows millions of vehicles to be connected and share information for various purposes. The main applications of IoVs are traffic management, emergency messages delivery, E-health, traffic, and temperature monitoring. On the other hand, IoVs lack in location awareness and geographic distribution, which is critical for some IoVs applications such as smart traffic lights and information sharing in vehicles. To support these topographies, fog computing was proposed as an appealing and novel term, which was integrated with IoVs to extend storage, computation, and networking. Unfortunately, it is also challenged with various security and privacy hazards, which is a serious concern of smart cities. Therefore, we can formulate that Fog-assisted IoVs (Fa-IoVs), are challenged by security threats during information dissemination among mobile nodes. These security threats of Fa-IoVs are considered as anomalies which is a serious concern that needs to be addressed for smooth Fa-IoVs network communication. Here, smooth communication refers to less risk of important data loss, delay, communication overhead, etc. This research work aims to identify research gaps in the Fa-IoVs network and present a deep learning-based dynamic scheme named CAaDet (Convolutional autoencoder Aided anomaly detection) to detect anomalies. CAaDet exploits convolutional layers with a customized autoencoder for useful feature extraction and anomaly detection. Performance evaluation of the proposed scheme is done by using the F1-score metric where experiments are carried out by exploiting a benchmark dataset named NSL-KDD. CAaDet also observes the behavior of fog nodes and hidden neurons and selects the best match to reduce false alarms and improve F1-score. The proposed scheme achieved significant improvement over existing schemes for anomaly detection. Identified research gaps in Fa-IoVs can give future directions to researchers and attract more attention to this new era.

With the significant technological growth that affected autonomous vehicles in the last decade, several consequences occurred as: human factor exclusion, entry and exit manoeuvres precision from roundabouts, and headway reduction. In this paper, it was carried out a microsimulation approach study that aims to evaluate benefits in terms of safety obtained with flower roundabouts in a scenario where traffic is characterized by conventional vehicles “CVs” and Connected Autonomous Vehicles “CAVs”. This study focused on the evaluation of CAVs and CVs operation with the presence of the so called “weak users” or rather, pedestrians and bikes. Then, simulated scenarios were characterized by the presence of zebra-crossings in main roads, positioned at 20 m from circulatory carriageway edges. Micro simulation choice is due to the absence of survey data collection because the presence of CAVs in ordinary traffic is still minimal. The micro simulation was carried out through VISSIM, so it was operated with a specific methodological path, consisting, in the application, of O–D matrix based on real cases, in order to achieve an assessment of potential conflicts in relation with the increase in CAVs. Simulation results showed that higher safety levels were achieved for special cases of O–D distribution and with CAVs present. Finally, considering crash absence in results related to CAVs presence, safety interventions of such roundabout types have to be thorough. There were 10 O/D matrices analysed through VISSIM considering parameters as: average tail length, maximum tail length, average speed, vehicles, and number of stops quantity. As reported in the conclusion section, O/D matrices that showed minimum conflicts and maximum dynamic performances were identified.