Interaction of autonomous and manually-controlled vehicles

The early days of highly autonomous vehicles co-existing with traditional vehicles could present road safety risks and create unanticipated consequences as other drivers may be surprised by the different behavior of autonomous cars. Manually-controlled vehicles might expect that vehicles with automated capabilities behave in a similar way. To this end, autonomous vehicles algorithms need to take into account human behavior and if appropriate, be able to reproduce it. The Interaction of autonomous and manually-controlled vehicles (IAMCV) project will study the best strategies to ensure that manual driven vehicles understand the decisions made by autonomous vehicles and vice versa. To this end, it will focus on investigating the actions that occur between vehicles and humans and that influence one another by observing road utilization. This knowledge will make possible to provide automation algorithms with the necessary data sets to make decisions that consider the actions that occur between vehicles and humans and that have an effect upon one another. The key objectives of IAMCV are: O1) To find out how to show the information related to the decisions performed by the automated system in an efficient way to other road users. O2) To tailor the automation according to the human behavior by studying intervehicular interaction and being able to predict the response of other road users. O3) Promote trust in automation by demonstrating the technology as a reliable and safe. These goals will be achieved by analyzing road users` behavior from naturalistic driving and floating car data from vehicles equipped with sensors. Different levels of automation will be considered focusing on awareness and diagnosis of previously defined road situations. Several scenarios will be replicated under real and lab conditions (i.e. intersections with multiple vehicles and roundabouts) to perform the pertinent experiments and determine the impact of the coexistence of AVs and other road users on general road safety. Information related to the situation will be conveyed by the automation to manual drivers to determine if this knowledge generates a degree of confidence that can compensate mistrust. The IAMCV project will help to provide a solid scientific base for advancements in automated vehicles technology, building knowledge and expertise in the field of ITS; It will contribute to extending the standards and regulations for designing autonomous vehicles, considering road user needs aiming at increasing user acceptance and trust in Autonomous Vehicles (AV) by guaranteeing road safety.

The Interaction of Autonomous and Manually-Controlled Vehicles (IAMCV) project centered on investigating communication and coordination between these two types of vehicles. This involved exploring ways to improve safety, efficiency, and overall traffic flow by ensuring interactions and transitions between automated and manual driving modes. The ultimate goal was to leverage this information to develop advanced automation algorithms and systems that could make decisions based on the results obtained and to disseminate the results in an accessible, timely and structured manner to help minimize the acceptance barrier between road users and automation. Vehicle behavior was investigated by means of decision-making processes as a result of awareness, diagnosis of road situations, and trust in the automation. The key objectives of IAMCV were: O1) To find out how to show the information related to the decisions performed by the automated system in an efficient way to other road users. O2) To develop automation systems that anticipate and adapt to human behavior basing on the analysis of inter-vehicle interactions and how road users communicate and predict the responses of other vehicles and pedestrians on the road. O3) Promote trust in automation by showcasing its reliability and safety through demonstration The O1 objective was achieved by designing intuitive visual displays and communication methods that effectively conveyed the decisions made by the automated system to other road users, ensuring clarity and safety. The O2 objective was met by developing decision making algorithms that anticipated and adapted to human behavior based on the analysis of inter-vehicle interactions. By understanding how road users communicate and predicting responses from other vehicles and pedestrians, the systems were designed to interact more intelligently and predictably in traffic scenarios. Regarding objective O3 trust in automation was promoted through demonstrations that highlighted the system's reliability and safety features. Real-world testing and simulations revealed the system's consistent performance and adherence to safety protocols, building confidence among users and stakeholders. The most representative achievements are listed below: Development and analysis of the IAMCV dataset for analyzing inter-vehicle interactions and advancing autonomous driving algorithms utilizing various sensors to capture diverse real-world driving scenarios. Exploring complex driving scenarios and interactions between autonomous and human-operated vehicles that revealed human behavior patterns and their importance for automated driving systems. Methodological advancements in data analysis, modelling, and simulation such as survival analysis and robust tools techniques for trajectory clustering algorithms, data analysis, modelling, and simulation within automated driving. Aligning subjective safety perceptions with objective safety metrics guiding the creation of algorithms and systems that better align with drivers' perceptions and behaviors. Providing practical insights and tools for autonomous vehicle enhancement to increase road safety, efficiency, and effectiveness.