Figure 1. Steps to conduct a FOT according to the FESTA methodology.
The main objective of BERTHA project is to develop a scalable and probabilistic Driver Behavioural Model (DBM) based mostly on Bayesian Belief Network (BBN) algorithms. In order to validate new, human-like driving systems like this, before being incorporated into next-generation vehicles, the real scenarios to which they will be subjected must be properly understood. To this end, the Field Operational Tests (FOTs) arise as a crucial tool. In short, FOTs are large-scale testing programs aiming at a comprehensive assessment of the efficiency, quality, robustness, and acceptance of Information and Communication Technologies (ICT) solutions used for smarter, safer, cleaner, and more comfortable transport solutions, such as navigation and traffic information and advanced driver assistance.
Along with data obtained in driving simulators, the data collected during the FOTs will then be used to correlate the performance of the DBM, ensuring that the driving behaviour obtained is as human-like as possible from the early stages of the development process. The FOTs provide essential support for developing Autonomous and Connected Vehicles as these tests ensure that ADAS and autonomous driving systems can handle the complexities of real-world driving scenarios.
In BERTHA project, CIDAUT is coordinating the FOT execution taking as basis for the definition of the test methodology the FESTA Handbook. The FESTA methodology divides the FOT execution into three phases: preparing, using, and analysing, with different steps within each phase, as can be seen in the following diagram.
For the practical implementation of the FOT, CIDAUT is looking to replicate the Use Cases defined in BERTHA, which are the following: “Collision risk avoidance on highway”, “Insertion on highway”, “Pedestrian crossing in urban area”, “Left turn at urban intersection”, and “Pull back in on urban highway”. To this end, CIDAUT has identified several areas in Valladolid city and its surroundings where the Use Cases will happen in a natural way while being safe for the drivers.
Several data will be collected during the driving trials, which will involve between 30 and 50 drivers of different gender, age, driving experience and driving style. A test vehicle has been equipped with cameras, radar sensors, GPS device and an on-board diagnostics device to record in real-time the state of the vehicle, the driver, and their immediate surroundings.
Figure 2. Equipment for data collection installed in the test vehicle.
The work mentioned above is part of the BERTHA project, funded by the European Union under grant agreement no. 101076360, within the framework of the Horizon Europe programme. Further information about the project can be found at https://berthaproject.eu/