For this work, the UPCT autonomous vehicle (UPCT-CICar [23 (
link)]), was driven by a human pilot in manual mode. CICar is a real-world prototype, based on a commercial electric vehicle, the Renault Twizy, which has undergone a series of modifications to provide it with the required functionality. The CICar has been equipped with multiple sensors, including LiDAR, cameras, IMU, GPS, encoders, etc., necessary for the vehicle to perform autonomous driving tasks. This platform setup integrates a perception system, a control system, and a processing system on board the vehicle.
Perception System: The purpose of a sensor system is to collect data from the surrounding environment of the AV and send that data to the control system. These sensors measure different physical quantities, which are typically selected to overlap each other, providing the redundant information needed to correctly merge and correlate the information. In our autonomous vehicle, two types of sensors are used to measure the environment: short-range sensors (up to 10 m) and long-range sensors. Installed short-range sensors include a Sick 2D laser ranging scanner and time-of-flight camera. The long-range sensors are a 3D LIDAR scanner and a camera in the visible spectrum.
Table 3 and
Figure 3 show the different devices involved in data acquisition during the tests, as well as the details of the variables involved in obtaining them.
Driver Biometric System: The drivers’ biometric signal collection system has been carried out using a non-invasive wearable device, bracelet type, called Empatica E4. The Empatica E4 is a wrist-worn top-quality sensor device considered a Class IIa Medical Device according to 93/42/EEC Directive. Empatica E4 device measures the acceleration data (ACC), as well as other physiological parameters, namely the Blood Volume Pulse (BVP), from which the Heart Rate Variability (HRV) and the Inter-Beat Interval (IBI) are derived as well, skin temperature (TEMP) and also changes in certain electrical properties of the skin such as the Electrodermal Activity (EDA). For the creation of our dataset, among the several measurements recorded by the Empatica E4, this signal was considered, since it provides information better suited for activity recognition. A summary of the technical specifications of the accelerometer sensor is detailed in
Table 4.
Control System: The main control systems of the Renault Twizy have been automated in order to allow the vehicle to be autonomously controlled. The modified systems are the steering wheel, the brake pedal and the accelerator pedal (see mechanical modification in
Figure 3). Despite the fact that all driving will be manual and not autonomous, the system will record the data with two controller drives through a CAN bus. The Compact Rio cRIO 9082 controls the accelerator, brake and steering wheel movements with the CAN-Open communication protocol, as well as I/O signals.
Processing System: Each sensor works with its own sample rate, and in most cases, this is different between devices. The achieve the synchronisation of the data and accurately reconstruct the temporal sequence, time stamps have been generated to synchronise the operating start and finish times. All of this is controlled and synchronised by the on-board processing system.
Rosique F., Navarro P.J., Miller L, & Salas E. (2023). Autonomous Vehicle Dataset with Real Multi-Driver Scenes and Biometric Data. Sensors (Basel, Switzerland), 23(4), 2009.