Instrumented PTWs. Based on the findings of the literature review and on the specific kinematics of PTWs, the project partners defined a common set of signals regarded as necessary for the observation of rider behaviour. In addition, each partner was free to include other sensors, capable of enriching the recorded information, while maintaining an unobtrusive instrumentation. The list of basic common in all four studies vehicle signals includes: • linear acceleration (three components); • roll, yaw and pitch angles; • longitudinal speed; • brake activation; • throttle position; • steering angle; • GPS position; • turn signals; • video: 2 cameras positioned to capture the frontal environment (required a minimum 90° field of view) and the rider’s face. Possible additions to the basic set of sensors could include: • pressure of the brake circuits; • rear and front wheel speed. During the brainstorming phase for the definition of the relevant signals, the possibility to install a radar or lidar unit, to derive quantitative measurements of the surrounding environment (e.g. distance from the leading vehicle), was considered. The final decision did not include these signals and the related sensors because: • the limited free space for the installation on the front of PTWs; • the power requirements, which could create problems regarding the electric power balance of the vehicles. The research group did not consider this information to be vital for the naturalistic riding study, although a future implementation may. However this decision will be revisited after the data processing phase, to make a final assessment about the necessity of this information and provide more precise guidance for future studies. Data acquisition of all signals is set to be recorded at an accuracy of 100Hz except for video signals which are sampled at 10Hz and GPS position which can be sampled at 1Hz (values indicate minimal requirements). The specification for sampling frequency is derived from typical riders’ reaction time measured independently by INRETS during previous research and by the PISa consortium (www.pisa- xxxxxxx.xx): an average value is 0.3-0.4 s and thus, for an appropriate instrumental description of the reactions, 15 signal samples are reputed necessary. The limitation for video frequency is set based on practical requirements for on board data storage. The rationale behind the selection of signals will ensure that the following data are captured: • riders’ evasive manoeuvres (necessity fo...
