Trustonomy has recently published the Deliverable D2.2 “Trustonomy Methodological Guidelines”.
This report contains the definition of the generic and specific guidelines for developing and producing the methodological frameworks of Trustonomy in the domains of Driver State Monitoring (DSM) assessment, Human Machine Interfaces (HMI) design assessment, risk assessment and ethics analysis, driver training, Driver Intervention Performance Assessment (DIPA) and trust/acceptance assessment, for autonomous road transport, taking also into account other studies and standards or best practices in the field.
This determines the questions to be investigated during the entire project, in order to refine the preliminary concepts and provide more detailed answers.
Secondly, the analysis of a set of 10 multimodal HMI designs was made regarding suitability and effectiveness and guidelines for identifying optimal combination of HMIs depending on the user and the context were produced. Furthermore, the ethically challenging questions for the Automated-Decision-Support framework were raised along with the beneficial effects that Trustonomy will have, including the optimization of the decisions and actions made by the autonomous systems of the car, as well as the maximization of the feeling of safety and trust in system.
During the analysis made for driver training, the need of a tailor-made ADS training was indicated and several useful guidelines to be used during the development of Driver Training Framework were described. The development of new curricula for Driver Training, or changes in the existing ones, will have multiple positive outcomes, additionally to the obvious result of the standardization of driver assistance systems naming and the scope of correct operation and warning. The main goal of Driver Intervention Performance Assessment (DIPA) is to ensure that the driver is able to take over the control of the vehicle and maintain its safety. Due to the nature of the problem, time is a critical parameter as calculations and decisions should be made immediately, so DIPA must work close to vehicle systems, assuring minimal latency of data transmission. Taking these (and several others) parameters into account, the presumed set of measures which will be considered in the research was defined.
Last but not least, guidelines for the measurement of trust and acceptance are provided. The trust and acceptance assessment is going to be based on four diverse types of data collected through driving simulations in a laboratory setup: physiological measures, which will provide the size/extent of trust as different driving scenarios unfold; driving performance measures, which will provide insight not only at a continuous level but also in reaction to specific events; vehicle performance measures; and subjective measures, elicited via questionnaires will provide measures of acceptance as well as trust.
Read the full Deliverable in our Library!
