The expected results of HAVEit outline in more detail the measures to assess the successful achievement of each challenge of the project, as these are set by the HAVEit objectives.
a) Architecture: Safety architecture including redundancy management
- Development, integration and verification of a safety-critical control-platform which ensures scalability and configurability transfer to the software approach, data-base based configuration environment and data-base translation engine environment.
- Communication establishment including vehicle to vehicle link between vehicles following each other (distance between 1 m and 20 m), roadside (traffic light/sign) to vehicle broadcast (distance between 1 m and 20 m) and geo-networked information exchange and neighbor awareness.
- Fast, smart and reliable actuators, including steering actuators and brake-by-wire actuators.
b) Situation adaptable, optimized task repartition in the joint system: driver - co-system
- Co-system: One input channel of the safety relevant decision control unit is represented by the co-system command provided by a virtual driver integrating the ADAS functionalities. A perception layer by means of sensors and a data fusion module will provide an environment model to the knowledge layer. Similar to a driver this layer will be performed by a multi-agent system which integrates different control tactics for different situations. Success criteria at the different project stages:
- Driver in the loop assessment: The driver state information forms an important input for the later decision on the optimum automation level. Both direct (e.g. monitoring the driver vigilance with a camera system) and indirect measures (evaluating driver activities, such as history of pedal, steering activities, vehicle behaviors) will be combined to achieve the required high confidence level for the driver state information. Relevant dimensions are at one hand attentiveness, drowsiness and stress level of the driver, on the other hand the richness and the validity of the driver's mental model of the system. Methods and parameters to assess the degree the driver is in the loop will be developed.
- Joint system design and validation: The interaction between the elements of the joint system, especially between the driver and the co-system, will be designed using established guidelines, assessing the expectations of future users, and enabling a clear migration path from today's ADAS systems to HAVEit-systems and beyond. This includes the visual, haptic and auditory feedback to keep the driver in the loop and to bring him/her back step-by-step into the loop in case the highly automated system comes to its limits. Special emphasis will be given to the mode awareness and to safe mode transitions that lead to more or to less automation.
c) Safety architecture applications
- FASCAR - Extended joint system demonstration: A generic vehicle will be used for the validation of the 2E steering activities (drive-by-wire applications) to show extended possibilities for improved ergonomics.
- EWB-Truck - Drive-by-Wire heavy good vehicle for open roads: This vehicle will focus on the platform concept and the EWB actuators. The results of the architecture modeling will be applied to this vehicle with homologation constraints.
d) Highly automated driving in public traffic
- Automated queue assistance: by integrating and improving longitudinal and lateral support the level of automation is augmented and the system will therefore represent a significant step forward compared to state of the art. Suitable key indicators, describing the state of both driver and vehicle, will be defined and measured to ensure that the system works according to specifications and also to assess the impact on driver comfort and road safety.
- Automated assistance in road works and congestions.
- Temporary auto-pilot: this application will be able to handle dedicated scenarios autonomously, thus allowing the driver to be out of the loop for a few seconds. This option opens a completely new field of applications, as a couple of situations can be handled autonomously by the system, thus allowing the driver to withdraw from the primary driving task for a while, being sure that any change in the situation would be signaled to him. It must be guaranteed that in cases where driver invention is necessary – e.g. to avoid or to mitigate a collision – an immediate return to driver control is possible.
- Active green driving: this application aims at increasing the efficient energy usage. Impact on efficiency will be assessed by comparing the active green system performance in predefined use cases to a baseline system. This assessment can be performed by measuring a set of key indicators such as fuel consumption or change of battery state of charge for some specified use cases. These indicators and use cases will be precisely defined during the requirements and specifications phase of the project.