Demonstrators

SCD 1.2: Robotaxi

VIF

TTTECH Auto

The main target of this demonstrator is to develop and integrate edge environment perception and vehicle intelligence algorithms in a robo-taxi to handle three urban use cases: 1. Overtaking a vulnerable road user 2. Stopping in front of a crosswalk 3. Picking up a customer in a fetch-up zone A proof-of-concept for bidirectional communication approaches to the cloud (SCD1.3) is also implemented within this demonstrator. The main hardware platform for this demonstrator is a Ford Mondeo equipped with lidar, IMU and GNSS sensors, an LTE modem and a drive-by-wire kit. The used main software platform is RTMaps which enables to integrate C++ (control), Python (intelligence) and Julia (perception) modules and provides interfaces to sensors and actuators (e.g. CAN Bus). The core processing unit is a power pc with dedicated GPU next to a communication module (TTTech Razormotion).

The real world demonstrator will be used in a lot of other projects that focus on shared perception, road monitoring, micro public transportation to name a few. The results gained in the development process (e.g. fast prototyping with Julia, embedding in different software platforms) will be used within other projects. The developed control and planning module RomPac is in use in several other projects with a focus on Co-Simulation. The perception module will be integrated in other platforms used for autonomous driving such as VifWare (based on AutoWare). The gained competence in closed-loop simulation using Carla based on standards like openDRIVE and openSCENARIO and in future openMATERIAL will contribute to other projects where Co-Simulation, generation of labelled learning data or closed-loop simulation is relevant. The Virtual Vehicle research GmbH has a strong focus on the software defined vehicle and thus the results achieved within this demonstrator will provide useful assets.