Desktop simulation is widely used to make millions of virtual test kilometres for data processing, sensor fusion and to check ADAS controller performance and robustness. However, a human driver is essential for the final stages in the ADAS controller development process.
Once the ADAS controller has situational awareness of the simulated environment and is in a position to take decisions, the method and level of severity of the interaction with the driver needs to be evaluated. The only way to include the driver into the control loop without letting them drive a real car is to use a driver-in-the-loop simulator. Furthermore, using a driving simulator in the development of ADAS algorithms provides a good opportunity to evaluate various algorithms and strategies with a human in the loop, much more efficiently, before signing-off (with a prototype) on test tracks and public roads.
Desktop methods have a strong focus on highly detailed (semi-) physical simulation of the sensors, data reduction, sensor fusion and objectivation to feed the ADAS algorithms and test performance and robustness of the overall systems in standardised benchmark situations. A DIL simulator is then used to validate and test the effectiveness of ADAS controller performance in real world scenarios with a human driver in the loop.
To evaluate this closed loop performance, it is desirable to use relatively simple built-in ideal sensor models to achieve well-defined input and create situational awareness for the ADAS controller. This way, the focus of the DIL simulator experiments lies with controller development and analysis of the human-machine interaction. Sensor failure and/or erroneous controller behaviour can be manually injected by the operator to simulate the effect of real world disturbances and imperfections, however interfaces to advanced ADAS development packages like Vires-VTD and TASS-PreScan do also exist.
ADAS engineers also need reliable subjective feedback to understand how the driver copes with the fact that a controller is interfering with their vehicle. This cannot be accurately achieved with desktop simulation tools. Cruden simulators offer dedicated automotive motion-cueing, high-fidelity vehicle dynamics models, realistic graphics and cockpit mock-ups. This ensures that the driver is fully immersed in a believable environment, which leads to useful subjective-objective correlation for our ADAS customers.
Cruden’s ethos of open architecture offers the flexibility to adapt the company’s solutions to the client’s needs. Panthera is compatible with many external software packages and supports LiDar-sourced scenario databases with OpenDRIVE and OpenCRG. Cruden experts’ involvement with both the automotive and motorsport markets for over 25 years has led to a strong background in chassis development and a highly accurate vehicle dynamics model. This is key to achieving realistic ADAS simulations and the company is already working on a wide range of ADAS projects, in automotive and commercial vehicles with European and North American OEMs.
Cruden will be providing live demonstrations of its simulator, set up for ADAS and autonomous vehicle controller development, at all three days of the Automotive Testing Expo (Messe Stuttgart; June 20th-22nd 2017; stand 1169).
Visitors will be able to experience how easily an ADAS controller can be developed and tested with Cruden’s toolset, via a Cruden A406-F3 6-DOF motion-based simulator, complete with a three-channel projection system using top-of-the-line industrial grade projectors from Esterline and a 6m diameter screen. Cruden’s open architecture simulator software, Panthera, will simulate a typical selection of ADAS scenarios including lane keeping and active cruise control for visitors to test drive.