Everything under Optimum Control
- PDF / 265,123 Bytes
- 2 Pages / 595.276 x 790.866 pts Page_size
- 37 Downloads / 198 Views
CONTROL TECHNOLO GY
Everything under Optimum Control
Automated Functional 8 Development for Automatic
Tracking for 18 Trajectory Automated Driving Functions
Transmissions Using Deep Reinforcement Learning
have mechanical steering”
Andreas Homann, Torsten Bertram
Interview with Roland Arnold
[TU Dortmund], Martin Keller [ZF]
Katja Deuschl, Thomas Doster,
[Schaeffler Paravan Technologie]
Gerd Gaiselmann, Stefan Studer [Mercedes-Benz]
6
next generation of 14 “The vehicles will no longer
The first 125 years of automotive history were really simple: Vehicles were exclusively driven by human drivers who had 100 % control over the vehicle. However, the increase in the number of ADAS systems introduced over the past 15 years to control the vehicle in critical situations now requires clever control algorithms on the level of that universal tool: the human being. A step in this direction is represented by partially automated driving vehicles at SAE level 2 that already support the driver in many sub-functions. In the meantime, automotive development is already on the cusp of the next step toward high and full automation. This requires huge advances not only in the development of AI-controlled vehicle electronics, but also of the sensors and actuators required. If a hailed robotic vehicle is to be punctual bookable and safe to use, then the focus of development no longer needs to be placed on the vehicle itself, but above all on intelligent control technology. At Mercedes-Benz AG, one project focusses on the question of how artificial intelligence methods can be applied to develop and optimize gear shifts in automatic transmissions. The use of deep reinforcement learning can result in new control concepts and help to identify the potential for the optimization of existing control concepts. Automated vehicles with higher levels of automation will represent a significant component of individual mobility in the future. To do this, they need to acquire their environment, recognize situations and autonomously control the vehicle. The TU Dortmund uses the example of dynamic pre-control of the trajectory angle to describe how the interaction of actuators can be used to realize trajectory control as good as a human driver. © unlimit3d | stock.adobe.com
In an interview, CEO Roland Arnold of Schaeffler Paravan Technology discusses Space Drive – according to him, the first road-legal drive-by-wire system that replaces the mechanical transmission of steering, braking and acceleration commands with electronic signals. Alexander Heintzel
ATZ electronics worldwide 09|2020
7
Data Loading...
