As if by magic, the truck transports its freight from one place to another at the port. A driver cannot be seen. The truck drives autonomously. A trained long-distance driver sits in a control center and controls the vehicle remotely in an emergency. This is what the logistics of the future could look like – at least when it comes to the Chair of Vehicle Technology at the Technical University of Munich (TU). At a performance show, the employees of the chair showed their approaches in the field of autonomous mobility. Start-ups were also represented, sharing their innovative ideas with the public. At a total of six different stations, visitors from politics and business were able to get an impression of the research at the TU.
A Garching team won the Indy Autonomous Challenge (IAC).
(Photo: Indy Autonomous Challenge)
Research begins in motorsport
High speeds, fast reaction times, different weather conditions and no road markings: In motorsport races, the driver’s attention is required. The TU is currently researching how driverless vehicles behave in such situations and where their limits lie. New algorithms are constantly being created that contribute to solving the problem. For example, how an autonomous vehicle should behave when overtaking. Simon Hoffmann from the AV Teleoperation team at the TU will demonstrate the result on a simulator at the performance show. There you can see how humans and the AI software compete with each other in a race.
Representatives of the TU were able to gain some experience at the “Indy Autonomous Challenge” in Indianapolis in 2021. In the racing series, teams from international universities competed against each other with robotic racing cars for the first time and competed in different disciplines. Although the team won the competition in the end, the race was the trigger for a new research approach at the Chair of Automotive Engineering.
The focus of the event in the USA was cooperation and data exchange between the universities. “We want to continue to channel the development and innovations of autonomous vehicles in the future,” says Simon Hoffmann from the AV Teleoperation team. Therefore, the TU wants to develop open source software that is accessible to all researchers, the algorithm of which will later serve as a platform for self-driving cars and trucks.
Phillip Karle explains the inner workings of the autonomous research vehicle “Edgar”.
(Photo: Florian Peljak)
Premiere for the autonomous Edgar
“Edgar” – that’s the name of the new autonomous research vehicle of the TU. The name of the blue and white van is derived from “Excellent Driving Garching”. Edgar has a number of features that set it apart from traditional minibuses. This is what makes autonomous driving possible at all. Phillip Karle from the AV Perception team explains: “The eyes of the vehicle are cameras. They perceive, for example, traffic signs, traffic lights and people. A laser sensor scans the surroundings in detail. However, this reaches its limits in bad weather, which is why an additional radar sensor is necessary We have installed antennas for GPS reception and the Internet.” According to Karle, all the data comes together in a computer located in the trunk of the vehicle. There, the perception of the environment, decision-making and action are calculated. The computer then passes on how the vehicle should move. “In order to make the development realistic, the autonomous minibus is to be tested in downtown Munich as early as next year,” adds Karle. “One year later, in 2024, Edgar will even drive around the Bavariaring at the Oktoberfest as an autonomous taxi shuttle.” In addition to the racetrack, the TU can also create algorithms for road traffic. In addition, the employees of the Chair of Vehicle Technology want to use the minibus to implement complete open source software for autonomous vehicles.
Research partner from Ingolstadt
One of the first research partners is the Technical University of Ingolstadt. In the field of autonomous driving, she focuses on critical urban scenarios and the expansion of intelligent infrastructure. In the course of the exhibition, the university showed different sensor systems and the associated software. “From data from infrared cameras and laser scanners, the basis for algorithms is created via AI-based real-time object recognition and sensor data fusion, which enable safe automated driving even in critical urban scenarios,” explains Andreas Huber from the press office of the TU.
Nothing works in autonomous driving without cameras and sensors.
(Photo: Florian Peljak)
When the robot car goes on strike
Autonomous driving functions often reach their limits. In such cases, teleoperated driving offers a reliable safety solution. Frank Diermeyer, Head of Teleoperation, gives an example: “If there is a branch on the road, the autonomous vehicle cannot continue its journey. A human operator is called in from a distance, who assesses the situation and decides whether the car can continue its journey may continue or not.” Remote-controlled driving is therefore a key technology for tomorrow’s mobility. According to Diermeyer, it will play a central role in the introduction of autonomous vehicles. The team at the Chair of Vehicle Technology is currently concentrating on researching safe, efficient and intuitive solutions for teleoperated driving. At the event on the TU campus, visitors had the opportunity to ride in the teleoperated research vehicle.
researchers and company founders
Two start-ups emerged from the research work of the TU on autonomous and teleoperated driving: Driveblocks and Fernride. Driveblocks launched six graduate students from the first 2021 Indy winning team. Among them Alexander Wischnewski and Stephan Matz. At the exhibition, the two presented an autonomously driving truck that is controlled by electronic signals. To ensure that the truck is controlled, the founders are developing a modular and scalable platform for autonomous driving. This includes various algorithms for object recognition, environmental perception, sensor fusion and decision making. The possible areas of application for this platform are diverse, for example on the motorway for transport between logistics centers, in a container terminal or in a mine.
From years of research in the field of teleoperation at the TU, three founders founded the start-up Fernride in 2019. There they are working on the implementation of remote-controlled electric vehicles. Teleoperators can already take over control from afar in difficult situations. For example, if the vehicle’s software is overwhelmed. But the start-up is pursuing a larger goal: “Fernride has set itself the task of solving the problems of the transport industry,” says co-founder Jean-Michael Georg von Fernride at the exhibition. An example: in Europe alone there is already a shortage of more than 400,000 truck drivers. Teleoperations could compensate for this number.
The special approach is the gradual autonomy through teleoperation. According to Georg, this is how drivers become teleoperators who can fully control their vehicles remotely. “As the winner of the German Mobility Award, the company now has partnerships with well-known car manufacturers, more than 40 employees and investments in the tens of millions,” explains Andreas Huber from the TU press office. “50,000 to 100,000 trucks are already teleoperated at ports, factory premises or in logistics centers.”
The exhibition was also about the transfer of knowledge to the economy.
(Photo: Florian Peljak)