News

Since the start in autumn 2015, Revere – Chalmers' vehicle laboratory – has established itself as a full-fledged research infrastructure for academia and the automotive industry in western Sweden. Originating from road traffic research, the scope of Revere now also includes marine vessels and electrified vehicles.

Self-driving vehicles, active safety and vehicle dynamics are the three areas in focus for Revere, Resource for Vehicle Research at Chalmers.

“Vehicle researchers and companies in the transport sector are welcome to contact us with their ideas and projects to get help to transform them from theory to reality”, says Fredrik von Corswant, Director of Revere.

In collaboration with Revere, technologies, theoretical models and algorithms can be developed and tested on real vehicles and in real traffic environments, or on the test track AstaZero.

“We write research applications in collaboration with researchers and industry partners, enabling a tailor-made test phase that gets the most out of the results”, he declares.

Full-scale vehicles and models
In the lab at Lindholmen in Gothenburg, Revere has several vehicles of different kinds that are used in the research activities, for example a full-scale truck as well as cars and also smaller radio-controlled model vehicles. The smaller vehicles are often used in the teaching of students.

By adapting the vehicles to the current projects and providing them with different types of sensors and equipment for data logging, communication technology and more, data can be collected and the theories refined and verified.

“Our flexible, in-house developed software platform OpenDLV is our greatest asset”, says Fredrik von Corswant. “Only imagination sets the limits for what it can be used for in vehicle research, I would say. For example, it is possible to connect remotely to the vehicles while test driving, in case there is something that you immediately want to adjust in the software. Another advantage is that the platform can handle large amounts of data and is able to compress video, without losing data that is important to the algorithms.”

Real tests give more reliable results
Revere often works with long vehicles and combinations of vehicles, for example a tractor unit that is connected to two semi-trailers and a converter dolly forming a High Capacity Transport, HCT vehicle. Such a combination is longer than what is normally allowed to drive on public roads in Sweden today. In the vehicle lab, research and tests are performed on, for example, how to stabilize long vehicle combinations so that they do not tip over.

“We are also developing protocols for the communication between vehicle units”, Fredrik continues. “For example, if the tractor uses the brakes, a connected electrical dolly should not continue to propel forward.”

Conducting tests in real life provides supplementary and more reliable information, than computer simulations alone can supply.

“Only full-scale tests cover all the factors that affect the vehicle. For example, there are often disturbances that interfere with signals from sensors, inertia due to the weight of the vehicle, and also effects such as delays in the system, limitations in computer capacity, and more. Tests can provide answers to questions concerning how sensors act in bad weather or how the grip of the tires on the road surface behave depending on road conditions. This is often very difficult to determine with theoretical models alone.”

Revere's software platform can also be used to perform simulations. Real traffic data can be mixed with simulated data. For example, how would a self-driving car behave if there is an unexpected object like a dustbin, or other obstacle, on the road?

“We also compile datasets, such as film sequences, that are collected and created by our sensors in traffic situations of various kinds. We then share the data as open source with those who want to test their algorithms.”

Data created in the test vehicles is automatically uploaded to Revere's cloud server. The latest addition is data from a bus, that in early 2021 will run in regular traffic between two cities in India. In connection with the cloud server, Revere also offers analyses of data in a computational cluster, which enables, for example, training of systems for machine learning.

At the lab, more humanistic aspects are also studied, such as research into driver behavior and how surrounding road users perceive the interaction with self-driving vehicles since there is no driver to make eye contact with.

Research also in marine settings
For a couple of years now, Revere has entered the field of marine vessels, mainly in collaboration with RISE. A pilot boat in the port of Gothenburg is available for research projects, and also a catamaran platform and some models.

“The sensor technology used does not differ much from land to sea, apart from the radar systems. We have transferred a lot of our existing systems into marine applications”, says Fredrik. “I think there is a great need and demand for research on automation in the marine settings, from academia as well as from the industry. This is definitely an area for further expansion in the future.”

A venue to be proud of
Looking back at the first five years of the lab, what is he as a Director most proud of?

“We have successfully carried out a number of demonstrations of cutting-edge research, and our in-house developed software platform stands out well compared to the equivalents being developed at vehicle companies”, says Fredrik von Corswant. “Today, Revere is a venue where researchers and developers from various organizations and disciplines meet. That provides exciting cross-border connections and creates ideas for new innovations.”

The demand for a vehicle research infrastructure remains stable for the foreseeable future. We have probably only entered the first phase of the societal development that self-driving cars and electrified vehicles of various kinds are bringing.

”I hope that Revere in the future can attract more researchers and companies to be active collaborating partners. Our goal is to continue to build competence in automation and active security in the region, to provide a good recruitment base for industry and research institutes”, concludes Fredrik von Corswant.

Examples of research projects

I-dolly, a self-propelled truck trailer without driver or car
Revere is testing, in collaboration with, among others, Volvo Trucks and researchers from Chalmers, an intelligent converter dolly, that has its own electrical propulsion and steering. The dolly autonomously transports trailers with containers the last kilometers from a distribution center to the end customer for unloading.

COPPLAR, a prototype car for safe navigation in complex inner-city environments
In collaboration with several companies and researchers from Chalmers, Revere has developed a test vehicle for research on various self-driving functions, with a special focus on urban environments and changing weather conditions. Self-driving vehicles that cooperate with each other enable more safe navigation in complex inner-city environments. Together with Ericsson, Revere also has performed a demo on the AstaZero test track to show how vehicles communicate with each other for safe passages through an intersection.

AutoFreight, extra long self-driving trucks for smarter logistics
Revere is working with about ten partners to develop solutions for self-driving trucks to travel from the port of Gothenburg to Viared, an undustrial area near Borås. Field tests are conducted on highway 40 using an extra-long vehicle combination (HCT) of nearly 32 meters, which enables the transport of two containers per truck compared to usually one.

Facts About Revere, Resource for Vehicle Research at Chalmers
Revere is a part of Chalmers' research infrastructure and is closely linked to the SAFER traffic research center. Other partners are Volvo Trucks and Volvo Cars. Region Västra Götaland contributes with funding.
Read more about Revere

For more information, contact
Fredrik von Corswant, Director of Revere
fredrik.von.corswant@chalmers.se

Text: Yvonne Jonsson