This article is part of the special topic "Future Mobility".
Digitization The value of digital twin technology to autonomous vehicles
Technological innovation is forever changing mobility. Research and testing are costly and take a long time to complete. In response, a digital twin offers a simulated environment to drive forth development in vehicle manufacturing and testing.
When you see the same test drives of autonomous vehicles, you might think the innovation is fragmented and slow. However, much of the real work is happening behind the scenes in another world called a digital twin. Digital twin refers to a digital simulation or a virtual model of a product, process, service or functionality. It's used extensively across a range of sectors for a variety of purposes including visualization, monitoring, and 3d representation.
I recently spoke with Oleksandr Vasylkov, Delivery Director, Automotive at Intellias to find out more. He notes that one one of the critical use cases for simulation is modelling autonomous driving to ensure their safety on the road:
"According to research, autonomous vehicles would have to be driven more than 11 billion miles to determine their reliability. With a fleet of 100 autonomous vehicles being test-driven 24 hours a day, 365 days a year at an average speed of 25 miles per hour, this would take 518 years—about a half a millennium.
This is exactly where the simulation comes to help. And it seems like the only option that will make the AVs possible."
Digital twin uses emerging tech to conduct R&D at scale
The use of technologies such as IoT, 3D mapping, cloud computing, big data, and augmented reality makes it possible to run virtual tests of both software and hardware in the loop to gain an accurate representation of how car parts and components are going to perform. This can include how vehicles operate under different road conditions which are difficult to test at scale.
In the US, rFpro has created a highly accurate digital twin of Mcity, a purpose-built facility at the University of Michigan for testing connected and automated vehicles and technologies under controlled, realistic conditions. The digital twin consists of a wealth of different roadside objects and materials to ensure accurate simulation of radar, LiDAR and camera sensors and enables vehicle manufacturers and technology suppliers to accelerate development using simulation.
A digital twin can also be used in automotive manufacturing and how various car components perform together. Oleksandr explained:
"Different companies are involved with the development of different electronic units with a vehicle, and sooner or later, you have to integrate them. And it becomes even more challenging when this development goes in parallel like for example, two different parts of the vehicle should be developed from scratch, and they should work with each other and interact with each other and the only option to make these is a simulation. So, we have conceptual software in the loop as well as the relevant hardware."
Intellias is currently working on a project with a Tier 1 OEM that simulates the whole environment, including the map, roads network, traffic, pedestrians, weather conditions. They are also focused on the vehicle itself, including the development of autopilot, computer vision, and self-learning algorithms as the safety integrity level as well as modelling and testing of mapping, infotainment systems, and V2X communication.
OEMs such as Tesla have heavily invested in the capabilities of a digital twin. It's been reported that Tesla creates a digital twin of every car it sells. This creates a capacity to detect problems and test software solutions. These can be turned into updates which can then be patched through to the vehicles.
The challenges of digital twin
Oleksandr notes that "It's not so complicated if we're just upgrading one piece of a car that already exists, because you have all the infrastructure and it's easy for you to make most simulations. In this instance, you already know the exact behaviour of the component or whole it shouldn't behave and also what hardware is available. But the majority of challenges starts when OEMs introduce a new platform with a new generation of a vehicle. In this case, almost all the hardware and software is being renovated. They are often looking for new suppliers, and you are in the position of having to make a new generation of software from scratch."
The value of modelling and digital twin for cross mobility platforms
Technologies such as digital twin are also used to plan and test transportation systems, explore mobility options, and overall city planning. In the US, the Open Mobility Foundation (The OCF)was launched in June 2019 to explore how digital twin models can enhance urban mobility. An example is an emergency vehicle that needs to respond to a crash by travelling from the local firestation to the location of the crash. Immediately, a routing simulation is run on the virtual city with its representation of current traffic. Once the ideal route is decided, the vehicle can be directed through both the real and virtual world. In the virtual world, intersection signals are notified just in time for the emergency vehicle to advance through the route and each signal's real-world counterpart is then set to green. The OCF suggests that in the future, digital proxies for autonomous vehicles will be notified of the changing context so that their physical twins can take appropriate action.
India's newest city, the Amaravati Smart City, is being developed in a digital platform which includes advanced mobility and traffic monitoring and simulations, as well as digital zoning and predictive traffic analytics. The use of digital twin simulation is an opportunity to see how a city moves across both verticals, spaces, and services and as all the different components that make a city livable connect and function in collaboration and alongside each other.