Digital twins have become somewhat of a buzzword in the past couple of years. But what exactly are they? A digital twin, as its name indicates, is a non-physical copy of a physical object. Just like a digital scan of a physical picture. This virtual element enables a real-time view of all relevant data coming from said object. Depending on the system being studied, specific sensors can be tracked and monitored. This allows for the replication of the system’s environment (adherence of the road, weather, surrounding objects or systems, etc). In this blog post, we will discuss digital twins and their use cases in automotive.
Digital twins have become somewhat of a buzzword in the past couple of years. But what exactly are they? A digital twin, as its name indicates, is a non-physical copy of a physical object. Just like a digital scan of a physical picture. This virtual element enables a real-time view of all relevant data coming from said object. Depending on the system being studied, specific sensors can be tracked and monitored. This allows for the replication of the system’s environment (adherence of the road, weather, surrounding objects or systems, etc). In this blog post, we will discuss digital twins and their use cases in automotive.
For automotive, the value of using digital twins lies mostly in running simulations. It’s easier (and cheaper) to simulate crash tests, autonomous driving and other scenarios in a simulator, rather than using physical vehicles.
With the use of artificial intelligence and machine learning (AI/ML), the virtual twin also can help identify issues before they appear on the physical twin. This makes it possible to apply fixes to the physical twin before any problem occurs in real life. Let’s explore the use cases in more detail.
Digital twins in the automotive industry can be used during the system design phase in multiple fields, from vehicles to robotic arms. From a vehicle perspective, digital twins allow for more reliable vehicle design and development.
Consider electric vehicles (EVs). Modelling energy consumption for new prototypes is very important. Having a clear view on how a vehicle behaves – from the battery management system (BMS) to the wheel and tyre pressure efficiency – allows engineers to perfect its design. Being able to enhance the placement of wires while limiting the thermal or electromagnetic impacts, can help reduce the weight of the vehicle as well as its cost. It’s true that the battery’s state of charge is the first element that comes to mind when thinking about power consumption. But vehicle aerodynamics have a huge impact on power consumption (in all vehicles and not just EVs). Thanks to computational fluid dynamics simulations, the vehicle’s aerodynamics can be highly optimised. This is how OEMs can obtain the lowest drag coefficient.
Digital twins can also help to optimise manufacturing flows. OEMs and suppliers have to take the whole supply chain, including manufacturing constraints into account, to streamline operations. From a factory point of view, the design of robotic arms, the development of the supply chain and conveyor belts are highly critical. Companies can also simulate their supply chain using extended digital twin models and running AI/ML models to test different scenarios. Being able to anticipate the best positioning of sensors while designing factory machines can save time during use but also increase savings related to component and material optimisation. Furthermore, companies can simulate their supply chain using extended digital twin models and running AI/ML models to test different scenarios.
Digital wins in the automotive industry can also be used during the serial life phase.
Thanks to digital twins, it’s possible to simulate autonomous driving (AD) algorithms using AI/ML computations in real time. Indeed, verifying AI/ML algorithms using a simulated environment allows engineers and developers to know if they are safe. Once said algorithm has been tested in a digital environment imitating the real world for over tens of thousands of kilometres, then it can be applied to a physical prototype. While physical testing takes time, it’s possible to speed up simulations and run them in parallel in order to generate thousands of hours of driving while keeping a realistic simulated environment with applied gravity, weight and physical collision anticipation.
Digital twins allow for better monitoring of such computations and can help identify specific scenarios that require more in-depth simulations. For example, some AD real-life scenarios are extremely difficult to reproduce but can go a long way to fine tune sensors and algorithms. Physical prototypes won’t go away, but having a digital model of a real physical sensor (ie. camera, lidar, etc) makes it unnecessary to run scenarios that were unforeseen during development in the open world. Not only does it allow for huge savings, it also limits the risk of accidents with other vehicles, pedestrians, etc.
Digital twins that use real-time vehicle sensor data make predictive maintenance achievable. Many automotive companies use them today in order to monitor any sensor (let’s say the airbag deployer), and obtain a status on the wear and tear of any part of a vehicle or factory machine. This allows for huge savings (no more downtime, stock and resource anticipation). It minimises the risk of accidents, whether it’s in the factories or from vehicle defects, and provides constant knowledge on the status of each critical (safety-related) and non-critical system element.
As you can see, digital twins offer vast potential for automotive companies at different stages of the vehicle’s lifecycle.
In the development phase, they help OEMs lower costs by simulating and optimising the vehicle’s conception, wiring, weight, aerodynamics and overall structure – instead of testing each of these variables on a physical prototype. In the manufacturing phase, they help optimise equipment locations, maintenance, and required movements for each step of the building process. In the serial life phase, they can help anticipate wear and tear, defects, but also be used for replaying real-life scenarios encountered by a physical vehicle.
In this post, we’ve only covered the tip of the iceberg! In upcoming blogs, we will delve into real-life use cases that your company may probably be facing right now. For those interested in the technologies that power up digital twins, we will discuss how vGPUs can help with virtual desktop infrastructures (VDI), high performance computing (HPC) and advanced autonomous driving simulations. Stay tuned.
Curious about automotive at Canonical? Check out our webpage.
Want to learn more about software-defined Vehicles? Download our guide.
One of the most critical gaps in traditional Large Language Models (LLMs) is that they…
Canonical is continuously hiring new talent. Being a remote- first company, Canonical’s new joiners receive…
What is patching automation? With increasing numbers of vulnerabilities, there is a growing risk of…
Wouldn’t it be wonderful to wake up one day with a desire to explore AI…
Ubuntu and Ubuntu Pro supports Microsoft’s Azure Cobalt 100 Virtual Machines (VMs), powered by their…
Welcome to the Ubuntu Weekly Newsletter, Issue 870 for the week of December 8 –…