Mon, 29 Oct 2018
In a world where AI is taking more control in human decision making, systems supported by digital twins can prove game-changing for the automotive industry. Digital Twins are key to the long-term success of autonomous vehicles in the future.
Digital Twins are used in the automobile sector for creating the virtual model of a connected vehicle. They can capture the behavioral and operational data of the vehicle and analyze the overall vehicle performance, delivering a personalized service for customers.
The digital twin is either viewed as an imitation of a model or as a sum of connected components. Engineers delve into AI to design the ideal automotive product way before it enters the production line. Simulations can be analyzed to see why and how future problems such as breakdowns occur. Using digital twins for autonomous vehicles will potentially save many unplanned costs and miles through road testing and maintenance.
Collaborating with operational teams and analysts is ideal for gathering data to start modeling processes. It enables engineers to observe how a vehicle will perform in any condition through gathering data on types of motor, suspension, structure, aerodynamic body they want to tap into, and the materials they are built from. This is why we are hosting an event called the automotive digital twin workshop 2019 with some of the biggest automobile manufacturers in the world where we talk about the most effective methods of incorporating virtual counterparts into automobile manufacturing to support this idea.
The next step would be setting up a system for the data to automatically be transmitted back to the manufacturer. The manufacturer can analyze the data and perform predictive analysis to help make driving experiences smarter and safer for car owners. Engineers are also able to predict performances of products within large systems such as; an individual wing on a plane, a rocket engine launching, an office building maintaining energy through the day, A race car engine which is about to burn, or even a driverless car navigating the road during rush hour can create a digital twin of every single autonomous vehicle it sells, enabling them to analyze how a car performs in its physical environment, and track the vehicle from its creation to the day it goes to the junkyard. Much of the sensor infrastructure is already in place in newly released vehicles with them containing up to 100 sensors monitoring critical systems. This data can be combined with design information so predictive analysis can be implemented. An example of where this could be useful is for the modeling and monitoring of airbag systems. Defects would be spotted far sooner, and many potential injuries would be avoided.
While being highly intelligent, digital twins are not fully autonomous and will require human intervention – we have come to this conclusion after executing a few studies at Challenge Advisory that helped us realize this. The manual testing of new features and modifications of physical assets through virtual replicas are a particular example. AI, while not necessarily providing more intelligence than humans, is certain to boost skills by implementing more efficient and productive analysis. While the automotive industry has the relevant processes to move towards further implementing digital twins, they will still need the ability to include individually derived data.
The full benefit spectrum of Digital Twin, a technology that is capable of simulating the future and the present is very vast and in-depth. Apart from its ability to eliminate failure conditions for automotive manufacturers, it can also help greatly enhance the shopping experience of car buyers. How you might ask? Well, the improvements can be categorized into 3 parts:
The car purchasing stage is by far the most important part of the entire automobile manufacturing industry. It is the final step that determines sales volumes and profit and if there is a way that simulation technology could serve it, it would be to eliminate all the possible friction from this process. For instance, local automobile providers tend to agree that car purchasing is an emotionally intense process for some people. Most individuals who are looking to get a personal vehicle already have a specific vision in mind regarding what they are looking for, however, car dealerships often create a context where prospects are faced with multiple variants and decisions. Digital Twin could help enhance the experience by manufacturing a simulation that could accurately represent how a specific model of a car takes place in an individual’s life. For instance, developing a virtual twin of the car model and putting it into a virtual representation of a person’s property (e.g. the garage) could help him bring life into his vision about the product. Simulating a virtual road, the car model and the person itself would help him visualize how he would appear driving the automobile, visualize the car’s handling and speed. By utilizing the ability of VR-based virtual twinning, the person would also be able to experience how it will feel like being in his future vehicle, get the hang of all controls and equipment. The alternative we have now is the ability to test drive a vehicle prior to purchasing it, however, due to the limited nature of this experience, prospects are not able to get a personalized view for every single car and model they would like to try out. Digital Twin allows a maximized personal experience for dozens or hundreds of car models to simulate in a matter of seconds.
Once the final buying experience is over, the benefits and features of digital twinning do not end there. Throughout the entire lifespan of a car, the technology can serve its purpose for monitoring the health of all parts, processes, and activities that happen in the engine, providing in-depth details and updates about the well-being of the vehicle to the driver. Although this feature has been implemented by quite a few car manufacturers out there such as BMW, Audi or Volkswagen, the overall result of monitoring automobile health is not a great success. Best case scenario, a few select car manufacturers have managed to use machine learning to predict the source of an error or issue, however, nobody has come close to monitoring engine work to extreme accuracy like Digital Twin platforms can.