Driving simulators have always attempted to connect real people with imaginary vehicles. The fundamental principle is not new; however, the technology used to make this happen is a moving target. Since the 1990s, driving simulators have taken on many forms — from small-scale gaming/ entertainment systems to the large-scale systems used by vehicle manufacturers for their product research and development (R&D) activities. The variation in driving simulator designs has mostly come about owing to the variety of use cases and the adoption of emerging technologies. For example, the objective of some simulators is simply to provide display monitor graphics from the driver’s seat perspective to allow visual participation in a simulation, while others aim to engage all the senses with various motion systems, projection graphics and sophisticated audio immersion. While simulation has been an element of vehicle engineering for decades, Driver-in-the-Loop (DIL) simulators seek to advance the technology. To learn more, we spoke to Kia Cammaerts, founder of Ansible Motion.
The following is an edited transcript of the conversation.
S&P Global Mobility: Could you tell us a little about your business?
Kia Cammaerts: We founded Ansible Motion with the sole objective of designing and manufacturing engineering-class automotive DIL simulators. In 2009, we released the first commercially available, high-dynamic, onset-cued motion system designed specifically for ground vehicles. Originally our customers were in professional motorsports including F1 teams and manufacturer-backed or “works” teams. The company has grown to offer a comprehensive product portfolio of turn-key DIL simulators to vehicle manufacturers, tier one and tier two suppliers, research institutions and universities alongside other motorsport customers that now include World Rally Championship, Indycar, NASCAR, Formula E, Super Formula, WEC and GT racing. Our product range features compact, stationary variants to full-size, dynamic simulators including our most successful product to date, the Delta series S3. In 2022, we became part of AB Dynamics’ global automotive testing and simulation group.
What are the main use cases for DIL simulators?
You can find DIL simulators in use around the world at automotive original equipment manufacturers, tier 1s, tire manufacturers, high-level motorsport organizations, universities and research facilities. The applications and use cases are extensive and include all the traditional areas of automotive product development that you might imagine.
Fundamental chassis development — suspension, steering, tires, ride, etc. — as well as active safety, advanced driver assistance systems (ADAS), and autonomous applications. There are several published customer case studies in the vehicle dynamics arena regarding the use of our simulators.
Other popular areas these days are noise, vibration and harshness (NVH), powertrain and especially electric vehicle (EV) development. The motorsport sector is still one we supply to (for race preparation and setup optimization), but much more of our business today revolves around road cars. Most of our recent projects have had elements of ADAS, active safety and autonomous vehicles, and customers are increasingly realizing that DIL is useful for human machine interface and infotainment user interface development as well as traditional human factor experiments.
There are also several applications that might be categorized as “specialized.” These include advanced mobility (research space related to exploring new types of vehicles and roadway/infrastructure systems), connected systems (a deeper dive into the vehicle-to-everything space, where the focus is on onboard vehicle communication systems and their interactions with other vehicles and infrastructure), human-vehicle interaction (research into human physiology, biomechanics and high-level, man-machine interfacing) and university research, which can be highly specialized (one example is the car sickness study we did with Vrije Universiteit [VU] Amsterdam‘s Department of Behavioral and Movement Sciences, research organization TNO’s Department of Perceptual and Cognitive Systems and UK’s Coventry University’s Center for Mobility and Transport).
DIL simulators can help sort out these issues and more, and experiments can be conducted repeatably and safely, within the confines of a controlled lab environment, so there is no risk to people or machinery. Infinite environment and scenario possibilities can be explored as well — many of which might be impossible or expensive to replicate in the real world.
For example, what happens when it is raining? What happens if there is a tire puncture? What happens if there is a control system or sensor malfunction, and the failsafe mode is engaged? In a DIL simulator lab, these cases are just a few computer keystrokes away.
The aforementioned specialized applications represent the front lines of automotive research, so it is interesting to imagine those areas growing and for new research areas to emerge in the coming years — all related to the concept of advanced mobility.
Increasingly the industry is looking for solutions to solve more than one area of validation, to be applied across multiple vehicle development areas such as vehicle dynamics; NVH; powertrain and e-mobility; ADAS and active safety; human factors; connected and automated vehicles, or CAV, as well as high-performance and motorsports uses. Customers that use our dynamic Delta series simulators as well as our static Sigma and Theta series simulators confirmed the technology serves multiple functional groups across various focus areas.
Use cases for simulators grow day by day. As more engineers become exposed to the ability and capability of today‘s simulation tools, they naturally find more use cases. Ansible’s prime objective is to connect real people with virtual vehicles and environments, wherever they want in the world. We want engineers to experience convincing, realistic virtual test drives. We have found that customers have no trouble identifying interesting use cases for DIL simulators — some of which we have never even imagined. DIL simulators are uniquely positioned to engage real people, real drivers and evaluators, with pre-prototype vehicles and onboard systems — and that is the value proposition. It is a powerful way to sort out development issues, often before any metal is cut, which may otherwise be missed.
Ultimately DIL simulation can be used as a tool to answer the ever-increasing number of what-if questions that simply were not explorable before. I think this is a use case that is not always considered.
How much fidelity do you need in your simulation of the road environment?
High-fidelity environmental content is best; this means both visual and engineering content. On the visual side, we are looking for high refresh rates — up to 240 Hz — along with highly realistic scene rendering. On the engineering side, we are looking for macro and micro road surface representations that can properly interact with tire models, as well as material and object physics that can interact with sensor models. In DIL simulation parlance, we work with “digital twin’” environments, such as those provided by rFpro driving simulation software.
How do your simulators take account of different pieces of hardware processing at different speeds?
We integrate all the software applications and hardware interfaces with our own AML DDB — Distributed Data Bus — environment. We developed this powerful, synchronous real-time computing environment with an open and modular architecture, and it effectively serves as a master, synchronous clock for the entire DIL simulator ecosystem. With a human in the loop at all times, it is, of course, necessary to maintain real-time execution across the board.
How do you set about simulating multiple inputs from multiple sensors?
Sensors — which can include any combination of real, hardware in the loop or software in the loop elements — are no different than any other input/output element, such as the primary vehicle or subsystem models. Much like tire models, sensors and sensor models require the presence of necessary environmental descriptions to function properly. For example, a lidar sensor used in the DIL simulator lab must be able to interact with a virtual environment rather than a real one — so the distance to virtual objects (from the sensor’s perspective) must be a part of the virtual environment’s data set.
Given that we are seeing multiple different sensors on a car, all with pros and cons, redundancy is key for OEMs. How do your simulators manage that?
Sensor redundancies and fail-safes are managed and interrogated in a DIL simulator laboratory much as they would be on a real car. The only difference — which is key for developing sensor fusion approaches — is that DIL simulator labs are by default, controllable and perfectly repeatable testing environments. Scripted virtual environments can include all the infrastructure, weather, traffic, etc. so development engineers have a much more comprehensive, explorable and safe sandbox than real-world testing could provide.
Could you give us a real sense of how much an OEM can save by using simulation versus real-world testing?
One good example involves Continental, which recently acquired our Delta series S3 simulator. It will enable Continental’s engineers to repeatedly and consistently test tires across a range of virtual terrains, locations, scenarios and seasons. The simulator is supporting the company’s goal to reduce real-world testing by up to 100,000 kilometers per year and use 10,000 fewer tires for development. Another example is from a long-standing customer Honda, which has said that Ansible Motion’s simulators advance its engineers’ understanding, support its young engineers to learn and allow it to observe how drivers interact with new technologies.
Could an OEM use a simulator to determine consumer preferences?
Cars are ultimately consumer devices, so it is well worth studying and understanding human behavior to deliver the best user experience. If humans are going to be experiencing the vehicles, you should be including them in the virtual test process to get the best user experience data to make decisions on any aspect of the vehicle.
There is no better, safer or more repeatable solution than a DIL simulator. People always interact in ways we cannot always predict or see in numerical data. You can peruse online forums or social media for a few minutes to get a sense of how intrusive Lane Keep Assist is found to be in some applications and the frustration that this evokes; however, by placing evaluators into direct contact with proposed intervention systems in a few virtual test-driving scenarios, you can pick up on such things before an OEM signs off on a vehicle. Additionally, it goes further than just subjective assessments; it is also an ongoing feedback loop with all the other inputs and how you interpret those through your senses.
If we look at autonomy, even when the task of driving is removed, you still have a human being interacting with a vehicle and deciding whether they are having a good experience and determining their preferences. Any tool such as DIL simulation offers a sneak preview of this interaction that will be extremely useful in developing the vehicle.
According to S&P Global Mobility’s Jeremy Carlson, associate director for the Autonomy practice, “Simulation is an important tool in the development of automated functionality, from discrete ADAS up to full-stack autonomous driving.”