As automotive technology becomes more intricate, vehicle developers face ever greater challenges in managing testing and development in the real world. In recent years, driver-in-the-loop simulators have emerged as an important tool to help carmakers in facing up to this challenge.
A range of systems by companies such as Ansible Motion, Wirth, Multimatic and rFpro are increasingly being used by OEMs and tier-one suppliers as a cost-effective and environmentally friendly method of testing products before prototyping. Vehicle simulators have changed beyond recognition in recent times, with developments in responsiveness and graphics, as well as improvements in vehicle and road surface modelling fidelity. So today’s driver-in-the-loop (DIL) simulators are potentially much more powerful and useful tools.
The modern breed of DIL simulation tools can be used for the subjective and objective assessment of vehicle and component performance in a growing list of scenarios where a human driver is required, including the testing of passive chassis designs and steering systems, as well as for the testing and calibration of chassis, drivetrain and engine control systems such as traction control, stability control, torque vectoring, and advanced driving assistance systems (ADAS).
According to Chris Hoyle, technical director at rFpro, a driving simulator that is capable of faithfully reproducing vehicle dynamics means that systems may be tested in safety, in controlled and repeatable conditions. rFpro pioneered the use of driving simulators for the testing of vehicle dynamics in Formula One and NASCAR racing. And the firm now includes among its clients automotive manufacturers and tier-one suppliers such as Ford, GM, Honda, Toyota, Jaguar Land Rover, Hyundai, AVL and Bosch.
Driver in control
“A few of our customers are also running driver-in-the-loop simulators connected in real time to drivetrain dynamometers,” says Hoyle. “This allows a human driver to be in complete control, via a plant model of the vehicle under test, of the entire drivetrain mounted on the dyno. The Volkswagen emissions scandal has recently refocused attention on this use case. Human drivers behave very differently from the automated tests that perfectly follow speed profiles.”
Ansible Motion produces ‘engineering class’ simulators for users in the automotive engineering and racing sectors, including Ford and Caterham Formula One. Ansible’s technical director Kia Cammaerts agrees that simulators serve as excellent engineering tools for vehicle dynamics and on-board systems evaluation – which he says is the key point of interest for automotive manufacturers and tier-one companies.
Role in validation
He adds: “We know our simulators are being used for validating hardware and software during the development cycle. One area of recent interest is supporting the validation of active safety and ADAS systems. Our driver-in-the-loop simulators are trusted tools for testing stability control systems, electric power-assisted steering, collision mitigation braking, intelligent systems and cooperative driving systems. Really any development area where people need to be brought into ‘early and often’ contact with on-board systems.”
One company that uses DIL systems for engineering applications is McLaren Applied Technologies. In 2008, it decided to build its own simulator facility for commercial purposes – as well as to continue the development of the technology beyond its earlier use for F1 by sister company McLaren Racing. According to Mike Phillips, commercial director at McLaren Applied Technologies, the company’s racing heritage means that its DIL simulator technology is excellent at modelling highly dynamic systems – with suspension, chassis and steering being the most obvious applications.
He also reveals that the technology has been used by the company to develop kinetic energy recovery systems and hybrid powertrain systems.
“Our own car company, McLaren Automotive, uses the simulator regularly to down-select concept designs at an early stage and also to develop more mature technology as an alternative to vehicle testing,” he adds.
Another company that has been heavily involved in the use of DIL simulators is Italian outfit Danisi Engineering. Last year, it acquired two driving simulators in recognition of the growing complexity of modern vehicle dynamics studies. According to chief executive Giacomo Danisi, the first thing the company noticed was that DIL simulators enabled it to conduct a high-level evaluation of the quality of a vehicle model very quickly.
“Test drivers work as extremely efficient post-processors, and when driving a new model are able to identify many modelling inaccuracies within a few minutes compared to the hours it would take for an engineer to post-process the data,” he says.
“However, the real strengths are the ability to perform complex real-world scenarios involving driver actions, environment interactions and active systems intervention in a repeatable and controlled environment which is something practically impossible to plan and execute with a running vehicle.”
For Hoyle of rFpro, the greatest advantages of using DIL systems are the savings in cost and time and the reduced project risk achieved by conducting more testing earlier in the engineering process. This enables design decisions and platform architectures to be evaluated with a professional test driver years before committing to a physical design.
Containing costs
“At each stage of the vehicle manufacturing process, from design, through testing and then validation, the cost associated with making changes increases,” says Hoyle. “Any change identified by testing while the new vehicle is still at the model-based stage will be significantly cheaper to make than if the same change is only identified once you reach the prototype stage.”
For Phillips at McLaren, the single biggest advantage of using driver-in-the-loop systems is the speed of development they enable, especially when compared to the building and testing of physical components. In his view, they also allow engineers to “test ideas and make design commitments with confidence early in the vehicle design project”. He believes that is especially important where these decisions fundamentally affect the way a car will ‘feel’ to the driver, for example in steering, gearbox shift and ADAS interventions.
“Poor designs only really manifest themselves in the later-stage prototypes when it can be really expensive to reverse-engineer out of trouble,” he says.
That said, he admits that the use of DIL simulators doesn’t come without some disadvantages, which he says mostly revolve around a “poor vehicle model that doesn’t correlate well to data from rigs and vehicles”.
Danisi also highlights the fact that DIL systems are quite complicated to run and argues that, particularly for hardware-in-the-loop projects, users need to pay close attention to all the decisions made in the simulator set-up to evaluate their effects on simulation accuracy.
Although the ability to make software changes very rapidly is “welcome,” he also warns of the danger that users could get “lost among the changes” without a robust configuration control of both tests and the simulator itself.
Linking test equipment
Looking ahead, Phillips predicts that another operational advantage will come from integrated test systems whereby the DIL simulator is linked to other test equipment – a key reason why McLaren Applied Technologies has teamed up with test equipment provider MTS Systems to develop a next-generation DIL system.
“Imagine a DIL simulator linked to, and using streaming data from, an actual dynamic tyre rig or damper rig with the real hardware being controlled in real time,” says Phillips.
Beyond modelling and simulating the vehicle for DIL systems, Phillips also says McLaren is interested in the wider possibilities of human-in-the-loop design and optimisation.
“With our history of understanding, and exploiting, human perception within simulated environments, we are able to expand the application space beyond automotive and into health and wellness, general transport and consumer product design development,” he says.
“All of this is in the pursuit of optimising a process or product around human interaction. We think the combination of advanced mechanical systems, realtime dynamic modelling and integrated testing in a human-in-the-loop virtual environment is the powerful design tool of the future.”
Hoyle agrees that the automotive engineering process will become increasingly virtual, and he reveals that rFpro already has one OEM customer that aims to have in place all the tools required to complete a virtual sign-off of a new car model by the year 2020.
“If you take that statement to its extreme then it means every single piece of real-world test equipment and every single test process applied before sign-off will need a virtual equivalent,” he says.
Unravelling complexity
In the future, Cammaerts from Ansible expects that biotelemetry measurements will play an even larger part of core vehicle development work – and involve new ways of monitoring drivers and vehicles during DIL simulator experiments.
“The data is already there, so it is now a matter of discovering new ways to study it and infuse it into the process,” he says. “This goes beyond human-machine interface studies, and into a deeper understanding of how we actually drive cars – which is an incredibly complex, intelligent activity.”
Another promising future application area is autonomous cars, where Cammaerts says DIL simulators are a “safe and consistent” laboratory tool for putting human drivers into unexpected or emergency situations while interacting with on-board vehicle systems.
“The handover between autonomous cars and human drivers is an excellent example where a DIL simulator is the ideal tool – and perhaps the only tool – to safely and efficiently explore what happens in such situations,” adds Cammaerts.
Did you know?
Driver-in-the-loop simulators are being used to measure the reaction times of drivers to interventions from the latest active safety and driver assistance systems in laboratory environments, reducing the amount of time cars spend on test tracks