Generating the most natural human behaviour is always the goal when specifying a driving simulator. In article 4 of 8 in our Motion Series, Cruden’s Dennis Marcus and Martijn de Mooij discuss how that target can be achieved with very different motion systems, depending on the application for the simulator.
A human in a driving simulator should always drive the virtual car as closely as possible to how they would on a real road or track. But different experiments require the simulated car to be driven in different ways and by different types of driver.
If the human is a professional driver, and especially if the driver is going to be exploring the performance limits of a new vehicle design – as in motorsport simulation or vehicle dynamics work – then the motion system should provide the driver with the right kind of feedback from the motion system. These educated drivers know what to look for and know how it feels to drive on the limit.
However, if you are seeking to immerse ordinary human drivers in a research study or see how they will handle a new HMI or ADAS function, they will require a different type of motion. We mere mortals are not as sensitive as the professionals – not as good at finding the limit of the vehicle, and nor are we looking for that in the simulator when testing ADAS. We just want to feel like we’re in a car, in order to experience the ADAS system as we would in real life.
This HMI and ADAS work is becoming the primary use for driving simulators at automotive OEMs. Even simulators originally implemented in vehicle dynamics departments are increasingly being used for HMI and ADAS experiments, especially now that electronic stability control technology is mature.
Motion for vehicle dynamics and motorsport
Humans have four primary feedback channels when driving a car – visual, vestibular, haptic and audio. Racing drivers vary in their reliance on some feedback loops over others. For some drivers, steering feel – a haptic channel – is more important than motion feedback, for example.
We’ve stressed before in our Content series of articles that the visual system is the most important feature of a driving simulator. You can drive a simulator without a motion system, but not one without a visual system. But from a physiological point of view, in human beings the vestibular system provides faster feedback than the visual system. Even if a driver – and especially a trained driver – can see that their car is sliding, they probably feel it first, either on the steering wheel or through their body (butt).
In vehicle dynamics or motorsport simulation, motion provides a very important cue. The moment the driver feels that the rear of the car is losing grip, for example, is a cue for them to react. Providing accurate vestibular cues makes the driver more likely to respond faster – and therefore more closely to how they would in the real world – than if they could only see the car’s movements.
This form of motion requires a very precise 3D model of the track built from laser point-cloud datasets. The driver needs to precisely feel the profile of the kerbs or the vibrations from a tram rail on a Formula E street circuit. It also requires a high-fidelity motion system to convey to the driver’s seat the details of that surface and the accelerations of the vehicle across it. Even with these elements in place, this type of work is still very difficult to do. A lot of effort must go into correlating the physics-based vehicle dynamics model, for example in Adams, to really make use of a DIL simulation for vehicle dynamics development.
Motion for immersion
If you’re doing vehicle dynamics research, and especially motorsport simulation, the way the driver feels the car behave is an essential element of the research you’re undertaking. That’s not the case for evaluating how an automotive ADAS system takes care of a lane-change, however. For this type of research with non-expert drivers, the car is driven in a normal way. The tires are never even close to saturation, so motion cues are less important than the visual and haptic ones, and not used by the drivers in these applications.
Increasingly, the development focus at OEMs is on HMI and how people experience driving a new car. What does the interior feel like? How do the infotainment systems work? Are they easy to operate while driving through traffic? Do ADAS systems work intuitively? Is the way in which they correct and assist the customer understood and appreciated?
All of these aspects are tested by large groups of regular drivers who drive in a normal way, not like they are trying to set a new lap record at the Nürburgring. Nevertheless, a degree of motion is always important for true immersion, even if it’s from a low-fidelity motion system. The key is for it to be believable motion.
It’s the same for yaw motion: even applying a small amount when taking a corner makes it feel more realistic, which in turn leads to more natural driver behavior.
If your simulator will only ever need to provide motion for immersion, a much less complicated motion system will often be sufficient – perhaps only three degrees of freedom, with heave, roll and pitch, rather than all six and a large workspace.
Fit for purpose
As always, a driving simulator should be specified according to the application. Customers should build something that’s fit for purpose in order to make the most of their investment.
A vehicle dynamics simulator of the type that would be used in motorsport might consist only of the part of the race car that the driver sits in – which will typically be small and light – plus a motion platform, and nothing else. The focus is on the motion cues that the driver needs to feel.
But if the emphasis is on human factors – researching how people interact with vehicle systems – then creating a natural environment with the highest possible level of immersion is more important. This might mean building a simulator with a full chassis mockup so that the driver sits inside an actual car. In such an application, a full chassis mockup with only three degrees of freedom in the motion system, combined with a high-end visual system, is a better way to spend the simulator budget than on a full motion platform that delivers accelerations that are not usually present in highway driving.
Cruden’s approach always begins by talking to the customer to figure out what the simulator will be used for. Those discussions shape the direction to go with that particular simulator, which means we can create a much better solution for the available budget. In particular, when it comes to motion for immersion, we’re convinced that simpler is often better – that customers can achieve what they want, with a lot less.
Irrespective of motion system layout, we always try to provide the information that the driver is looking for – only those aspects of the car’s movement that matter to the driver. We’ll return to this theme again in a future article on motion cueing.
For more information, please contact Dennis Marcus via email@example.com or on +31 20 707 4646.
If you think a colleague would be interested in receiving the articles in this series, you can sign them up to our newsletters here.
Links to subsequent articles will be added below as they are published.
View all articles in our Motion Series of articles: here.
Article 1: Driving Simulator Motion Systems 101