In the final installment of Cruden’s Motion Series articles, we explore the role of auxiliary motion systems in driving simulators with product manager, Bastiaan Graafland
This Motion Series of articles has discussed everything from workspace options to advanced motion cueing, chassis mockups and different platform geometries. We’ve seen how different driver-in-the-loop (DIL) simulator solutions work best for different applications – human behaviour research, motorsport, NVH or ADAS development, for example. To complete the picture, now it’s time to look at auxiliary cueing systems – anything from shakers to helmet loaders. How does this less obvious hardware contribute to the success of a driving simulator?
Let’s begin with the most prominent of these unsung heroes, the steering control loading system. Cruden already fits steering position controllers instead of torque controllers for some ADAS simulations, but until the advent of fully autonomous vehicles, every driving simulator will retain a steering wheel as the main input for a driver to a simulator, as it is to a real-world vehicle, so accurate steering torque simulation is a must.
Control loading motors provide force feedback to the driver to replicate the steering feel of a specific car. The wheel is not only used to provide steering input, but also feeds back the reaction forces in the torque of the wheel and other phenomena from the vehicle model like the end stop or understeer in a FWD car on a slippery road.
Control loading motor for steering feedback
All of this makes the steering control loading system not only a prominent haptic feedback system to a driver, but one of the most accurate elements of the entire simulator. Steering feel is crucial to driver confidence and is characteristic to a particular vehicle, so it’s important for the simulated setup to match the virtual car being driven.
As we explored in our earlier article on cascaded systems, the limitations of a motion system mean it cannot always perfectly reproduce the experience of a real car, but a control loading motor can put the exact same torque on the steering wheel that you would otherwise get from the steering rack. So long as there is sufficient fidelity in the motors that create the torque and the vehicle model behind it, it can and should be 100% realistic.
“We always try to keep the connection between the shaft of the motor and the steering wheel as direct as possible, because the control loop that will create the feedback, benefits from a very stiff connection between the two,” explains Cruden product manager, Bastiaan Graafland. “Compliance will result in unrealistic steering feel and controller artifacts. Typical car steering feel should be provided by the simulation model. So instead of mechanically replicating the play in the steering wheel and the joints between the wheel and the pinion on the steering rack, a stiff connection is preferred. The performance of the controller is compromised if part of the compliance is in the hardware and the rest in the model.
“Cruden has extensive understanding on the mechanical side to make it as direct as possible,” he expands, “even when we have to compromise for packaging reasons. We have lots of experience in different linkage systems to solve these issues and make it as stiff as possible. That enables you to accurately model the steering torque characteristics.”
Vibration units or shakers are another common type of auxiliary device. These are added to a motion system to vibrate the mock-up, or certain areas of it. They may be installed on the chassis mockup or upper platform, or in specific areas touched by the driver such as on the seat, steering column, pedal box or shifter.
Shakers are most often implemented in a driving simulator for immersion purposes. Even in a static simulator, the simple addition of background engine vibrations or road noise can bring the structure to life and make a real difference to the driver’s immersion level.
For more targeted experiments, signals from the vehicle model may be used to realistically simulate vibrations that would also be experienced in the real world. NVH studies to compare different sub-system designs each with specific vibration transfer paths between the suspension and the car, for example, requires powerful and precise shaker solutions that are much more complex and expensive than shakers for immersion.
These targeted, high precision shaker applications are also dependent on getting the right signals from the vehicle model. In Cruden’s experience, simulator customers often need to go down a development path with their vehicle model, to make proper use of these kinds of systems.
“As with other simulator features, you should only put these shakers in if the use case demands them,” says Graafland. “If you don’t have the signals or use case for them, then it makes no sense. But you can go quite far in bringing these local vibrations to where you need them for your experiments.
In our earlier motion article on workspace, we discussed how space restrictions mean that driving simulators rarely match the real-world in sustained accelerations and decelerations. One solution to replicating a longer deceleration that cannot be met with motion system movements is to add another auxiliary device: a belt or harness loader. A tug on the belt under braking can help convince the driver that they are decelerating, especially in a motorsport simulator where they are already tightly strapped in.
A belt loader is unrealistic because it’s the opposite of what happens in real life, where the driver moves forward into the belt under braking. Its impact is subjective: some drivers find it makes the simulator more realistic, others dislike it.
Car seat with belt loader
Cruden fits harness loaders to most of its motorsport DIL simulators but in common with other auxiliary devices, it needn’t be specified from day one and can be easily retrofitted. The company has also installed motorsport setups that keep the belt fixed but push the seat forward a few centimeters under heavy braking, thereby pushing the driver into their belts in realistic fashion.
Another solution to the same problem is to use a G-seat, commonly known as a pressure seat in automotive applications. This auxiliary system is a crossover from flight simulation and uses inflatable seat pads to push the driver into the belt, or left or right to suggest cornering forces.
“Some users are very enthusiastic about this solution but in my opinion, a pressure seat’s contribution does not often justify the additional simulator complexity,” says Graafland. “One of its drawbacks is latency, which is bad for immersion. Another is that you need individually tailored adjustments to make it work properly. This might not be a problem for a one-driver, one-setup motorsport situation, but in automotive, where you want people of all different sizes in your simulator seats, it is much more difficult to calibrate it precisely.”
In Graafland’s opinion, adding a simple 3DOF motion system is a better way to indicate lateral acceleration than adding a pressure seat to a static simulator.
Braking feedback is increasingly communicated by another auxiliary system, the brake pedal loader. These actuators behind the pedal have a role in ADAS testing, where systems may push back on the throttle pedal to warn the driver of a hazard ahead, or in motorsport, where each driver might have his or her specific pedal feel. Real-life hardware differences in the pedal can be modeled in the actuator controller such that driver changes in the simulator needn’t result in a corresponding swap of the pedal system, just changes in a model.
“As with the steering system, Cruden has good knowledge of designing these linkages between the motor and the pedal to make the solution as light and stiff as possible,” says Graafland. “That overcomes any negative effects for the control loop from introducing compliances and inertia.”
Whether it’s a pedal loader, steering control loader, shaker or one of the more niche systems like a wind blower or helmet loader, auxiliary systems can enhance the driver’s immersion and make a driving simulator more realistic – but only if the resulting stimulus is properly synchronized with the wider motion and visual systems.
“Latency is a factor in all the systems we’ve discussed,” he cautions. “If we’re going to install it then we must make sure that the timing is exactly right, that the device is not doing something that is disconnected from what you would expect. Timing is more important than scale or an exact representation of what you would feel in the real world. The feedback has to come at the moment you expect it so if you can’t do it right, don’t do it.
“Before you add more systems, make sure you understand their impact on the human brain that will sense them. As we’ve said before, the main stimulus is the visual system, so that is the one to get right first, then the motion system and finally these auxiliary systems. But none of them make sense if you don’t get the latencies low and the synchronization right. Adding more systems will not automatically make your simulator better, but adding systems properly, will.”
For more information, please contact Dennis Marcus via firstname.lastname@example.org 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.
View all articles in our Motion Series of articles: here.
Article 1: Driving Simulator Motion Systems 101