In article 2 of 8 in our Motion Series of articles, Cruden’s Dennis Marcus and Martijn de Mooij discuss the role of workspace in simulator performance and the true benefit (or not) of adding more
In our first, introductory article to this series we noted that many people assume that the biggest possible workspace for a driving simulator is always the right way to go. Workspace is the most visible specification of a motion system so it’s understandable that it tends to attract a lot of attention – perhaps more than it should.
In theory at least, motion cues can last longer with more workspace, which leads to the argument that bigger is better. The more workspace you have, the longer you can accelerate at a certain level, but only up to the simulator’s limited maximum velocity.
What is “workspace” anyway? It is the commodity that a motion simulator has available to generate movement. Part of that movement is the actual motion cue (acceleration) and another part deals with the consequences of that initial cue (mitigating built-up velocity and position offset). Basically, the workspace determines how much movement can be generated by the platform. There is a clearly visible limitation in position, but just as important are physical limits in velocity and acceleration. After all, it is the acceleration that is felt by the human driver.
Simulator workspace is always characterised by a combination of position, velocity and acceleration. Consider that F = m × a. The acceleration a is what the driver feels, so that’s what the motion system must provide, with velocity and a position offset as by-products. This isn’t as easily represented in a picture. We can all visualise something measured in metres, but it’s much harder to imagine the relation between workspace, velocity and acceleration.
The problem is that workspace on a driving simulator is always limited, especially in comparison to a real vehicle. Consequently, motion cues on the simulator tend to be smaller and do not last as long as those in the real vehicle for most maneuvers.
In a real car, if you hit the brakes hard at 100 km/h, you will decelerate for several seconds until the car comes to a full stop. An unscaled (imaginary) sustained brake cue in a simulator would result a travelled distance of at least 40 m and, even worse, with a residual velocity delta of 100 km/h. In a practical simulator you will however reach the end of the workspace or hit the velocity limit very quickly. Doubling the longitudinal stroke from, say, 1m to 2m, will make the simulator’s footprint significantly larger, but the extension of the experienced acceleration cue is only marginal.
If you want to add a significant amount of duration to the cue, then many metres of stroke must be added. Yet if you don’t adjust the maximum velocity as well, then the simulator could reach its maximum velocity without even using the extended workspace. It’s acceleration that the simulator driver experiences and without increasing the maximum velocity in conjunction with increasing the stroke, there will be only a limited increase in cueing potential. Hence a simulator’s workspace should feature a balanced combination of position, velocity and acceleration limits
When discussing workspace, it’s easy to think purely in terms of longitudinal or lateral movement, but in fact it covers each of the simulator’s six degrees of freedom. Some are easier to replicate than others. For example, it’s relatively straightforward for the motion platform to handle the roll that you experience in a car. It’s the same for pitch and, to a certain extent, heave.
It is mainly the longitudinal, lateral and yaw directions that are challenging, independent of the type of motion system. A yaw table on a simulator might provide 360° of continuous, infinite yaw, in conjunction with a 360° visual system. Generous lateral workspace can be harder to accommodate, but if the simulator is focused on highway driving experiments, then it makes sense to have a lateral workspace that is as wide as a highway, such as at Daimler’s simulator in Sindelfingen. On these large simulators, a driver can experience unscaled lateral accelerations without the risk of false cues, caused by washouts.
The explanation for this is quite simple; when driving on the far left lane, the simulator is positioned at the far left side of the lateral workspace, and when driving to the centre lane, the simulator moves to the middle of the lateral workspace.
Whatever the application, longitudinal workspace can be a thorny issue. Under braking, one would ideally require several hundreds of metres of workspace in the longitudinal direction, but no viable simulator configuration can provide this amount of stroke.
The Daimler simulator is an example of how the workspace should always be matched to the application. By contrast, a motorsport simulator typically has much smaller lateral workspace, because it is not about lane changes. In a motorsport simulator the driver needs to feel when and on what wheel the car is losing grip in a high-speed corner where the sustained lateral g-forces are well beyond the capabilities of any motion platform.
While stroke is easy to visualise, it doesn’t provide the full picture on the usability of the workspace.
We know that if we increase the workspace, we can accelerate for a slightly longer period of time. But since the workspace is limited, we must still manage the acceleration properly to avoid hitting the workspace’s end-stop. A ‘washout’ is therefore required to reduce the velocity to zero before running out of workspace. In fact, the majority of the workspace is often required not for the initial motion cue, but for the washout.
In an ideal world the drivers will feel the cue that they require and expect – timely initial acceleration – and not feel the washout of the motion cue, for example, the platform being brought to an abrupt standstill. Every human being has a perception threshold, below which soft touches or small movements might go unnoticed. Motion cues obviously have to be above the perception threshold, whereas washouts should be executed below the perception threshold.
It’s hard to keep the washout below this level because human perception thresholds are quite low – and often lower still for racing drivers. To stay below it, the ratio between an acceleration cue and washout can easily be 1:10. Accelerating for a tenth of a second might then demand a full second to return to a standstill. As a result, if you add 1 m of longitudinal workspace, most of that will be used up for washout, not for increasing the duration of the actual acceleration cue.
A driving simulator is much more than a motion platform. Depending on your use cases, investments in other areas than the motion system might have a larger impact on the immersion of your simulator. We generally recommend that if you do add workspace for certain use cases, make sure you add enough to allow for position cueing rather than acceleration cueing. Cruden advises customers to always understand the cost of adding workspace versus the potential benefit. We discuss at length what the customer is trying to achieve with the simulator, establish the use cases and map them to the simulator requirements and specifications.
BMW’s new simulator park at its research and innovation centre (FIZ) in Munich is a great case study for matching the simulator to the application. On the ground floor are very large simulators with a lot of longitudinal, lateral and yaw workspace to simulate the vehicle dynamics of autonomous driving in urban situations. On the third floor are multiple, smaller simulators that will focus on driver immersion for HMI research. There are no ‘halfway house’ simulators. BMW did its own research and came to the same conclusion as Cruden; that workspace should match with the use case for the simulator and you either do workspace really well, or put the primary focus on other aspects of the simulator.
Cruden customers derive huge benefits from simulators that have even a limited workspace. No matter how large or small the workspace is, we always aim to get the maximum out of a simulator for what it is you are trying to achieve. We’ll return to this philosophy again when we discuss motion cueing in a later article. Until next time!
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.
Links to subsequent articles will be added below as they are published.
Article 1: Driving Simulator Motion Systems 101