Dr. Adrian Simms

Dr. Adrian Simms

Modern driving simulators are delivering some unprecedented levels of realism and can extend the role of the simulator in areas of vehicle dynamics and ADAS development but, to ensure that a simulator will deliver on its promises, OEMs must take a rigorous approach to evaluation. We spoke to Dr Adrian Simms, Business Director (Laboratory Test Systems) at AB Dynamics about how to assess a simulator effectively.

What makes it so challenging to assess the latest generation of driving simulators?

Until now, the range of simulator applications has been constrained by the physics of the motion platform, in particular, its inertia. Simulators have been most successfully utilised for less performance demanding applications such as human-machine interfaces and ergonomic studies. The high performance required for vehicle dynamics applications was not previously available.

The latest generation of simulators, such as AB Dynamics aVDS (advanced driving simulator) offer more precise on-time cueing (audio, visual, haptic, vestibular), greater travel for effective motion cueing, and low latency ensuring the driver's senses receive this input at precisely the right time to increase realism and eliminate motion sickness. Different suppliers use different technologies and it can be difficult for the customer to establish which will be the most satisfactory, and to specify the best system for their application.

Comparing inappropriate metrics can lead to costly and incorrect choices being made, as several vehicle manufacturers have discovered. For example, where a hexapod architecture is used, impressive travel in one axis often hides limited excursion in combined directions. Systems with additional mechanisms to overcome this inherent limitation will have higher mass and inertia, which is detrimental to the simulator's performance.

What kind of personnel need to be involved in the assessment?

Once the decision has been made to invest in a driving simulator, it is essential to commit a highly skilled test driver for sufficient time to perform a rigorous assessment. This may seem self-evident but it is surprising how many companies, for whatever reasons, fall short in this area.

The driver should be capable of evaluating changes to a car's parameters in the order of 1-2 percent on the track, while also being familiar with tuning a vehicle to this level. If the company already uses a simulator, the new simulator should be evaluated using an existing vehicle model with which the driver already has extensive experience.

How should an evaluation be structured?

Does the simulator feel like driving a real car?

The allocation of sufficient time for the driver to spend in the simulator is key to a successful selection process. Initial acclimatisation will probably require at least an hour just to establish a good 'feel' for the vehicle. First impressions would include an assessment of realism – does the simulator feel like driving a real car? Using a proven and familiar model of one of the company's vehicles, the driver should evaluate whether the simulator experience matches the car that has been modelled in all respects, or whether there are shortcomings.

The next step should be to introduce small changes (1-2 percent) to the vehicle model that are 'blind' to the driver, who should then perform a variety of manoeuvres on different road types and surfaces. For an adequate assessment of simulator performance, the customer should include changes to as large a range of parameters in the vehicle model as possible, as part of their testing. These might include spring rates, damping, anti-roll bar stiffness, tyre parameters and others.

What differentiates the better performing simulators from others?

In the best simulators, the driver should be immediately able to detect the changes made to the model and identify which parameters have been changed. Moreover, he or she should be able to correctly identify the approximate magnitude of the change. Such is the level of fidelity now available, the effect of the change should feel identical to making a similar change on the real car.

Poor simulators often give a 'floating' or disconnected feel, similar to that experienced when driving on ice.

Poor simulators often give a 'floating' or disconnected feel, similar to that experienced when driving on ice. In low-quality simulators, large changes to some parameters – even 100 percent – are not always detectable. In some cases, the driver detects the change, but incorrectly perceives its magnitude, thinking it is either much smaller or larger than the actual alteration. Test drivers have also reported that, on some simulators, small changes can be felt for some parameters but not others, which is why it is important to evaluate such a wide range of parameters.

These deficiencies in the poorer simulators clearly place severe limitations on the potential for the simulator to contribute to vehicle development and underline the importance of taking sufficient care over the evaluation process.