Open top menu
18 Jul 2015

A few weeks ago I was fortunate enough to be invited along to Ansible Motion, who've took up residence in the Hethel Technology Centre, around the corner from the original Team Lotus factory in Norfolk

I sat down with their owner and technical director, Kia Cammaerts, for an in-depth discussion about simulator technology in general, the companies direction and of course their product.  However, before we get into that I would just like to point out that Kia's career has spanned a wide range of engineering endevours, honing his skills with the likes of Team Lotus in the early 90's, as an aerodynamacist, before freelancing on a wide variety of F1 based projects.  He founded Ansible Design in 1999 where he focused on the provision of software, for simulation and analysis purposes. Whilst his latest project, Ansible Motion, founded in 2009 is the focus of our attention and is a ground up re-think on the driver simulator.

Kia explained that, as all good engineering ideas do, their simulator journey started with a conversation over a pint, with sketches on the back of bear mats providing a "this is how I'd approach that challenge" debate.

Well that planted the seed, now they needed to water it and watch it flourish, as engineers they approached the challenge with a blank canvas and pragmatic/engineering point of view but also looked at how others were dealing with the same challenges.

Kia went on to explain that the go-to solution for F1 teams and the like is a variant built around a hexapod platform (see below) as used in the aerospace industry before them.
Image Credit Red Bull Content Pool

The problem with the hexapod layout is it has limitations, as the six actuating arms work in pairs to provide x,y,z movements.  In the case of a race car, such as an F1 car, this unravels when you want to perform extreme movements in multiple directions, as the actuating arms have a finite limit on the length of their travel.

Once you have realised this limitation you can either reduce the travel for each movement, leaving a buffer for transitions, which of course then limits the effect you're looking for or you have to think bigger (like Moog did with the Dallara/Ferrari simulator below).  This is why simulators now seem to take up huge expansions of space, as each compete to produce something that closely aligns with the real world experience but inevitably lose latency in doing so.

Moog simulator as used by Dallara and Ferrari who commissioned its design

As you go bigger though you suddenly realise how much more power is needed to generate the forces, which in-turn increases the latency of the movements.  Some of the bigger teams and road car manufacturers will then add an additional layer to reduce the latency, taking away one of the big factors faced by the hexapod design (see below).

The Daimler driver simulator has a slide track, which allows the hexapod to move longitudinally, imitating braking and accelerating forces, but just look at the size of the room it's housed in.

Ansible's Delta series

Ansible Motion's original delta series simulator (S1)

The delta series simulator is, for the purpose of this article, our focus, as it offers dynamic movement like the other simulators we have looked at.  However, Ansible also have their Theta and Sigma series systems too, providing different levels of experience based on the users demands.  When seeing the Delta series sim the first thing you're struck by is its scale, it clearly takes up less floor space than the large hexapod variants, whilst its vertical stance is diminutive too.  This means energy consumption is also reduced, as the dynamic platform needs less room with which to manoeuvre.  Surely that means less definition then? Absolutely not, the patented design is still based on the six degrees of movement principle, it just achieves it in a different way.

The motion platform is a composite design, which features 3 layers, the base platform deals with longitudinal movements, the second platform facilitates rotation, whilst the upper platform provides the vertical movements.  Sounds logical doesn't it? It also gets around that stickiest of problems that the hexapod designs have, movements that require multiple direction changes.

This leads me onto the next phase of simulator engagement, not the physical design but the way in which we perceive the actions of the simulator.  This is an area where Kia and his team continue to work tirelessly in order to achieve a better result.


There are many challenges to creating a virtual world that immerses the user enough to replicate real-world environments, including but not limited to movement, haptics, vision and audible cues.  Ansible concentrated their efforts on movement with the delta series sim, paying particular attention to how the vestibular system reacts to inertia.
For those not versed in it, the vestibular system is a series of semi-circular canals and otoliths in the inner ear which help us to recognise both rotational and linear accelerations (imagine a gyroscope).  Therefore, the vestibular system provides us with spatial awareness and a sense of balance, it's why even in a darkened room when turned left we will know we have been turned in that direction.  However, in isolation our perception can be skewed, for example, if we were asked to walk in a straight line for lets say 20 meters in a darkened room although we could manage the straight line we would invariably come up very short.  The movement we experience in this instance is over heightened as we can't use vision to assist our pre-conceptions of space.

This is because the brain processes the vestibular system and proprioceptive senses* in a timescale of milliseconds but vision and audible cues are processed much slower.

Field Sobriety Test
*Proprioceptive sense is how we interpret the relationship between the parts of our body, think along the lines of muscle memory or to coin a phrase "driving by the seat of your pants" or the test conducted by American police officers when they suspect a drunk driver.  That test requires the driver to walk along a white line, whilst touching their nose with their finger, which establishes that the brain is interpreting the body parts relationships correctly.

Although our brain processes the vestibular information rapidly it reacts to accelerations rather than steady-state motions, meaning movements below a certain threshold do not transmit signals to the brain, with us relying on visual stimulation to bridge the gap.  This is a plus for the hexapod simulator as it gives a timelapse in which the platform can 'washout', returning to neutral ready for the next movement.  The problem however, is the more G you want to simulate the smaller the 'washout' period, which in-turn requires low latency, something which as we've already discussed the larger hexapod designs struggle with.  Ansible's delta series meanwhile can offer a rapid movement that simulates higher G, whilst retaining low latency, improving the operators experience, making it ideal for Motorsport and road car development alike.

Ansible use other cueing techniques in order to fool the mind into believing it is in motion, there is no one size fits all solution, as everyone has differing reactional speeds to the transient conditions we experience.   These can come in the form of haptic controls, such as force feedback delivered through the handwheel to the application of seatbelt tension, whilst low frequency applied to instruments and controls such as shifters and pedals can even enhance the users experience and are currently an area of expansion.

This brings me nicely to the next topic: Environment

I was lucky enough to witness both the S1 and S2 installations as although the S2 is now in full-time use the S1 installation is still tucked away in another part of the technology centre.  The original S1 is encompassed by a smaller screen (5 metres, set back 2 metres from the central focal point using 3 projectors) than its sucessor.  The S2 utilises a 8 metre screen, set back 3 metres from the central focal point, which uses 5 projectors to seemlessly deliver the visual landscape on which you'll drive.  This is an incredibly important factor, with the accuracy of the overlaying images paramount to our perception of the experience.
The S1 has a replica F1 tub mounted on the platform and has been used by drivers from all disciplines, although Kia explained that on occasions this did initially make some drivers retiscent, as they weren't used to sitting with their feet above their bum.  However, most racing drivers can adapt and even those that didn't initially like the driving position soon defaulted to 'racing driver mode', complaining more about setup than their comfort.

This is an important factor for simulators though, as a close representation of the driving environment will only improve feedback from the driver.  Ansible clearly understand this, as the larger S2 platform that I was fortunate enough to try features a lightweight tub, emulating a mid size saloon car (I like to think it was a Touring car or Nascar).

This ability to easily switch out the cockpit makes Ansible Motion's Delta series ideal for roadcar manufacturers, with them able to continually make design changes to the cockpit environment based on driver feedback.  Human drivers, Kia proclaims, do things that they shouldn't be able to, this makes SIL (software in the loop) redundant past a certain point and requires human interaction and feedback to improve the design envelope, until such point we are using autonomous vehicles (perhaps a discussion for another day) we want the drivers reactions to have an impact on how the design is led.

The simulator provides a cost effective solution for both road car and motorsport design with parts and setup changed with a few keystrokes, rather than taking hours of manual labour.  In terms of F1 it forms another layer of correlation as the teams look to validate aerodynamic information from their CFD & Wind Tunnel and before, during and after track testing.  Furthermore, it provides critical information to both the drivers and engineers in this ever more complex world of energy recovery systems.

Being able to fully simulate grand prix distances gives a greater appreciation of the scope of the challenge during the race, be it from the ICE or ERS.  Helping the driver to understand the dynamic changes they'll encounter during a race as the fuel load comes off and BBW (brake-by-wire) settings need to be changed to accommodate this.  Furthermore, fuel and energy management can be explored based on the drivers inputs, as although simulation work can be conducted as we have already discussed drivers often do things that shouldn't be possible and others can't.

Talking about people doing things they shouldn't....

My time at Ansible led to me trying out their Delta S2, which was a little disconcerting when you know these guys have had world class drivers at the helm of their sim.  As I climb aboard and strap myself in, Ian, who'll be my engineer for the session, gives me some pointers so that I acclimatise to the cockpit. 
Everything feels as it should do but of course your mind knows that this isn't a real car and so whenever I drive a sim I always have to remind myself to drive it like I'd expect a real car to drive.  I'm reminded of the training sequence in the first Matrix film as Ian steps back into the control room, then tells me through the headset he's going to load up a couple of test scenarios.  A short pause as the room goes dark and the platform and screen calibration take place, then I'm transported to a proving ground, aboard a generic family sized saloon.  My first task is to drive across the proving ground toward a section of the track, which Ian explains is between two visual markers.  Once there I find a selection of cones laid out in a straight line in front of me, spaced eqadistance apart I must now navigate my way through the cones at a given speed, trying to negotiate them as precisely as possible.
A secondary version of the same test is then undertaken with the cone spacing changed and speed increased in order to ascertain how competant I am and also to check my perception to the cueing methods.  I'm then asked to find my way across the proving ground to the skid pan where I must follow a circular line on the skid pan, increasing my speed until such point I'm on the edge of adhesion and having to correct the slide with some opposite lock.  These processes are important as they not only help me acclamatise to the environment, controls and scenery but also help the engineers correlate the environment to my perception.

After the initial proving ground tests, Ian suggests I try out a single seater, in order that I get an idea of the scale between the different vehicles.  I'm transported to Belgium and the familiar Pirelli logo appears on the sidewalls of the tyres, I'm encouraged to shift up to 1st and bury the throttle as otherwise the car will just stall.  I can immediately feel how much more difficult the car is to turn as I exit out of the garage.  Unfortunately as Ansible don't actually build the car models that are used on the simulator this 'borrowed' single seater model has a tendancy to crash the sim if too much kerb is taken and guess what my first instinct through Eau Rouge is...  Several attempts later, with me seemingly unable to dial myself out of the apex, Ian suggests we head to some old airfield in Northampton instead.

Having been transported the 400 miles almost instantly I find myself sat waiting to pull out of the garage under Silverstones 'Wing'.  Now I must admit I've never really been a fan of the new Silverstone layout (I used to hoon my track car around the old layout, so have a sentimental connection to it) and although I know the corners I'm not fully aware of the apex speeds and braking zones, so I decide to take it a little easier at first, especially as I don't want to find anymore of those kerbs..

The problem you have is that you can't really take it easy, you want to, but you just find yourself with your foot planted and commited.  Now I'm no racing driver, don't get me wrong I can hold my own on a circuit and in a kart but I don't believe I'd ever make a racing driver, although I'd love to spend say 6 months working alongside one of the elite, training and experiencing what they do, if nothing more than to write an in-depth column about it.  Lap by lap I could feel the improvement coming through, but I'm guessing my times were less than as heroic as they felt.  Furthermore, on the odd occasion I'd run out of talent, miss my apex and struggle to re-align the car, just as we see the real drivers do and what's more I could  feel that gut wrenching moment when you know you've stepped over the line, the axienty creep in as I try to correct and the agonising sigh as you realise that's a lap ruined.

A few laps later and Ian suggests we make another trip, this time to Spain where the Circuit de Catalunya awaits.  Like anyone involved with F1 I know the track pretty well, or at least I think I do until I'm driving it in the single seater.  The corners creep up on you so quickly and the lap is over well before I can blink, this time I've kept away from the kerbs but I've already realised which corners are the most tricky and where I believe I'll pick up the most time if I improve there. T4 is particularly tricky as you straighten the car out of T3, short burst of acceleration before unsettling the car for corner entry, this is made worse by the circuit camber, perhaps something we disregard when watching F1 drivers tackle each circuit.  Something I note on Catalunya's straight I hadn't before was the amount of buffering the car is under and the effect this has on you.

All in all I get out of the sim feeling wholly dissatisfied with my own ability but awash with admiration for the experience that Ansible Motion have just bestowed on me.  I could probably go on and on about the whole topic as not only do the F1 applications fascinate me but so do the road car and autonomous vehicle ones, but I'll leave it here for now, with the option of future articles and or edits to this one a distinct possibility.  I'm also hoping that at some point one of the F1 teams will allow me to visit their sim to correlate my experience with Ansible and expand upon the piece I have written here.

See below for a few additional images with comments that I feel may have been missed in the text of the main article.

Powering the experience Ansible have built a custom array of PC's that control everything from the projectors to haptic motors, the team use RFactor Pro to simulate the visual world.
The view from behind the 8 metre screen that helps to emerse the driver in the visual world they're testing in
The 5 projectors that provide the crucial visual overlay
An operations room lies just behind the simulator suite, where a cluster of PC's and engineers study the data being generated and assist the driver.
As we can see in this image the driver is conducting a city scape simulation, which is achieved through the use of RFactor Pro, which enables the end user to switch out the environment and model data for their specific needs.  Formula One teams work tirelessly to improve the experience for the driver, with laser scanned tracks and in-depth tyre modelling providing the desired FEEDBACK.

Special thanks to Elan PR for setting up my meeting with Ansible Motion and making this article possible.

Here are a few video's that help to explain the use of simulators in Formula One too

David Coulthard on Red Bull's simulator

Natalie Pinkham joins Daniel Ricciardo for a look at the Red Bull simulator

Mark Webber explains the Red Bull simulator

Ole big hands Gary Paffett on McLaren's simulator

Charles Pic using the Ansible Motion S1 for Caterham

James Allen gives a tour around the Wirth Research simulator



  1. we found this weblog very great and we wanna thank you for that. We hope you keep up the great work!
    Motion Simulator

  2. Great Post, I liked it and enjoyed reading it, Thanks for sharing this useful and valuable information with us.
    Projectors dealer in chennai


Total Pageviews