- Home
- Applications
- Driving Simulator Motion Platform
Application · DIL rigs, training simulators and vehicle development
Driving simulator motion platform, bought for the first half second of every event.
A road vehicle produces accelerations that last for seconds. No platform reproduces those, and the arithmetic that says so is on this page. What a platform delivers is the onset — the first few hundred milliseconds, when the driver decides whether the car is real — and knowing which half of each event you are buying is most of what a specification should say.
Looking for a complete driving simulator? We supply the motion subsystem underneath one. The cab, the steering system, the seat shaker and the vehicle model are three or four other suppliers — the road cue crosswalk shows which cue comes from where.
White racing simulator on a motion platform in a clean industrial workshop
What the rig is actually for
Eight driving rigs, and the one cue each of them is judged on
These eight buy very different machines, and the separating question is rarely fidelity. It is which cue decides whether the rig works: the onset of a manoeuvre, the texture of a road surface, the moment a driver takes back control, or the fact that fifty people a day get in and out of it.
Training
Driver training and licensing
Rigs that teach hazard perception, emergency handling and conditions a learner should not meet on a public road for the first time.
ADAS
ADAS and automated driving human factors
Takeover studies, trust and mode-awareness work, where the interesting data is what a driver does in the two seconds after the car asks for help.
Chassis development
Ride and handling development
Subjective evaluation of chassis, suspension and tyre changes before a prototype exists, driven by the customer’s own vehicle model.
Commercial vehicles
Truck, bus and commercial vehicle training
Heavy vehicle cabs where the cue set is slower, larger and dominated by body roll, load shift and a much longer response to steering.
Two-wheelers
Motorcycle and powered two-wheeler simulation
A different problem wearing the same clothes: the rider leans, the machine rolls, and roll angle is the primary cue rather than a secondary one.
Building a rig these eight do not describe? Send the one manoeuvre that decides it and we will work the machine out of that →
Four cues, four suppliers
What a driver feels in a car arrives from four places, and only one of them is a platform
Ask a driver what a car feels like and they describe one sensation. Build a rig and it turns out to be four separate engineering problems, each with a different frequency range, a different device and a different company on the invoice.
The platform owns chassis body motion — the roll, pitch, heave, surge and sway a vehicle model produces, up to the point where a person stops perceiving motion as motion. Road texture lives an octave or more above that and belongs to a small vibration device, not to a machine carrying half a tonne of cab. Steering torque is a loop of its own. Seat pressure and belt tension are hardware in the seat.
Getting this split wrong is the most expensive mistake in a driving simulator budget, and it usually runs one way: a larger platform is bought in the hope that it will produce the road surface, which it cannot, because moving mass and high frequency are opposites.
Nobody buys a bigger platform to get a better road. They buy a shaker, and a smaller platform than they were about to order.
Not sure which of the four your missing cue belongs to? Describe what feels wrong and we will name the device →
The 4-Way Road Cue Crosswalk Which device produces which sensation, and what happens when the wrong one is asked
| Cue | What produces it | Roughly where it lives | If you ask the platform to do it |
|---|---|---|---|
| Chassis body motion — the car leaning, squatting, pitching, sliding | The motion platform | The band where a person still reads motion as motion | Nothing goes wrong. This is the job |
| Road texture — surface coarseness, expansion joints, rumble strips | A vibration device under the seat or the floor | Well above the band a loaded platform can reach | You buy a larger platform, get a duller road, and pay twice |
| Steering feel — self-centring, kickback, the moment grip goes | The steering force-feedback system | A loop an order of magnitude faster than the motion loop | It cannot. This cue never passes through the platform at all |
| Seat and belt — pressure, bolster, belt tension under braking | Seat hardware and a tensioner | Static and quasi-static, felt as force rather than movement | The platform tilts to fake it, and the driver sees the world tilt |
Row two is where budgets go missing. A vibration device costs a fraction of the platform upgrade it replaces, and the two are not substitutes in either direction — a shaker cannot lean the car and a platform cannot resolve a road surface. Where a rig has to have both, the split is agreed before either is sized, because it changes the mass the platform is carrying.
The number that settles most driving simulator specifications
No platform reproduces braking, and the arithmetic takes about a minute to check
Braking is the manoeuvre every driving rig is judged on and the one no motion system delivers. It is worth working through in public, because a supplier who lets a customer believe otherwise has sold a machine that will disappoint on the first day it is used.
Take a firm stop: 0.8 g held for three and a half seconds, which is roughly 100 km/h brought to rest. To produce that by moving the cab backwards, a platform would need about 15.7 m of surge travel. Nothing built for a room has that, and nothing ever will.
So the cue is faked with tilt. Pitch the cab nose-down and a component of gravity presses the driver into the belts the way deceleration would. It works, and it is what every simulator on earth does. The catch is the rate: washout filters hold tilt to somewhere around 3 °/s so the rotation is not felt as rotation, and the angle that fakes 0.8 g is 38.7°. At 3 °/s that takes 12.9 seconds to reach.
The stop lasted 3.5 seconds. The tilt is still on its way when the car has been stationary for nine.
What the platform actually delivers is the first part — a fast surge that produces the real specific force for a few hundred milliseconds while the driver decides whether the car is behaving. That is not a compromise anybody chose. It is the physics of a room, and it applies to every supplier equally.
The onset is the product. Everything after it is a washout filter being tactful.
what the car doeswhat surge deliverswhat tilt has added by 4 sseconds from the start of brakingspecific force at the driver (g)
The Braking Cue Ladder What each level of braking would cost, in travel and in time
| Braking level | Specific force | Surge to hold it for 2 s | Tilt that fakes it | Time to reach that tilt at 3 °/s | What you actually get |
|---|---|---|---|---|---|
| 0.2 g — easing off for a junction | 1.96 m/s² | 3.9 m | 11.3° | 3.8 s | Onset, then a tilt that mostly arrives in time |
| 0.3 g — firm, everyday braking | 2.94 m/s² | 5.9 m | 16.7° | 5.6 s | Onset, and a tilt still arriving after the event |
| 0.5 g — hard braking | 4.91 m/s² | 9.8 m | 26.6° | 8.9 s | Onset only. The tilt is now visible in the scene |
| 0.8 g — emergency stop | 7.85 m/s² | 15.7 m | 38.7° | 12.9 s | Onset only, and nothing else is on offer |
| 1.0 g — a stop on a good surface with ABS | 9.81 m/s² | 19.6 m | 45.0° | 15.0 s | Onset only. A cab tilted 45° is a fairground ride |
This is arithmetic rather than a property of our machines — surge from ½at², tilt from tan θ = a/g, arrival time from the rate limit — and every supplier’s platform is subject to the same five rows. What genuinely differs between suppliers is how fast the first two hundred milliseconds arrive, how smoothly the surge washes out afterwards, and how honestly the rest is described before an order is placed.
The reason driving is harder than flight
Tilting the cab tilts the world, and in a car the world is right there
In an aircraft the horizon is far away, pitch attitude is part of the manoeuvre anyway, and a few degrees of cab tilt disappear into a scene that was going to move regardless. In a car the road surface fills the lower half of the field of view and the horizon sits at a height the driver has known since they learned to drive.
So the cue that solves braking creates a second problem: pitch the cab 10° nose-down and the driver sees the road arrive where it should not be. The fix is for the image generator to rotate the scene by the platform’s actual attitude, which sounds trivial and is not — it means the visual system has to be told the true pose, continuously, and that becomes another channel in the interface document and another entry in the delay budget.
A rig where the compensation is missing feels wrong in a way drivers describe as “floaty” without being able to say why. A rig where the compensation is late feels worse than one with no tilt at all.
In a driving simulator the motion cue and the visual cue are not two subsystems. They are one cue delivered by two suppliers.
What this means for a specification
Specify the onset, then decide what happens to the rest
Three decisions follow from the table above, and they are worth making deliberately rather than inheriting from whichever washout filter arrives with the software.
How fast the onset arrives. This is the part you are buying, so it is the part to test. It is set by moving mass, actuator bandwidth and the total loop delay, and it is measurable in a way that “realism” is not.
How much tilt you allow. More tilt gives a stronger sustained cue and a more visible scene rotation. Where the number sits is a judgement about your drivers and your content, not a property of the platform, and it should be a tunable parameter rather than a constant somebody buried.
What the washout does on the way back. The return to centre is where badly configured rigs give themselves away, because the same sensitivity that makes a quick onset convincing makes a quick return obvious.
Have a manoeuvre list but no view on how much of it is reachable? Send the list and we will mark each one onset-only, partial or full →
How driving rigs actually divide
The 4-Configuration Driving Platform Stack
Four configurations. What separates them is not size — it is which requirement ends up overruling the others once drawings start, and that is usually decided by who is sitting in the rig and what they are being asked to judge.
D1
Compact driver-in-the-loop rig
Onset governs · seat, wheel, pedals and screens, commonly under 300 kg
Small, stiff and quick. Bought where the value is in how fast the first movement arrives — motorsport, limit handling, tyre and grip work — and where travel matters far less than crispness.
Inside the supply
- The moving assembly, its drives, its controller and the cabinet as one delivered item
- Onset measured at the driver’s head rather than at the frame
- Command interface to your model, with the axis convention fixed in writing
- Loaded acceptance driven from your model rather than a demonstration file
Not included: the seat, wheel and pedal hardware, the force-feedback system, the screens and the vehicle model.
D2
Full-cab driving simulator
Load and eye point govern · cab section, occupants and screens as one load
A real cab, or a body section with real trim, carrying a driver and often an instructor. The moving mass grows, the head sits high above the interface, and both facts eat travel before any cue is produced.
Inside the supply
- Envelope stated at the driver’s head, with the manoeuvre list checked against it
- Structural interface designed for the cab you are actually mounting
- Access and boarding arrangement designed with the rig, not added afterwards
- Duty sizing against sessions per day rather than a peak demand
Not included: the cab or body section, the visual system, the instructor station, and the seat and restraint hardware.
D3
Vehicle dynamics development platform
Interface honesty governs · usually a buck with real controls
Here an engineer’s subjective verdict has to describe a model change rather than the rig producing it. So the rig’s own behaviour has to be measured and written down rather than simply signed for.
Inside the supply
- The platform’s measured response, issued so your model can be corrected for the rig
- Repeatability given as the spread across repeated runs, never as a best figure
- Axis convention, units and channel signs frozen and verified channel by channel
- State returned to your host at the rate agreed at the interface review
Not included: the vehicle model and its tyre data, your cueing filter where it is yours, and every conclusion drawn from what the rig produces.
D4
Platform on a long linear axis
One direction needs travel a hexapod cannot give
Where the programme genuinely needs metres rather than hundreds of millimetres in surge or sway, the answer is a platform carried on a linear axis rather than a bigger hexagon. It is a different machine with different civil requirements.
Inside the supply
- Platform, carriage and drive as one controlled system with a single command interface
- Combined envelope stated for the pair, not for each half separately
- Interface loads issued for the rail foundation designer to work against
- Travel and payload confirmed per project before anything is quoted
Not included: the rail foundation and civil works, the building, the safety fencing and the site approvals.
D3 is the configuration most often bought as D2, and it costs a programme a year. A development rig that has not had its own response characterised cannot separate a model change from a machine artefact, so every subjective verdict carries an argument about whether the rig did it. That characterisation is cheap at build time and impossible to add retrospectively with any authority.
Torn between a development rig and a training rig because the programme serves both? Tell us who will be judging results and we will name the configuration →
Which mechanism sits under the buck
The platforms these driving rigs are built on
Road vehicle work has a particular demand that flight work does not: braking and cornering arrive in the same second, and both are transient. That pushes almost every serious driving rig towards six axes, and it pushes the specification towards bandwidth rather than travel.
Three axes
3DOF motion platform
A serial mechanism giving heave, roll and pitch. Suits a rig whose task lives in one plane — a heavy vehicle trainer dominated by body roll, or a study where the motion only has to load the driving task rather than reproduce it.
Six axes
6DOF motion platform
The normal answer for road vehicle work. A car that leans correctly through a corner while failing to squat under the brakes reads as wrong long before a driver can explain why, and the two cues arrive together.
Hexapod geometry
Stewart platform
Six legs laid out for stiffness rather than reach, with the pivot placed by geometry instead of software. Chosen for limit-handling and motorsport rigs, where the entire value sits in how quickly the first movement arrives.
The Axis Read from the Driving Task Let the driving task pick the mechanism, before any brochure does
| What the driving task demands | What to specify | Why |
|---|---|---|
| Braking and cornering arriving in the same second, which is most road driving | 6DOF | Two transients in different directions cannot be produced by three axes without one of them lying |
| One plane dominating — heavy vehicle body roll, or motion used only to load a task | 3DOF | The three axes you do not command are three you do not pay to move |
| Onset quality above everything, with travel a secondary concern | Stewart | Stiffness and low moving mass decide how quickly the first movement reaches the driver |
| Metres of surge or sway rather than hundreds of millimetres | Platform on a linear axis | Stretching a hexagon for one long direction spends the other five to buy it |
The axis argument in full is on 3DOF vs 6DOF, and payload, stroke and acceleration explains why every kilogram added to a cab is subtracted from the onset. For driving rigs specifically, electric vs hydraulic matters because a rig usually lives in an office building with engineers working next to it.
Know the manoeuvres but not the mechanism? List three manoeuvres that decide the rig and we will return the axis count →
Recorded with a driver in the seat and a model in the loop
Three recordings that show what a datasheet has no way of saying
All three of these are things a rig either does or does not do, and none of them can be argued about afterwards once they are on film.
The honest version of a stop. The model’s deceleration and the specific force actually measured at the driver’s head, on one time axis, so the onset and the washout are both visible.
What the platform is not doing. One lap run twice, with the seat vibration device disabled and enabled, so the road cue can be heard and felt as a separate thing.
The afternoon nobody films. Each channel driven in isolation against the signed convention sheet, so a left-hand bend rolls the car the way a left-hand bend should.
Want the braking recording before you look at any specification? Ask for the recorded stop, and the method that produced the trace →
The input that eats commissioning weeks
Which way is positive? Two standards disagree about two of the three axes
A vehicle model outputs numbers. A platform consumes numbers. Whether those are the same numbers depends on a convention that neither side usually states, because inside each company it has been obvious for years.
The classic SAE J670 vehicle axis system puts x forward, y to the right and z downward. ISO 8855 puts z upward, and to stay right-handed, y to the left. One of the three axes agrees. Two do not, and the rotations that follow from them do not either.
A rig with a sign error in the lateral channel rolls the car the wrong way in bends. Drivers notice within a lap, but not always consciously — several report that the car “feels like a boat” before anybody thinks to check a sign. A sign error in the vertical channel turns every bump into a dip. Neither is a hard fault, nothing alarms, and both can survive months in a rig that is otherwise working.
So it goes on one page, signed, before manufacture, and it is verified channel by channel on the day the model is first connected. It takes an afternoon.
The Axis Convention Crosswalk Where the two common standards disagree, and what the disagreement feels like
| Quantity | Classic SAE J670 | ISO 8855 | What a mismatch feels like in the rig |
|---|---|---|---|
| Longitudinal x | Forward positive | Forward positive | Nothing — this is the axis that agrees |
| Lateral y | Positive to the right | Positive to the left | The car leans into the corner instead of out of it |
| Vertical z | Positive downward | Positive upward | Every bump becomes a dip, and the rig feels hollow |
| Rotations | Right-handed about the axes above | Right-handed about the axes above | Braking raises the nose; steering yaws away from the turn |
The 7-Line Interface Convention Check Seven questions, asked once, before anything is manufactured
| Line | The two answers we see | What a wrong assumption looks like |
|---|---|---|
| Axis directions | Classic SAE J670, or ISO 8855 | The car rolls the wrong way in bends, and bumps read as dips |
| Rotation signs | Right-handed about the chosen axes, or a house convention | Braking raises the nose, which nobody believes and everybody argues about |
| Angular units | Radians, or degrees | A factor of 57. Obvious in the first minute, which makes it the good case |
| Linear units | m/s², or g | A factor of 9.81. Also obvious immediately, and also harmless |
| Reference point | Vehicle centre of gravity, or the driver’s head | Cues the right size and the wrong character. Nobody spots this one, which makes it the bad case |
| Gravity in the vertical channel | Included in the specific force, or removed | The platform sits an entire g low, or spends its life holding an offset it should not have |
| Update rate and late-packet behaviour | Hold the last value, extrapolate, or fault | A rig that stutters under load while everyone blames the network |
Lines five and six are the expensive ones. A model that reports accelerations at the vehicle centre of gravity is describing a point roughly half a metre below and behind the driver’s head, and the difference between those two points is exactly the roll and pitch content a driver reads as character. Gravity is worse: whether the vertical channel is a specific force or a pure acceleration changes it by 9.81 m/s², and a rig can be commissioned around the mistake by trimming it out.
Not certain which convention your model was written in? Send one second of output for a known manoeuvre and we will tell you →
Where the platform sits in a driver-in-the-loop rig
Three loops close through the driver, and the tightest one is not ours
A DIL rig is usually drawn as one loop: model to platform to driver to controls to model. In practice there are three, they run at different speeds, and the platform is the slowest of them. That is not a weakness — it is what the human body allows.
Motion is bounded by perception. Past roughly the point where a person stops reading movement as movement, extra bandwidth in a machine carrying a cab buys nothing a driver can use. The visual loop is bounded by frame rate. The steering torque loop is bounded by the hand, which resolves vibration far higher in frequency than the inner ear resolves motion — which is why steering force feedback runs an order of magnitude faster than anything we build.
The consequence for a specification is narrow and useful: the platform does not have to be the fastest thing in the rig, but it must not be conspicuously later than the other two. A car whose wheel loads up half a beat before the body leans feels disconnected in a way drivers dislike immediately and describe badly.
Three suppliers means three delay budgets and one driver who experiences the sum of them. Asking each supplier for a figure in the same units, early, is the cheapest thing a rig integrator can do — and the answers are rarely comparable unless somebody insists on the format.
A driving rig is not one loop with three suppliers. It is three loops with one driver closing all of them at once.
The 3-Loop DIL Map What closes each loop, roughly how quick it has to be, and whose it is
| Loop | Closed through | Bounded by | Owner | What goes wrong if it is the late one |
|---|---|---|---|---|
| Motion | The platform, then the driver’s vestibular system | What a person still reads as motion | Us, with whoever runs the cueing filter | The car feels detached from the wheel and the picture |
| Visual | The image generator, then the eyes | Frame rate and the chain’s own latency | The visual system supplier | The scene lags the seat, which is the sickening order |
| Steering | The force-feedback wheel, then the hands | What a hand can resolve, which is far higher | The steering system supplier | Grip loss arrives after the driver has already reacted to it |
The third row is the one integrators underestimate. Steering is the fastest channel a driver has and the first place a rig gives itself away, which is also why it is worth asking the wheel supplier for their latency figure before anybody argues about the platform’s.
Assembling delay figures from three suppliers? Ask for ours in the format the other two can be written into →
Why the figures on a datasheet are not the figures you get
The driver’s head is a long way from the platform, and roll and sway share a budget there
In a car buck the driver’s head sits well over a metre above the platform interface. Every figure quoted at the interface arrives at the head multiplied by that lever arm, which is good news for roll cues and bad news for everything competing with them.
The 8-Line Driving Rig Brief Eight answers, and a quotation stops being guesswork
| Line | What it decides on the platform | If it is left blank |
|---|---|---|
| What the rig is for, and who judges the result | Whether the platform’s own response has to be characterised and handed over | A training rig is quoted and a development rig is needed |
| The driver’s eye point above the platform interface | Every travel figure, and how much sway each degree of roll costs | Figures quoted at the frame, and manoeuvres that stop fitting later |
| Gross moving load: cab or buck, seat, wheel, pedals, screens, driver, cabling | Actuator force, onset bandwidth and how much travel survives | Sized for a bare buck and delivered under a trimmed cab |
| The vehicle model: outputs, frame, rate, units, and gravity | The interface, the convention sheet and half the commissioning time | A rig that moves correctly in five channels out of six |
| Where the cueing filter runs, and who tunes it | Controller architecture, and the phone number a driver’s complaint reaches | An argument at commissioning with nothing written down to settle it |
| The manoeuvre list, marked for which must feel complete rather than started | Whether the programme needs a long axis, or a smaller platform and honesty | An expectation no motion system on the market meets |
| Road texture: is a vibration device in scope, and whose is it | The platform’s mass budget and the whole cue split | A larger platform bought to do a job it physically cannot do |
| Duty: sessions per day, participants per day, and the building it lives in | Thermal design, service interval, noise and standby power | A demonstration machine installed in a training centre |
Eight lines, and a driving rig enquiry usually arrives carrying three of them. Mark the other five “unknown” rather than filling them in optimistically: the unknown ones come back as questions, and the optimistic ones get manufactured.
Cannot fill in line four because the model belongs to another department? Send what you have and we will write the questions for them →
From a vehicle model to a rig somebody will sign for
Five steps, and the last one runs on your model rather than ours
Driving rig programmes fail in two familiar places: a convention nobody wrote down, and an acceptance test run on a demonstration file. These five steps close both, and the fifth is the one that matters most.
The 5-Step Model-to-Acceptance Path Each step, its output, and the person waiting for it
ST 01
Read the model
What it outputs, in which frame, at what rate, in which units, and whether gravity is in the vertical channel. One second of real output answers most of it.
Out: model interface note
ST 02
Sign the convention
Every channel, its unit, its sign and the standard it follows, on a single page signed by both sides. This is the page that gets checked on connection day.
Out: convention sheet
ST 03
Split the cue set
Which cue comes from the platform, which from a texture device, which from the steering and which from the seat — agreed before anything is sized, because it changes the mass.
Out: cue allocation
ST 04
Check at the head
The manoeuvre list run against the envelope at your eye point and your load, with each manoeuvre marked onset-only, partial or complete.
Out: manoeuvre report
ST 05
Accept on your model
Acceptance driven by your vehicle model with a driver in the seat, repeated with the spread reported, rather than by a profile that proves only that the machine moves.
Out: acceptance record
Model locked to your network, as most are? Say so at enquiry and the acceptance route is designed around it →
The seam between four suppliers
What we build, what the integrator has to own, and what we turn down
Driving rig scopes go wrong in a familiar way: four suppliers each assume somebody else owns the cue nobody quoted for. These four zones are written into the proposal, so any gap appears while it is still on paper.
Ours
- The mechanism, controller, drives and cabinet
- Motion generated from whatever command level the interface document names
- Mechanical limits, workspace protection and a stop circuit that is hard-wired
- The envelope stated at the driver’s eye point
- Measured onset, and our share of the delay budget
Never ours
- The vehicle model, tyre data and simulation software
- The cab, buck, seats, screens and instrument cluster
- The steering force-feedback system
- The road texture device, unless quoted as a separate item
- Whether the rig trains, validates or proves anything
Outside our class of machine
- Complete driving simulators supplied turnkey
- Certified seats, harnesses and restraint systems
- Rail foundations, civil works and buildings
- Image generation, projection and dome systems
Signed by both
- The axis convention and the sign of every channel
- Where the driver’s eye point sits, and the envelope rewritten at it
- The cue split across platform, texture device, steering and seat
- The manoeuvre list, each entry marked reachable or not
Four suppliers, one driver, and exactly one document that says who owns which sensation. We would rather write it than inherit it.
Integrating four suppliers and not sure where the seams are? Send the architecture and we will mark our edges on it →
Where to go next
Reference reading for a driving rig specification
The application above this one, the sibling that shares its physics, the axis decision and the vocabulary a specification has to use.
Parent application
Motion platforms for simulators
The general case: what a person can detect, how much you may get away with, and how late the whole chain is allowed to be.
Sibling
Flight simulator motion platforms
The same physics with a different horizon, a different acceptance regime, and an authority at the end of it.
Axis decision
Three axes or six, decided by the driving task
What three axes leave out, what six add in moving mass, and which side of that line road vehicle work sits on.
Vocabulary
Motion platform glossary
Gross moving load, motion reference point, washout, tilt coordination, specific force — defined once each, so a specification means the same thing in two companies.
Two outside documents are worth having open while a driving rig interface is being written, and neither of them is ours: SAE J670, vehicle dynamics terminology, and ISO 8855, road vehicle dynamics and road-holding ability vocabulary. Between them they define the axis systems in the crosswalk above. Which one your vehicle model follows is normally printed in its own documentation, and finding out takes five minutes now or three weeks at commissioning.
Asked before a specification is written
Driving simulator motion platform FAQ
Do you supply a complete driving simulator?
No, and you will hear that in the first reply rather than after a quotation. We supply the motion subsystem: mechanism, controller, drives, cabinet, the interface and the records that describe them. The cab or buck, the visual system, the steering force feedback, the seat and the vehicle model are four other suppliers, and each of them knows their part better than we would. Where a customer wants one company for the whole rig, we are the wrong company and it is cheaper to establish that in week one.
Can a motion platform actually reproduce braking?
No, and the arithmetic is on this page so you can check it rather than take our word for it. Holding 0.8 g for a three and a half second stop would need about 15.7 m of surge travel. Tilting the cab to fake it needs 38.7°, and at the usual 3 °/s washout limit that angle takes 12.9 seconds to reach — nine seconds after the car has stopped. What a platform delivers is the onset: a fast, correct specific force for a few hundred milliseconds while the driver decides whether the car is behaving. Every supplier is subject to the same table. Some are clearer about it than others.
3DOF or 6DOF for a driving rig?
Six, for almost all road vehicle work, because braking and cornering arrive in the same second and three axes cannot produce two transients in different directions without one of them lying. Three axes make sense where one plane genuinely dominates — a heavy vehicle trainer built around body roll, or a human-factors study where the motion only has to load the driving task rather than reproduce it. The honest test is your manoeuvre list: if two cues in different directions have to be true at the same instant, the decision is already made.
Our vehicle model uses a different axis convention from yours. Is that a problem?
It is a question, not a problem, and it is the one we ask first. The classic SAE J670 system has y to the right and z downward; ISO 8855 has z upward and y to the left. Two of three axes change sign between them, and the rotations follow. We adopt whichever convention your model uses, write every channel with its unit and sign on one page, both sides sign it before manufacture, and each channel is verified in isolation on the day the model is first connected. A sign error found that afternoon costs an afternoon. Found six months later it has usually been compensated somewhere else first.
Do you supply the steering force feedback and the seat vibration device?
No to both, and they are worth separating because they solve different problems. Steering torque is a loop an order of magnitude faster than anything a motion platform does, closed through the driver's hands rather than their inner ear, and it belongs to a specialist. Road texture lives well above the frequency band a machine carrying a cab can reach, and the right answer is a small vibration device under the seat or the floor rather than a bigger platform. Where a rig needs both, the cue split is agreed before either is sized, because it changes the mass we are carrying.
Can the platform run from our own vehicle model in real time, and do you need to see it?
Yes to the first, and no to the second. The command level is fixed at the interface review — you can send vehicle state and let our controller produce the cue, or run your own cueing filter and send platform commands. What we need is the model's output specification: channels, units, frame, rate and behaviour at the edges. We do not need the tyre data, the parameter set or anything else a competitor would want, and no project has required it. Where the model cannot leave your network at all, say so at enquiry and the acceptance route is designed around that.
What latency should a driver-in-the-loop rig have?
Ask for three figures rather than one, because there are three loops. The motion loop is bounded by what a person perceives as motion and is the slowest of the three; the visual loop is bounded by frame rate; the steering torque loop is the fastest by a wide margin because a hand resolves far higher frequencies than an inner ear. Our share of the motion loop is a fixed property of the controller and the axis count and is issued as a number in the delay budget. What matters more than the absolute figure is that the platform is not conspicuously later than the other two — a car whose wheel loads up before the body leans feels disconnected immediately.
A competing quotation offers the same travel and payload for less. Where does the difference sit?
Compare the conditions before the figures. Ask where the travel was quoted — the mounting interface or the driver's eye point — and at what load, including the cab, screens and occupant. Ask what the onset time is and how it was measured. Ask whether acceptance runs on your vehicle model or on a demonstration profile. Ask whether the axis convention work and the channel-by-channel verification are included or billed as commissioning support. Those four account for most of the gaps we see. Where a competing quotation genuinely covers the same conditions and costs less, we would rather hear about it than guess.
Commercial questions — lead time, payment terms, spares, installation — are answered in the full FAQ.
Send one second of model output. We will tell you what the rig can feel like.
A short sample of real model output, the manoeuvres that have to convince a driver, the load including the occupant and how high the head sits are enough for a first pass. If the programme needs a complete simulator rather than a motion subsystem, that is what you will hear back.
What comes back first
- Which convention your model output is in, read from the sample rather than assumed
- Your manoeuvre list marked manoeuvre by manoeuvre: onset only, partial, or complete
- The envelope restated at your driver’s eye point, with the height it assumes
- The cue split — what belongs to the platform and what needs a different device