Motion cueing is the software that decides how a motion platform moves. It takes the forces a simulation calculates - a long banked turn, a curb strike, the lateral load of a slide - and translates them into movements a platform with a few centimeters of travel can actually produce, while still convincing the inner ear that the full force is there. It is the layer that turns raw telemetry into a sensation the body can learn from.
Two platforms with identical specifications can feel completely different because of this layer alone. Same actuators, same degrees of freedom, same payload rating. One feels smooth and natural; the other feels abrupt and artificial. The difference sits in cueing quality, and cueing quality is the one thing a specification sheet does not show. This article explains what motion cueing does, how the algorithm splits and reshapes the signal, why a fixed-travel platform can suggest a turn that lasts far longer than its stroke, and why cueing matters more than axis count. The examples come from two product lines built by Motion Systems: Qubic System, an industrial-grade compact motion system for sim racing, VR simulators, driver-training, and R&D builds, and the Professional Series for heavy training cabins.
Key takeaways
- Motion cueing translates large, sustained real-world forces into short platform movements while preserving the sensation - it is the reason a few centimeters of travel can feel like a sustained turn.
- Tilt coordination uses gravity as a substitute force, tilting the platform so the inner ear reads a lasting acceleration the actuators could never sustain on their own.
- A well-tuned 3DOF platform outperforms a 6DOF platform running a generic profile; cueing quality, not axis count, decides whether motion sharpens a skill or only shakes the seat.
What is motion cueing in a simulator?
Motion cueing is the translation layer between what the simulation calculates and what the platform physically does. The simulation produces a continuous stream of forces and accelerations as if the vehicle were real. The platform can only move within a fixed mechanical envelope. Motion cueing is the algorithm that bridges that gap, frame by frame, deciding which forces to reproduce directly, which to fake with tilt, and which to discard.
Think of it as a real-time translator. A literal translation - moving the platform exactly as the vehicle moves - is impossible, because a real vehicle travels across hundreds of meters while the platform has centimeters of stroke. So the algorithm translates the meaning of the motion instead: the onset of a brake, the build of a corner, the texture of a rough surface. Done well, the machinery stays invisible and the operator feels a vehicle. Done poorly, the operator feels a machine working against the scene.
That is why cueing sits at the center of platform quality. Actuators and degrees of freedom set the physical ceiling for what is possible. The cueing algorithm decides how much of that ceiling the operator actually feels, and how faithfully it maps to the forces of real operation.
How does a motion cueing algorithm work?
A motion cueing algorithm works by running incoming telemetry through a four-step chain and then splitting each axis into separate signal paths tuned to the physics that axis reproduces. The chain is the same on every platform; the quality of the split is what differs.
The four steps run continuously while the simulator is active:
- Telemetry. The simulation streams real-time data: velocity, acceleration, angular rate, orientation, terrain contact, impacts. The refresh rate of this stream sets the ceiling for everything downstream. A feed at 100 Hz carries ten times the detail of one at 10 Hz, which is the difference between a curb strike arriving as a trackable curve and the same strike arriving as one blunt event.
- Motion cueing. The algorithm reshapes those forces into commands the platform can execute without hitting its limits. The engineering complexity concentrates here, and so does the variation between platforms.
- Actuator execution. The cueing output becomes position commands sent to each actuator many times per second, synchronized to the visual display.
- Perception. The operator receives forces aligned with the scene, and over repeated sessions the body builds muscle memory for the real vehicle.
Inside step 2, the algorithm does not treat all telemetry the same. Each input channel - surge, sway, heave, pitch, roll, yaw - passes through its own signal path, because a sharp jolt and a slow sustained load need opposite treatment.
Definition - signal path: the dedicated processing route an algorithm applies to one axis of motion. Splitting motion into separate high-frequency and low-frequency paths lets a platform reproduce a sharp bump and a long corner at the same time, instead of averaging them into one muddy movement.
High-pass and low-pass motion filters: splitting the work
A motion cueing algorithm splits each axis into two main paths - a high-pass filter for short, sharp cues and a low-pass filter for long, sustained ones - plus logic that keeps the platform inside its limits. This split is the core mechanical trick of cueing.
- High-pass path - onset cues. A high-pass filter keeps the sharp, short-duration part of an acceleration and discards the long tail. This is what the platform reproduces by physically moving: the snap of a gear change, the jolt of a curb, the first instant of braking. These transients are exactly what the body uses to time its reactions, and they are short enough that the platform can produce them directly with stroke to spare.
- Low-pass path - sustained forces. A low-pass filter keeps the slow, long-duration part - the steady lateral load of a long corner, the sustained pull of acceleration down a straight. The platform cannot reproduce these by moving, because they last far longer than its travel allows. So this path feeds tilt coordination instead, covered in the next section.
- Limit-handling logic. This monitors position and velocity against the actuator envelope and smoothly eases commands as the platform approaches its limits, rather than letting it slam into a hard stop and produce a visible jolt.
- Cross-axis coordination. A braking corner combines surge, sway, and roll at once. Without coordination, those axes fight each other and the motion turns incoherent. Good cueing keeps them working together.
The quality of these filters is everything. A platform with weak filtering produces false cues - movements in directions the scene never indicated - which is the fastest way to make a simulator feel wrong and to make trainees sick. Good filtering stays invisible: the platform responds, returns to neutral below the threshold the inner ear can detect, and is ready for the next cue with full range available.
What is tilt coordination in motion cueing?
Tilt coordination is the technique that lets a fixed-travel platform reproduce a sustained acceleration by tilting into it and borrowing gravity as a substitute force. It is the single most important idea in motion cueing, and the reason a platform with a small stroke can suggest a turn that lasts far longer than its actuators could ever reach.
Here is the physics, in plain terms. A platform can only sway sideways for a few hundred milliseconds before it runs out of travel, so it cannot reproduce the sustained lateral load of a long corner by moving alone. But it can tilt the operator. When the platform rolls into the direction of the perceived acceleration, the operator's head is no longer perpendicular to the floor, and a component of gravity now pulls along the body axis - the same direction a real cornering force would pull. The inner ear cannot distinguish that gravity component from genuine lateral acceleration, so the brain reads it as a sustained turn. The tilt happens slowly enough to stay below the threshold the vestibular system notices, so the operator feels the force without feeling the platform rotate.
This is the mechanism behind one of the most common buyer questions.
Why can a motion platform not reproduce sustained g-force?
Sustained g-force requires sustained acceleration, and sustained acceleration requires distance - a real cornering car covers hundreds of meters while the load holds. A platform has centimeters of travel, so it physically cannot maintain the displacement that a real sustained force demands. No ground-based system can. Tilt coordination is the workaround: it does not produce the real force, it produces a gravity cue the inner ear accepts as that force. This is also why a motion rig sometimes moves more than the real vehicle's suspension would - it is reshaping forces for perception, not copying suspension travel.
The honest limit matters here. Tilt coordination convinces the inner ear, not the body's full sense of load. It reproduces the direction and onset of sustained force convincingly; it cannot reproduce the crushing magnitude of high sustained g the way a centrifuge does. For training, that is usually enough, because the skill being learned is recognizing and reacting to the onset and direction of force - which is exactly what tilt coordination delivers.
Classical, adaptive, and predictive cueing approaches
Most motion cueing in service today is a variant of the classical washout approach: fixed high-pass and low-pass filters with tilt coordination, tuned once per platform and use case. It is well understood, predictable, and reliable, which is why it dominates. Its weakness is that the tuning is a compromise - filters set for a hard-braking scenario are not optimal for a gentle cruise, and the algorithm cannot tell the difference.
Adaptive and predictive approaches try to close that gap. Adaptive cueing adjusts filter parameters on the fly based on how much travel remains, squeezing more usable motion out of the same envelope. Predictive (model-based) cueing looks ahead at the upcoming demand and pre-positions the platform so the next big cue has range available. Both can improve fidelity, and both add complexity and tuning effort. For most training and sim-racing applications, a well-tuned classical approach delivers the sensation the task needs without the extra risk - which is the recurring lesson of cueing: a simpler algorithm tuned well beats a sophisticated one tuned poorly.
How motion cueing is tuned for a specific simulator
Cueing is not a single setting switched on once. Each axis carries its own filter parameters - cutoff frequencies, gains, tilt-rate limits - and those parameters are matched to two things: the platform's physical envelope and the forces the application produces. A sim-racing profile tuned around a roughly ten-degree roll envelope emphasizes sharp onset cues and fast return. A flight-cabin profile with a larger envelope leans harder on sustained tilt for long, slow maneuvers. The same algorithm, tuned differently, produces a different vehicle.
Engineering origin matters here. On both Motion Systems lines the cueing runs on an algorithm stack called ACE (Acceleration Control Engine), a modified classical washout that operates at 250 Hz. Most simulation hosts output telemetry at around 60 frames per second; ACE interpolates that stream up to 250 Hz, so the platform moves on smooth sub-frame steps instead of jumping between frames. Each channel passes through independent high-pass, low-pass, limit-handling, and tilt-coordination blocks, with per-channel parameters set for the specific platform model. Because the hardware and the cueing come from the same team, the tuning is a model-specific calibration rather than a generic profile dropped onto whatever platform happens to run it. The same engine scales from compact sim-racing rigs, where frame-to-frame consistency decides whether a fast lap feels repeatable, up to professional training cabins and high-fidelity research devices.
Why cueing quality matters more than axis count
Cueing quality matters more than axis count because the human sensory system audits coherence, not specifications. A well-tuned 3DOF platform with a quality cueing algorithm outperforms a 6DOF platform running a generic washout filter. The reasoning runs in three directions.
On the hardware side, 6DOF extends the motion envelope: the platform can now reproduce lateral slides, linear accelerations, and yaw rotations that 3DOF cannot. That expansion only pays off if the cueing can use the extra axes coherently. A 6DOF platform running an algorithm that does not coordinate all six axes leaves the extra three either unused or poorly synchronized with the primary three - paid for, but not felt.
On the perception side, the operator cannot tell that a platform has six actuators. The operator can tell whether the motion feels natural. A 3DOF platform with cueing tuned to handle pitch, roll, and heave convincingly produces the subjective experience of a well-behaved vehicle. A 6DOF platform with rough cueing produces the subjective experience of machinery wedged between the scene and the body. The axes are invisible; the quality of motion is not.
On the outcome side, training transfer depends on whether simulated forces align with real ones. Quality cueing produces that alignment; poor cueing produces misalignment, and misalignment degrades training even when every axis is present. The worst case is well documented in motion research: a poorly cued platform that drives the body in the wrong direction produces more discomfort than a static simulator, which means it produces worse training outcomes than no motion at all. The investment becomes a liability until the cueing is right.
The practical implication for a buyer is to treat cueing as a first-class question. Pick the degrees of freedom the task requires - and no more - then spend the rest of the evaluation on cueing: at what rate it runs, whether it is tuned per model or generically, how it coordinates simultaneous cues, and whether the supplier can show evidence it has been validated against comfort measurements. A supplier who answers those questions with detail has engineering depth on the side that actually decides motion quality. A supplier who deflects to axis count and payload numbers does not.
How latency and cueing work together
Cueing decides how the platform moves; latency decides when it moves relative to the scene. Both have to be right. The best cueing in the world feels broken if it arrives late, because the body feels the corner a beat after the visual cues show it, and that loss of synchronization is exactly what triggers sickness. This is why response time is a hard design target, not a marketing line.
The numbers differ by line, and the attribution matters. Across the Qubic System range, end-to-end latency runs below 8 milliseconds (the QS-H13 layered system runs below 10), fast enough that a sim racer feels the rear step out the instant the screen shows it. The Professional Series runs below 30 milliseconds end-to-end, measured, which keeps a full training cabin comfortably inside the threshold where motion stays an asset. Cueing latency is one component of those totals - the algorithm has to compute its output fast enough not to eat the budget. Strong cueing and low latency are two halves of the same goal: motion that the body reads as part of the scene, not a reaction to it.
Frequently asked questions
What is motion cueing and why is it needed?
Motion cueing is the software that translates a simulation's calculated forces into movements a motion platform can physically produce. It is needed because a real vehicle generates forces across hundreds of meters of travel, while a platform has only centimeters of stroke. Without cueing, the platform would hit its limits within seconds and stop producing any usable motion, so the algorithm reshapes large forces into short, contained movements that still feel right.
How does tilt coordination simulate g-forces?
Tilt coordination tilts the platform into the direction of a perceived acceleration so that gravity pulls along the operator's body axis, in the same direction a real force would. The inner ear cannot tell that gravity component apart from genuine lateral or longitudinal acceleration, so it reads the tilt as a sustained force. The tilt is applied slowly enough to stay below the threshold the vestibular system can detect, so the operator feels the force without feeling the platform rotate.
Why does a simulator move differently than the real vehicle?
Because cueing reshapes forces for perception rather than copying the vehicle's motion. A platform reproduces the sharp onset of a cue directly, then quietly returns to neutral - a movement the real vehicle never makes - so it has range ready for the next cue. It also tilts to fake sustained forces. The goal is to deliver the sensation the body uses to react, not to mirror suspension travel or chassis movement one-to-one.
Can a motion platform reproduce sustained cornering force?
Not as a literal force, but it can reproduce the sensation convincingly. Sustained cornering force requires sustained lateral acceleration, which requires more distance than any platform has in travel. Tilt coordination solves this by using gravity as a substitute, so the inner ear perceives a lasting turn. The direction and onset feel accurate; the raw magnitude of high sustained g cannot be reproduced by a ground-based platform, only approximated.
What is the difference between cueing and washout?
Motion cueing is the whole translation system - filters, tilt coordination, limit handling, and cross-axis coordination working together. Washout is one technique within it: the slow return of the platform to its neutral position between cues, timed below the threshold the inner ear notices, so range is always available for the next movement. In short, washout is part of cueing, not a separate alternative to it.
How is motion cueing tuned for a specific simulator?
Each axis gets its own filter parameters - cutoff frequencies, gains, and tilt-rate limits - matched to both the platform's physical envelope and the forces the application produces. A sim-racing profile emphasizes sharp onset cues within a small roll envelope; a flight-cabin profile leans on sustained tilt within a larger one. The strongest tuning comes from suppliers whose cueing and hardware are engineered together, so the parameters fit the specific model rather than a generic average.
