- Home
- Technology
- How a 6DOF Motion Platform Works
How a 6DOF Motion Platform Works
A 6DOF motion platform controls the position and orientation of a rigid moving platform in X, Y, Z, roll, pitch and yaw. Six actuators connect that platform to a fixed base; a controller converts each commanded pose into six actuator lengths, then synchronized servo loops move and correct the actuators until the requested motion is produced.
6 controlled DOF
Three translations and three rotations of one rigid body.
6 coupled actuators
No actuator belongs to only one platform axis.
Inverse kinematics
Closed-form: pose in, six lengths out.
Forward kinematics
No closed-form solution — must be solved numerically.
What Is a 6DOF Motion Platform?
A six-degree-of-freedom motion platform is a mechatronic system that controls the position and orientation of a moving rigid body. The term 6DOF describes a motion capability. By itself it says nothing about the mechanism, the actuator technology, the control architecture or the application.
Direct answer. A complete 6DOF platform combines a load-bearing parallel mechanism, six actuators with drives, a motion controller running kinematics in real time, position feedback, software limits, a host interface and an agreed verification method. The mechanism produces useful motion only when all of those layers behave as one coordinated system.
6DOF Describes Motion
X, Y and Z define translation. Roll, pitch and yaw define orientation. Together these six generalised coordinates fully describe the pose of a rigid body in three-dimensional space — no more coordinates are needed, and no fewer will do.
Stewart Describes a Mechanism
A Stewart or hexapod mechanism connects one moving platform to one fixed base through six variable-length limbs arranged in parallel. It is the most common way to achieve 6DOF, but not the only one.
Motion Platform Describes a System
The platform includes structure, actuators, drives, controller, software, host interface, limits, feedback and the acceptance method that proves it does what was specified.
Two distinctions worth keeping: not every 6DOF mechanism is a Stewart platform — serial robot arms and crank-arm designs also reach six axes — and not every Stewart platform is used as a simulator. Precision positioning hexapods use the same architecture for micron-level alignment. This guide explains the parallel-kinematic architecture used in CSCMotion 6DOF platforms.
The Six Degrees of Freedom: Surge, Sway, Heave, Roll, Pitch and Yaw
The axis names come from naval architecture and are used throughout simulation and vehicle dynamics. They are only useful once the coordinate frame, positive direction, rotation convention and reference point are stated — different host systems genuinely disagree about these.
| Coordinate | Type | Common name | Physical description | Everyday example |
|---|---|---|---|---|
| X · Translation | Linear | Surge | Movement forward and backward along the longitudinal axis. | A car accelerating or braking |
| Y · Translation | Linear | Sway | Movement left and right along the lateral axis. | A vehicle sliding sideways |
| Z · Translation | Linear | Heave | Movement up and down along the vertical axis. | A lift starting or a wave lifting a hull |
| Rx · Rotation | Angular | Roll | Rotation about the longitudinal axis. | An aircraft banking into a turn |
| Ry · Rotation | Angular | Pitch | Rotation about the lateral axis. | A nose rising on take-off |
| Rz · Rotation | Angular | Yaw | Rotation about the vertical axis. | A vehicle rotating about its centre |
Why six, and why exactly six?
An unconstrained rigid body in three-dimensional space has exactly six independent ways to move: three translations and three rotations. Any motion, however complex, decomposes into these six. This is why 6DOF is the complete case — a platform with six controlled degrees of freedom can reproduce any rigid-body motion within its physical limits.
Axis labels are not a standard
Positive pitch may be nose-up in one system and nose-down in another. Rotation order (for example Z-Y-X versus X-Y-Z) changes the resulting orientation for the same three angles. Units, origin and reference point vary too. All of this belongs in the project interface document, agreed before integration — not discovered at first power-on.
The Parallel Mechanism: Six Limbs Carrying One Platform
In a parallel kinematic machine, every limb connects the base directly to the moving platform. Load is shared across all six paths rather than passing through a chain of joints, which is why these mechanisms are stiff and accurate relative to their mass — and why the mathematics is harder than a serial robot arm.
Almost every published explanation of a Stewart platform assumes spherical (ball) joints at both ends of each limb. That is the textbook form, and it makes the mathematics clean. It is not what we build.
CSCMotion platforms use Cardan (universal) joints. Combined with the prismatic and rotational freedom of the piston inside the cylinder barrel, this gives each limb a T-P-S kinematic chain — which is where the designation 6-TPS comes from. The distinction is not academic; it changes the load path, the stiffness and the workspace boundary.
Spherical joint · 3 rotational DOF
Free to rotate about all three axes, including about the limb’s own axis. Clean to model, but that free spin means the limb cannot resist torsion, and off-the-shelf ball joints with adequate stiffness at high load become large and expensive.
Cardan joint · 2 rotational DOF
Constrains rotation about the limb axis. The result is a determinate load path and higher stiffness under alternating load — the governing condition for fatigue life in a motion base. The trade is a more restricted articulation range, which must be respected during workspace analysis.
Moving platform
Carries payload, fixture or cabin
Moving platform
Carries payload, fixture or cabin
Moving platform
Carries payload, fixture or cabin
Moving platform
Carries payload, fixture or cabin
Why parallel instead of serial?
A serial robot arm stacks joints in a chain, so each motor carries everything downstream of it and errors accumulate along the chain. In a parallel mechanism the six limbs share the load and the errors average rather than accumulate. That is the source of the stiffness and repeatability advantage — and the reason a hexapod has a small workspace relative to its footprint.
Geometry sets the trade-off
Joint circle radii, limb inclination and the angular offset between upper and lower joint pairs decide how workspace, stiffness and force capability are balanced. Widening the base improves tilt stiffness but consumes floor space and reduces the achievable rotation range.
How a 6DOF Motion Platform Works: From Command to Measured Motion
The clearest way to understand a 6DOF platform is to follow a single target pose from the host all the way to a measured result. Hardware and software responsibilities vary between projects, but this engineering path stays the same.
01 · Define the Target Pose
A host, trajectory file, test controller or cueing system requests X, Y, Z, roll, pitch and yaw over time.
02 · Validate the Command
Units, frames, machine state, timing, pose limits and rate limits are checked before the command is accepted.
03 · Solve Inverse Kinematics
The controller converts platform pose into six target limb lengths using the calibrated geometry.
04 · Generate Axis References
Interpolation plus velocity, acceleration and jerk limits turn target lengths into executable synchronized motion.
05 · Move the Actuators
Servo drives and motors extend or retract the electric cylinders to follow the axis references.
06 · Correct With Feedback
Measured actuator position is compared with the reference and the servo loop closes the error every cycle.
07 · Verify Platform Motion
Response is checked at an agreed reference point, with the approved load and an agreed test method.
Step 01 · command source
The host is not always doing the same job
A test system may replay a predefined trajectory. A research model may generate motion in real time. A professional simulator inserts motion cueing between its vehicle model and the platform command. These are genuinely different application layers, and confusing them is a common source of specification error.
- Who owns the command, and at what update rate
- Which units and coordinate frame it uses
- Who enforces each limit
- What happens on fault or loss of communication
Step 02 · verification
“The platform moved” is not an acceptance criterion
Useful evidence compares commanded against measured motion for a stated payload, reference point, trajectory, instrument and tolerance. A fast no-load demonstration video proves only that the mechanism is capable of moving.
- Approved moving assembly, not an empty platform
- A representative combined trajectory
- An independent instrument where required
- A traceable, repeatable test record
Why Six Actuators Do Not Equal Six Independent Platform Axes
Surge, sway, heave, roll, pitch and yaw describe the moving platform. Limb length describes the mechanism. In a parallel platform these two coordinate sets are related through geometry, never by a one-to-one assignment.
A change in one limb length contributes to several platform motions at once. Conversely, a requested pure platform motion normally requires several or all six limbs to change together — and the pattern shifts with the current pose and the attachment-point geometry.
Pure request
Heave
All six limbs extend together, but not by equal amounts — the pattern depends on limb inclination and current pose.
Pure request
Roll
Limbs on one side extend while the others retract, in a coupled pattern set by geometry, pivot location and pose.
Combined request
Translation + Rotation
The controller solves the full pose as one problem. Isolated per-axis maxima cannot simply be added together.
Mechanism response
Limb Contribution
One limb does not own one DOF. Each contributes to the force and motion balance of the whole platform.
How Inverse Kinematics Converts a Target Pose Into Actuator Lengths
Inverse kinematics answers the real-time control question: given the required platform position and orientation, what length must each limb have? For a parallel platform this is the easy direction — it is a direct geometric calculation with a single answer, and it runs every control cycle.
For limb i, the controller works from its base attachment point, its platform attachment point, the target translation and the target rotation.
Principle equation
bi— base attachment point, base framepi— platform attachment point, moving framet— target translation vectorR— target rotation matrixli— limb vector in the base frameLi— required limb length
Read plainly: take the platform attachment point, rotate it into the commanded orientation, shift it by the commanded translation, then subtract the fixed base point. What remains is the vector the limb must span — and its magnitude is the length the actuator must reach. Repeat six times and the pose is fully resolved.
Why this direction is the easy one
Each limb is solved independently — limb 3 does not need to know what limb 5 is doing. Six separate vector subtractions, six magnitudes, done. There is exactly one correct answer and it is reached in a fixed number of operations, which is what makes real-time control at high update rates practical.
What the controller does next
Target lengths alone are not executable. The controller must still generate a time-parameterised trajectory for each axis, respecting velocity, acceleration and jerk limits, and keep all six axes synchronized — otherwise the platform passes through poses that were never commanded.
Calibration matters more than the formula
The equation assumes the attachment points are exactly where the drawing says. In a real machine they are not. Identifying the as-built geometry and compensating it in the controller is what turns a correct equation into an accurate machine.
The Forward Problem: Six Lengths Back to One Pose
Run the question backwards. You have six measured limb lengths from the encoders. Where is the platform? Most published introductions to Stewart platforms move straight past this point. It deserves attention, because it is the harder half and it shapes what a platform can honestly claim about its own position.
Inverse · pose → lengths
Closed form, one solution
Each limb solved independently by vector subtraction. Deterministic, fast, exactly one answer. This is what runs in the real-time control loop.
Forward · lengths → pose
No general closed-form solution
Six coupled non-linear equations must be satisfied simultaneously. For a general Stewart geometry there is no formula that returns the pose directly, and the system admits up to 40 mathematically valid solutions — most of which are physically impossible assemblies the mechanism could never reach.
The asymmetry is a property of parallel mechanisms in general, and it is the mirror image of the serial robot case. For a robot arm, forward kinematics is trivial (chain the joint transforms together) and inverse kinematics is the hard, multi-solution problem. A hexapod inverts that difficulty exactly.
Because there is no formula, the forward problem must be solved numerically — and numerical solvers need a good starting estimate, or they converge slowly, or to the wrong root, or not at all. In a real-time control loop, “not at all” is not an acceptable outcome.
This is worth knowing when you read any platform’s specification. A stated pose is either the pose that was commanded, a pose calculated from a mechanism model, or a pose measured by an external instrument. Those are three different claims, and only the third is independent of the model’s assumptions.
How Closed-Loop Control Keeps the Six Actuators Synchronized
Closed-loop control is often described as though one feedback value proved the complete platform pose. In practice, actuator feedback, calculated platform state and independent measurement are three separate evidence layers that answer three different questions.
Actuator Feedback
Encoders or linear position sensors report actuator-side motion to the servo drive, which closes the error between the axis reference and the measured response every control cycle. This is what keeps the six limbs synchronized with each other.
Calculated Platform Pose
The controller estimates platform pose from the mechanism model and the six actuator positions — the forward kinematics problem. The result is only as good as the geometry, calibration, joint behaviour and modelling assumptions behind it.
Independent Verification
An external measurement system verifies motion at the platform or payload reference point, under an approved load and trajectory. This is the only layer that does not depend on the mechanism model being correct.
Why the distinction matters: structural deflection, joint compliance, mounting flexibility, backlash, payload dynamics, residual geometry error and cable forces can all create a real difference between the commanded pose, the pose the controller calculates, and the motion an instrument measures at the payload. On a well-built platform these differences are small — but “small” is a measured result, not an assumption.
What Limits the Usable 6DOF Workspace?
A parallel platform does not offer maximum translation and maximum rotation on every axis at the same time. The usable workspace is a coupled volume shaped by geometry, load, motion and safety constraints at once — which is why it cannot be described by a simple list of per-axis maxima.
01 · Limb Length
Minimum and maximum actuator lengths define part of the reachable boundary.
02 · Joint Articulation
Cardan or spherical joints must stay inside their permitted angular range at every pose.
03 · Mechanical Clearance
Limbs, joints, platform, base, payload and surrounding equipment must not interfere.
04 · Force & Speed
A pose can be geometrically reachable while exceeding actuator force, velocity or acceleration limits.
05 · Payload & Centre of Gravity
Mass distribution and inertia change limb loading and dynamic response across the workspace.
06 · Conditioning
Near singular configurations, force transmission degrades and controllability can be effectively lost.
07 · Cables & Utilities ·
Routing, bend radius, drag and strain relief reduce the practical motion envelope.
08 · Safety Margin
Software limits should retain margin from every mechanical, joint, load and control boundary.
Do not add isolated maxima
Each figure in a specification table is normally measured with the other axes near neutral. Maximum surge, maximum pitch and maximum heave are not simultaneously available, and adding them describes a pose the platform cannot reach. Before any specification is agreed, a representative combined trajectory has to be checked through the actual geometry and load model.
≠
MAX X + MAX Y + MAX Z
+ MAX ROLL + MAX PITCH + MAX YAW
Where Motion Cueing Fits — and Where It Does Not
Motion cueing is not inverse kinematics and it is not servo control. A cueing algorithm converts vehicle or flight-model dynamics into bounded platform motion intended to create convincing perceptual cues inside a workspace far smaller than the motion being represented.
It sits above everything described in this guide. Cueing decides what pose to request; inverse kinematics then converts that request into limb lengths, and the controller still has to validate, synchronize and execute it.
Many platforms never use cueing at all. Industrial test systems normally command a defined physical trajectory directly, because the goal is to reproduce a specific motion accurately, not to create a perception of motion.
Industrial testing · usually no cueing
Test profile or real-time test controller → platform pose command → inverse kinematics → servo motion. The commanded motion is the goal in itself.
Research motion · project dependent
Experiment model, recorded field data or scripted trajectory → optional processing → platform pose command.
Professional simulation · cueing normally present
Vehicle or flight model → motion cueing → bounded platform pose command → platform subsystem. Sustained acceleration is represented by tilt coordination within a finite envelope.
Common Misunderstandings About 6DOF Motion Platforms
Most specification mistakes start with a correct idea applied without its engineering conditions. These six distinctions prevent the working principle from turning into a misleading requirement or product claim.
Misconception 01
“Each actuator controls one DOF.”
Correction: actuator motion is coupled through geometry. Several or all six limbs normally contribute to any single requested platform motion.
Misconception 02
“6DOF means unlimited motion.”
Correction: every physical platform has coupled limits in stroke, angle, force, speed, joint articulation and clearance. Six axes describes completeness, not magnitude.
Misconception 03
“Actuator stroke equals platform travel.”
Correction: platform translation and rotation are geometric results of all six limb lengths and the attachment-point layout. Stroke is an input to that calculation, not its output.
Misconception 04
“Encoder feedback proves payload pose.”
Correction: encoder data proves actuator-side response. Independent measurement at the payload reference point is a separate and sometimes necessary evidence layer.
Misconception 05
“Every 6DOF platform needs motion cueing.”
Correction: cueing is application-side processing used mainly in simulation. Many test and research systems command trajectories directly and never need it.
Misconception 06
“The mechanism alone determines performance.”
Correction: geometry, structure, actuators, load, controller, calibration, installation and measurement conditions all act together. A good mechanism poorly calibrated is a poor machine.
6DOF Motion Platform FAQ
Short answers to the questions engineers most often ask about 6DOF working principles.
What is a 6DOF motion platform?
A 6DOF motion platform is a mechatronic system that controls the position and orientation of a moving rigid body in six degrees of freedom: three translations (surge, sway, heave) and three rotations (roll, pitch, yaw). In the common parallel-kinematic form, six actuators connect a moving platform to a fixed base, and a controller converts each commanded pose into six actuator lengths.
How does a 6DOF motion platform work?
A host requests a pose in X, Y, Z, roll, pitch and yaw. The controller validates that command against its limits, solves inverse kinematics to obtain six target limb lengths, generates synchronized axis trajectories, and drives the servo actuators. Encoder feedback closes the loop on each axis, and platform-level motion is verified against the command using an approved load and instrument.
Does each actuator control one degree of freedom?
No. In a parallel mechanism the actuators are coupled. A single requested platform motion such as pure heave normally requires several or all six actuators to change length together, and the coupling pattern changes with the current pose and the attachment-point geometry. Treating the platform as six independent linear axes produces incorrect motion and can overload joints.
Why do CSCMotion platforms use Cardan joints instead of ball joints?
Most published descriptions of Stewart platforms assume spherical or ball joints. Our platforms use Cardan (universal) joints, which is where the 6-TPS designation comes from. A Cardan joint constrains rotation about the limb axis, giving a determinate load path and higher stiffness under alternating load — the condition that governs fatigue life in a motion base. The trade-off is a more restricted articulation range that has to be respected during workspace analysis.
Why is forward kinematics harder than inverse kinematics?
Inverse kinematics — pose to six limb lengths — is a direct closed-form calculation, solved independently for each limb. The forward problem — six measured lengths back to one pose — has no general closed-form solution for an arbitrary Stewart geometry and admits many mathematically valid roots, most of them physically unreachable. It must be solved numerically. We seed a numerical iteration with a neural-network estimate and validate the result against simulation. See Hexapod Kinematics.
Can I add the maximum values from a 6DOF specification table?
No. Maximum surge, heave and pitch are each measured with the other axes near neutral. The usable workspace is a coupled volume, so maxima on different axes are not simultaneously available. A representative combined trajectory has to be checked through the actual geometry and load model before a specification is agreed. See Payload, Stroke & Acceleration.
Does every 6DOF motion platform need a motion cueing algorithm?
No. Motion cueing is an application-layer function used mainly in vehicle and flight simulation, where sustained accelerations must be represented inside a finite workspace. Industrial test systems normally command a defined physical trajectory directly, with no cueing layer at all.
Does encoder feedback prove the payload reached the commanded pose?
Encoder feedback proves actuator-side position. Structural deflection, joint compliance, mounting flexibility, geometry error and payload dynamics can all create a difference between commanded pose, model-calculated pose and the motion an instrument actually measures at the payload reference point. Independent measurement is a separate evidence layer.
Is a Stewart platform the same as a 6DOF motion platform?
Not quite. Six degrees of freedom describes controlled motion; Stewart platform describes a mechanism architecture. A Stewart platform is the most common way to achieve 6DOF, but serial robot arms and crank-arm designs can also reach six axes, and Stewart mechanisms are also used for precision positioning rather than simulation. See Stewart Platform Design.
Related Motion Platform Engineering Guides
This guide covers the working principle. The next questions — geometry, mathematics, sizing and architecture choice — each have their own guide.
Geometry
Stewart Platform Design
Joint circle radii, limb angles and platform ratios, and how each geometric choice trades workspace against stiffness and force capability.
Mathematics
Hexapod Kinematics
The full inverse and forward kinematics treatment, the Jacobian, singularities and how workspace envelopes are computed.
Selection
3DOF vs 6DOF
What three axes can and cannot reproduce, and how to decide from the motion task rather than from the budget.
Sizing
Payload, Stroke & Acceleration
Why headline numbers are not comparable in isolation, and which inputs actually size a platform.
From principle to specification
Define What Your 6DOF Platform Must Actually Move
Understanding the working principle is the first step. A platform configuration can only be reviewed once the real moving assembly, the combined motion profile and the integration boundary are defined. Send us those and we will tell you what is missing before anyone discusses a model number.
What an engineering review needs
The eight inputs that turn a working-principle question into a reviewable specification:
Physical motion objective
Complete moving assembly
Total mass and centre of gravity
Inertia, where available
Required translations and rotations
Reference point those values apply to
Representative combined trajectory
Duty cycle and host interface