Motion integrity
Every motion has a history.
Normal motion systems treat motion as a command: send it and hope the return value means what you think. KineticON treats important motion as a transaction — with an identity, a persisted life-cycle, a receipt and an answer to the only question that matters after anything goes wrong: what actually happened?
In short: MotionProof records each move from RECEIVED to TRANSACTION COMPLETE inside the controller; StateWitness returns every position with the sensor, time, frame and trust state behind it; FrameProof gives each coordinate frame a version and a dependency graph so the right things — and only those — lose trust when an event occurs; LinkPulse and EventStream make communications a single authoritative stream instead of a crowd of pollers.
MotionProof™ · transactional motion
Motion with a receipt.
Transaction 018F2: MOVE Z TO 84.650 mm. The controller persists each critical transition with the command hash, the machine-configuration hash, the firmware identity, the axis configuration, the coordinate-frame identity, the starting position, the target and the motion profile — and then the final position, the encoder result, the following error, the limit state, the drive status, the brake state and any stop or fault reason.
Now cut the supervisory PC’s link at STARTED. The controller does not care: it owns the motion. When the PC returns, it does not infer truth from whether a DLL call returned. It asks.
- RECEIVEDt+0.012s
- VALIDATEDt+0.192s
- ACCEPTEDt+0.372s
- STARTEDt+0.552s
- TRAJECTORY ACTIVEt+0.732s
- DECELERATIONt+0.912s
- COMMAND POSITION REACHEDt+1.092s
- ACTUAL POSITION SETTLEDt+1.272s
- STATIONARY VERIFIEDt+1.452s
- TRANSACTION COMPLETEt+1.632s
Every critical transition is persisted. If the PC disappears mid-move, it reconnects and asks the controller for the receipt.
// what a conventional API returns Z = 84.650 mm // what KineticON knows about that number value 84.650 mm source load encoder (linear scale), not motor counts timestamp hardware capture, TimeFabric clock generation controller session 41 · config 7c2e… frame HOLDER_624 / rev 4 trust VERIFIED reason encoder continuity maintained since physical frame verification
StateWitness™ · provenance
A number without its provenance is a rumour.
Most APIs return values. KineticON returns values with the evidence behind them: which sensor produced the reading, when, under which controller generation, in which coordinate frame — and a deterministic trust state.
VERIFIEDKNOWNSUSPECTUNKNOWN
This is not an AI probability. VERIFIED means continuity since a physical verification. UNKNOWN means UNKNOWN — and a failed read is never formatted as a number, never reported as an arrival and never saved as a position.
FrameProof™ · coordinate frames with histories
A socket reconnect does not magically restore frame truth.
Machine, home, work, holder, fixture, tool and product frames each carry a version, a source, a timestamp, commissioning evidence and a place in a dependency graph. When an event invalidates a transform, everything that depends on it becomes appropriately untrusted — and nothing else does.
Every frame is VERIFIED at revision 4. Pick an event to see exactly which trust it destroys — and which it does not.
LinkPulse™ & EventStream™ · communications
One authoritative stream. Many subscribers. Nobody hammering the controller.
Communications should be so reliable that nobody thinks about them. The controller produces timestamped state and event streams — position, velocity, drive state, I/O, transactions, warnings, diagnostics, temperature, power, condition metrics — and the UI, the logger, diagnostics, the twin and remote monitoring all consume the same acquisition.
- Normal traffic is proof of life; an adaptive heartbeat fills only genuine idle gaps.
- Transaction latency, jitter, queueing, packet errors, retransmissions, link state and controller load are tracked.
- Explicit reconnect with session generations — no silent command replay, no duplicate execution, transaction de-duplication.
- Dual host interfaces, physical network redundancy and an independent diagnostic link on the architectures that need them.
One timestamped acquisition, published once. Every consumer reads the same authoritative moment.
- STOP AXIS
- VERIFY STATIONARY
- ENGAGE BRAKE
- VERIFY BRAKE
- RELEASE DRIVE TORQUE · when permitted
Stop, prove the stop, engage, prove the engagement, and only then release torque.
Safe brake intelligence
A brake is not a stop.
Vertical axes deserve native architecture. The controller understands motor torque, stationary state, the brake command, brake feedback, the measured release delay, the measured engage delay and the safe torque state — and makes the correct order a built-in primitive instead of a convention the application must remember.
Release runs the mirror image: establish drive holding torque, release the brake, prove the release, permit motion. And STOP, BRAKE, POWER LOSS, STO and SAFE STOP stay what they physically are: five different events.
Motion Capsule™ · controller-owned execution
Hand the controller a sealed program — and let the PC disappear.
A Motion Capsule is a signed, versioned local machine program: the sequence, the trajectory, the permitted frames, the tool and I/O requirements, the limits, the safety policy and the configuration hash it was written for. The controller validates the capsule before accepting it. After launch the supervisory computer may disappear, according to policy, because the deterministic controller owns the execution.
COMPLETE SAFELY
The operation finishes under controller ownership and the transaction closes normally.
STOP SAFELY
The operation is brought to a controlled stop with the reason persisted.
HOLD SAFELY
The machine holds position under proven conditions until the supervisor returns.
The choice is explicit per operation — never an accident of whichever thread happened to be running.
Origin case · PleniDose
Why these subsystems exist.
PleniDose is Panacea’s precision dosing and crimping gantry for Lyoprester dual-chamber cartridges, running on a commodity multi-axis controller reached through a vendor library. Each item below is a real engineering episode from building it, and each is now a KineticON subsystem.
| What PleniDose went through | What it taught | KineticON answer |
|---|---|---|
| The controller link dropped for seconds in the middle of a dry run; every library call returned an error code until it reconnected. | Communication loss must be reported fast, and position after it is unknown until proven. | MotionProof · LinkPulse |
| A read that failed returned a buffer that had never been written — which printed as “X 0.0 mm” and looked like an arrival. | An unknown value must never be formatted as a number, reported as a completion or saved. | StateWitness |
| A frame was re-anchored after software had moved an axis past its soft limit, baking a large error into the next run. | A frame has an identity and a history; events, not reconnects, decide its trust. | FrameProof |
| A Z brake had to be protected from the very software that commanded it: stop, prove stopped, then engage. | The safe order must be a primitive, not a convention. | Safe brake intelligence |
| The machine ran warm while idle and the power-supply fan kept cycling. | Holding strategy must follow load, brake state and orientation. | QuietHold |
| Many interface panels wanted controller status at once and queued behind one library lock. | One acquisition, many subscribers. | EventStream |
Frequently asked
Frequently asked questions
What is transactional motion control?
What does a motion controller do when it loses communication mid-move?
What does VERIFIED mean for a position?
Which events invalidate a coordinate frame?
Why is braking a vertical axis a special sequence?
References
Sources
- Database transaction — Wikipedia (the transaction model MotionProof borrows for motion)
- Data lineage — Wikipedia (provenance, applied here to machine state)
- EtherCAT Technology Group — EtherCAT technology
- Kollmorgen — SafeMotion functions including Safe Brake Control
- PleniDose — Panacea Bio Chem
Current literature
Trending in the field
Recent literature where precise, provable machine motion meets biology and the laboratory — retrieved from PubMed, 17 September 2026.
- Automated Biomedical Research Laboratories: Development, Current State, and a Roadmap for Adaptation into Shared Research Resources — J Biomol Tech, 2026
- Automated implementation of the SwabSeq COVID-19 diagnostic assay on the opentrons flex liquid-handling robot — Diagn Microbiol Infect Dis, 2026 Nov
- Cell-free systems as complementary Test layers for protein engineering in biofoundry workflows — Curr Opin Biotechnol, 2026 Sep 15
- Automated carousel-based electrochemical sensing toward microbiological and oncological settings — Anal Chim Acta, 2026 Sep 22
- Rapid and robust laser-frequency auto-locking using Bayesian-optimization and discrete-wavelet-transformation algorithms — Rev Sci Instrum, 2026 Sep 1
This week in the field
11 – 17 September 2026
Newest PubMed records in robotic manipulation, automated laboratories and precision control — refreshed weekly. Listing only; inclusion is not endorsement.
- A versatile dexterous robotic wrist: lightweight, modular, and pose-adaptive — Bioinspir Biomim, 2026 Sep 15
- Cell-free systems as complementary Test layers for protein engineering in biofoundry workflows — Curr Opin Biotechnol, 2026 Sep 15
- Bibliometric and knowledge-map analysis of research on robot-assisted vascular interventional surgery (2015-2025) — J Robot Surg, 2026 Sep 12
MOTION INTEGRITY ARCHITECTURE























