UR Reality
Universal Robots ships six current-generation arms. The UR3e (3 kg payload, 500 mm reach) targets benchtop and laboratory work. The UR5e (5 kg, 850 mm) and UR10e (12.5 kg, 1300 mm) cover the bulk of small-and-medium-manufacturing pick-and-place, machine-tending, and assembly applications. The UR16e (16 kg, 900 mm) extends payload for screwdriving, palletizing of heavier packages, and end-of-arm tooling that itself weighs several kilograms. The UR20 (20 kg, 1750 mm) and UR30 (30 kg, 1300 mm), the most recent additions, push the platform into pallet-scale logistics and heavier assembly tasks that were previously the province of fenced industrial robots.
All six arms run PolyScope, the UR teach-pendant runtime, which exposes a single programming surface across the line and supports URCap extensions from a sizable third-party integrator ecosystem covering grippers, vision systems, force/torque sensors, conveyors, and application-specific software (Robotiq, OnRobot, Schmalz, MiR for mobile-base composition, and several hundred others). Teradyne has owned the company since 2015, and the strategic context is acquisition-portfolio coordination with Mobile Industrial Robots (MiR) and the broader Teradyne automation thesis.
Safety today is governed by ISO 10218-1/-2 (industrial robot safety) and ISO/TS 15066 (collaborative-robot specification, including biomechanical force and pressure limits). UR arms are designed for PFL (power-and-force-limited) operation per 15066, and integrators routinely deploy them in collaborative cells without perimeter fencing. The architectural reality, however, is that the safety mode is established at integration time by a risk assessment specific to that cell, that tooling, and that workpiece. Changing the task, different payload, different end-effector, different human proximity, requires re-running the risk assessment and frequently re-validating the cell. Mode is a deployment-time property, not a runtime property.
Architectural Fit
Governed actuation, as disclosed in the provisional, places a graduated-actuation mode selector and a composite admissibility evaluator ahead of each actuator. The evaluator scores a proposed actuation over credentialed observations, an authority taxonomy, freshness, and policy, and emits one of admit, gate, defer, solicit, reject, or escalate. The mode selector then maps that determination onto a governance-policy-defined set of actuation modes for the actuator class, ranging in the disclosure from disabled, simulated, advisory, and consultative through constrained, stage-gated, and full, with an emergency-accelerated mode reserved for authorized preemption. This is a continuous, bounded mapping rather than a binary permit-or-deny, so a falling admissibility score degrades the arm toward less autonomous modes instead of forcing a stop-or-run choice.
Applied to a cobot, the arm's own ISO/TS 15066 collaborative safety envelopes, hand-guiding, power-and-force-limited operation, and speed-and-separation-monitored operation, become the actuator-class modes the selector governs, rather than settings fixed once at integration. A proposed motion is admitted only if the credentialed evidence of human proximity, workpiece identity, and tooling state supports the mode it would run in; if the evidence is stale or thin, the evaluator can defer, or solicit additional observations, before committing. A commitment-point evaluator prefers reversible actuation paths where feasible, and every evaluation, mode selection, and preemption is written to a lineage field, so the reason the arm entered a given mode is reconstructable after the fact. The disclosure does not itself define the 15066 force and pressure limits; it defines the governance layer that decides, per actuation, which of the cobot's compliant modes is admissible.
For a UR arm, the substrate composes naturally with PolyScope's existing safety configuration. Today PolyScope supports safety planes, joint limits, tool-flange force limits, and reduced-mode triggers; governed actuation reframes these as the constituents of declared mode-commitments rather than as standalone configuration. A URCap that surfaces governed-actuation commitments would let an integrator declare, for a given cell, the set of modes the arm is permitted to enter, the conditions under which each mode is admissible, the supervisory inputs that gate transitions, and the audit record that each transition produces. Risk assessment becomes a declaration the cell carries, not a binder the integrator files.
Regulatory direction is consistent with declared, auditable, mode-bound operation: the EU AI Act sets obligations for high-risk systems, and the revised ISO 10218 tightens the integration-and-safety boundary. Those are external market conditions, not claims of the filing. Governed actuation supplies an architectural shape aligned with them, and PolyScope is a runtime where it could be expressed without rebuilding the platform.
UR-Specific Fit
Universal Robots is positioned to absorb governed actuation more easily than any of the larger industrial-robot incumbents. Three properties converge. First, PolyScope is a single runtime across the entire arm line, so a substrate change propagates uniformly from UR3e to UR30 rather than fragmenting across product families. Second, the URCaps ecosystem is structurally ready for substrate-level extensions: integrators are accustomed to declaring capabilities through URCap manifests, and governed-actuation commitments are a natural extension of that declarative model. Third, the deployed base spans precisely the application classes, laboratory, SME manufacturing, light logistics, where regulators are tightening collaborative-mode requirements fastest and where cell reconfiguration is most frequent.
The Teradyne ownership context matters here. Teradyne's automation thesis is that the bottleneck to industrial-robot adoption in SME and mid-market segments is integration cost, not arm hardware. Governed actuation directly attacks integration cost by making cell reconfiguration a declarative change against a known substrate rather than a re-engineering exercise. The MiR mobile-robot composition is a second axis: a UR arm on a MiR base operates across multiple physical zones, each with potentially different supervisory regimes, and the governed-actuation primitive is what lets the combined system commit to an admissible mode at each zone without per-zone manual reconfiguration.
Competitively, UR's larger industrial peers, FANUC, ABB, KUKA, Yaskawa, operate on heterogeneous controllers with separate collaborative product lines (FANUC CRX, ABB GoFa/SWIFTI, KUKA LBR iiwa/iisy, Yaskawa HC). A substrate change on those platforms requires per-line work. UR's single-runtime position is the architectural lever, and governed actuation is the substrate that uses it.
UR Position
Universal Robots gains, under governed actuation, an architectural substrate that converts the cobot's defining property, safe operation alongside humans, from a per-cell certification into a runtime commitment carried by the arm itself. The product implications run across the line. UR3e gains a substrate fit for laboratory and pharmaceutical applications where mode discipline is regulatory rather than industrial. UR5e and UR10e gain a deployment substrate that lets a single arm move between collaborative-mode pick-and-place and speed-and-separation high-throughput operation as workpiece flow varies. UR16e gains a screwdriving and palletizing substrate where transition between hand-guided teaching and PFL execution is a settled commitment. UR20 and UR30 gain the substrate that legitimizes their entry into pallet-scale logistics: the larger payload makes mode discipline mandatory, and governed actuation provides it.
For Teradyne, the strategic position is that governed actuation differentiates the UR platform on an axis the larger incumbents cannot match without multi-year controller convergence work. For the URCaps ecosystem, governed actuation creates a new class of integrator value: cells that declare their mode regime are easier to commission, easier to audit, and easier to reconfigure, which is exactly the cost surface integrators sell against.
The architectural commitment is to treat actuation mode as a settled, auditable property of the arm at runtime rather than a deployment-time configuration choice. Universal Robots' single-runtime architecture, declarative URCap ecosystem, and SME-facing deployment posture make it a strong candidate for adoption; ISO 10218, ISO/TS 15066, and the EU AI Act point market and regulatory expectations toward this shape.
Building It
A skilled integrator can reduce this to practice on existing UR hardware without new arm mechanics. The composite admissibility evaluator is a policy-configured function that consumes credentialed observations (human-proximity signals from a safety scanner or vision system, workpiece identity from a fixture or barcode, tooling state from the end-effector), scores them against an authority taxonomy and a freshness bound, and emits admit, gate, defer, solicit, reject, or escalate. The graduated mode selector is a table mapping each admissibility band to a permitted actuation mode drawn from the actuator-class set, and it can be surfaced as a URCap manifest that declares, per cell, the modes the arm may enter and the conditions gating each transition. The evaluator and selector can run on the safety-rated controller, on the PolyScope runtime, on an adjacent industrial PC, or split across them, and the observation channel can be the arm's own safety I/O, a fieldbus, or a mesh link. Embodiments include a single fixed cell, a UR arm on a MiR mobile base crossing zones with different supervisory regimes, and a fleet where per-cell declarations converge to a shared policy. The lineage record can be a local append-only log or a mesh-contributed governed observation. None of these variations depend on a specific vendor, controller, or sensor, and the same governance chain applies whether the actuator is a UR arm or any other governed manipulator.
Closing
Universal Robots already carries the architectural properties, single runtime across the arm line, declarative integrator ecosystem, and a deployed base concentrated in the regulatory-tightening segments of cobot use, that make governed actuation a natural fit. The primitive reframes mode as a runtime commitment rather than a deployment-time configuration: a composite admissibility evaluator gates each proposed actuation, a graduated mode selector maps the result onto the cobot's own compliant modes, a commitment-point evaluator prefers reversible paths, and a lineage field records why each mode was entered. Adopting it preserves UR's existing PolyScope and URCaps surface while giving integrators, end users, and regulators a single substrate to declare against.
Disclosure Scope
The invention described here, Governed Actuation, is disclosed in U.S. Provisional Application No. 64/049,409. The claims of the filing concern the composite admissibility evaluator, the graduated actuation-mode selector, reversibility-aware commitment-point evaluation, preemption budgets, and lineage-recorded actuation provenance, as set out in that application. References to Universal Robots, its UR3e through UR30 arms, PolyScope, URCaps, MiR, Teradyne, and to ISO 10218, ISO/TS 15066, and the EU AI Act are provided as external market and standards context to situate the disclosure, and are not claims of the filing. Universal Robots and the other named products are the property of their respective owners; product descriptions reflect publicly available information and are used for accurate technical comparison only. This article is a dated public disclosure tied to the above provisional application.