Skip to content

Standards

This page lists the binding standards and technical specifications CHAMBER and CONCERTO measure against. It is tier (b) in the three-tier evidence convention described on the literature page: peer-reviewed publications are tier (a) and live in literature.md; industry signal (humanoid factory pilots, logistics fleets, surgical robotics) is tier (c) and lives in adr/international_axis_evidence.md.

Every standard below is identified by its exact edition / part number / release in docs/reference/refs.bib. Public ADRs MUST cite by that key, not by a bare "ISO 10218" or "Release 17" shorthand.

The standards below are grouped into two stacks — machine functional safety and deterministic networking — followed by the axis-to-standard-to-measurable-variable flowchart that ties each ADR-007 heterogeneity axis to its ADR-014 report-table column.

1. Machine functional safety

ISO 10218-1:2025 [iso10218part1_2025] and ISO 10218-2:2025 [iso10218part2_2025] are the binding industrial-robot-safety standards since their 2025 revision. The 2025 edition of ISO 10218-2 absorbs ISO/TS 15066:2016 [isots15066_2016] (the original collaborative-robot technical specification on biomechanical limits and force-pressure tables), making the limits themselves part of a binding standard rather than a technical specification. ISO 13849-1:2023 [iso13849part1_2023] covers the safety-related parts of control systems and assigns Performance Levels (PL a–e) based on risk parameters (severity, frequency, avoidability). IEC 62061:2021 [iec62061_2021] is the parallel functional-safety standard targeted at machinery and its electrical / electronic / programmable-electronic control. The two map onto the broader IEC 61508-1:2010 [iec61508part1_2010] framework, which establishes the Safety Integrity Level (SIL 1–4) hierarchy that ISO 13849-1:2023 PL grades against.

CHAMBER references these standards directly. The Safety axis (SA) in ADR-007 varies per-vendor compliance level — heterogeneous force-limit and SIL/PL pairs across simulated controllers — and the violation columns in ADR-014 Table 2 report against the ISO 10218-2:2025 biomechanical limits absorbed from ISO/TS 15066:2016. The standards do not certify simulation; they define what the simulation must approximate to count as a faithful proxy for compliance-relevant evaluation.

Standard Title Edition / status Cite-key
ISO 10218-1:2025 Robotics — Safety requirements — Part 1: Industrial robots Binding, 2nd ed., 2025 [iso10218part1_2025]
ISO 10218-2:2025 Robotics — Safety requirements — Part 2: Robot applications and integration Binding, 2nd ed., 2025 [iso10218part2_2025]
ISO/TS 15066:2016 Robots and robotic devices — Collaborative robots Absorbed into ISO 10218-2:2025 [isots15066_2016]
ISO 13849-1:2023 Safety of machinery — Safety-related parts of control systems — Part 1 Binding, 4th ed., 2023 [iso13849part1_2023]
IEC 62061:2021 Safety of machinery — Functional safety of safety-related control systems Binding, 2nd ed., 2021 [iec62061_2021]
IEC 61508-1:2010 Functional safety of E/E/PE safety-related systems — Part 1 Binding (SIL framework), 2nd ed., 2010 [iec61508part1_2010]

ISO 13849-1:2023 [iso13849part1_2023]

ISO 13849-1:2023 specifies the design and integration of the safety-related parts of a machine's control system (SRP/CS). It normatively requires that each safety function be assigned a required Performance Level (PL a..e) derived from a risk graph over severity of injury, frequency / duration of exposure, and possibility of avoidance; the realised PL is then verified against the required PL using a categorical architecture (Categories B, 1, 2, 3, 4) that constrains structural redundancy, diagnostic coverage, and resistance to common-cause failures. The standard is the dominant claim path used by collaborative-robot integrators when arguing the safety-rated monitored-stop and power-and-force-limiting functions called out by ISO 10218-2:2025 §5. In CHAMBER, ISO 13849-1:2023 maps onto the SA-axis PL component of the per-vendor SIL/PL pair: a simulated controller's PL rating is one half of the vendor-compliance label that populates the ADR-014 Table 2 per-condition row when ADR-007 Open Q #4 resolves toward axis decomposition.

IEC 62061:2021 [iec62061_2021]

IEC 62061:2021 is the machinery-sector application of the IEC 61508 functional-safety framework to safety-related electrical, electronic, and programmable-electronic (E/E/PE) control systems. It normatively requires that each safety function be assigned a required Safety Integrity Level (SIL 1..3 for machinery — SIL 4 is out of scope for this standard), with the SIL derived from a tolerable risk argument over severity, frequency, probability of occurrence, and probability of avoidance; the realised SIL is then verified against systematic capability requirements and a quantitative target failure measure (PFH for high-demand mode). IEC 62061 is the parallel claim path to ISO 13849-1: an integrator may certify a safety function under either or both, and the two are explicitly cross-mappable. In CHAMBER, IEC 62061:2021 maps onto the SA-axis SIL component of the per-vendor SIL/PL pair and is the source of the SIL labels reported alongside PL grades in the ADR-014 Table 2 vendor-compliance rows.

IEC 61508-1:2010 [iec61508part1_2010]

IEC 61508-1:2010 is the cross-sector functional-safety standard for E/E/PE safety-related systems, from which both ISO 13849-1's PL framework and IEC 62061's machinery-SIL framework are derived. It normatively defines the Safety Integrity Level (SIL 1..4) hierarchy in terms of probability of dangerous failure per hour (PFH) and average probability of failure on demand (PFD), and it fixes the overall safety-lifecycle process: hazard and risk analysis, safety requirements specification, design, integration, validation, operation, and modification. The standard does not bind a specific industry, but its SIL semantics are the conceptual anchor that gives both PL grades and machinery SIL grades a common interpretation. In CHAMBER, IEC 61508-1:2010 is the upstream reference for the per-vendor SIL/PL pair reported by the SA axis and for the assumption-row label "ISO 10218-2:2025 SIL/PL precondition satisfied" that ADR-014 Table 1 adds (as A4) if the safety axis decomposes.

2. Deterministic networking and 5G-TSN

CHAMBER's fixed-format communication channel (chamber.comm) is anchored to the IEEE Time-Sensitive Networking (TSN) family and the 3GPP Release 17 URLLC profile. IEEE Std 802.1AS-2020 [ieee8021as_2020] specifies generalised-precision time-synchronisation (the profile of IEEE 1588 PTPv2 used in TSN); IEEE Std 802.1Qbv-2015 [ieee8021qbv_2015] defines scheduled traffic via time-aware shaping; IEEE Std 802.1CB-2017 [ieee8021cb_2017] defines frame replication and elimination for reliability (FRER) — used to obtain seamless redundancy across redundant network paths. 3GPP Release 17 defines URLLC and the 5G system architecture that integrates a 5G network as a virtual TSN bridge; 3GPP TS 23.501 v17 (Release 17) [threegpp_ts23501_r17] is the canonical 5G-system-architecture spec and the entry point for the 5G-TSN integration model; 3GPP TS 24.535 v17 [threegpp_ts24535] and 3GPP TS 24.519 v17 [threegpp_ts24519] specify the DS-TT/NW-TT translator protocol aspects. 5G-ACIA (5G Alliance for Connected Industries and Automation) publishes the industry-side 5G-TSN integration white paper (2019) [fiveg_acia_2019_5gtsn] that translates 3GPP and IEEE primitives into deployable factory configurations.

The six pre-registered URLLC profiles in chamber.comm.URLLC_3GPP_R17 (ideal, urllc, factory, wifi, lossy, saturation) are parameterised against the numeric envelopes these standards define: URLLC's 1 ms latency at 99.9999% reliability target, 802.1Qbv's microsecond-grade scheduled-traffic jitter, and the factory-jitter measurements reported in the 5G-TSN industrial trials (see the tier-c evidence sweep linked above).

Standard / spec Title / scope Role in CHAMBER Cite-key
IEEE Std 802.1AS-2020 Timing and Synchronization for Time-Sensitive Applications Per-tick clock alignment in chamber.comm [ieee8021as_2020]
IEEE Std 802.1Qbv-2015 Bridges and Bridged Networks — Amendment 25: Enhancements for Scheduled Traffic Anchors jitter bounds in URLLC profiles [ieee8021qbv_2015]
IEEE Std 802.1CB-2017 Frame Replication and Elimination for Reliability (FRER) Reference for the redundancy variant of the degradation wrapper [ieee8021cb_2017]
3GPP TS 23.501 v17 (Rel. 17) System architecture for the 5G System (5GS); Stage 2; URLLC + 5G-as-virtual-TSN-bridge model Integration model for the comm stack [threegpp_ts23501_r17]
3GPP TS 24.535 v17 (Rel. 17) DS-TT to NW-TT protocol aspects TSN translator protocol reference [threegpp_ts24535]
3GPP TS 24.519 v17 (Rel. 17) TSN AF to DS-TT/NW-TT protocol aspects TSN translator protocol reference [threegpp_ts24519]
5G-ACIA 5G-TSN white paper (2019) Integration of 5G with Time-Sensitive Networking for Industrial Communications Cross-check on factory-floor parameterisation [fiveg_acia_2019_5gtsn]

IEEE Std 802.1AS-2020 [ieee8021as_2020]

IEEE Std 802.1AS-2020 specifies the generalised precision time protocol (gPTP), the profile of IEEE 1588 PTPv2 used inside time-sensitive networks. A conformant bridge propagates timing information end-to-end through a master-slave hierarchy elected by the best-master-clock algorithm (BMCA), and the standard normatively requires that path delay and residence time at each hop be measured and corrected so that participating end-stations share a common time reference accurate to sub-microsecond on a single-segment factory deployment. CHAMBER's fixed-format channel reads AoI timestamps against this assumed shared time-base: without an IEEE Std 802.1AS-2020-grade primitive, AoI is not comparable across agents, and the conformal prediction layer's input is no longer well-defined.

IEEE Std 802.1Qbv-2015 [ieee8021qbv_2015]

IEEE Std 802.1Qbv-2015 defines time-aware scheduled traffic via gate control lists (GCLs) at every bridge output port. Each traffic class is assigned an open transmission window inside a repeating cycle, and the bridge normatively guarantees that frames belonging to a scheduled class are forwarded only during their window, free of contention from best-effort traffic. The standard's guarantees apply to per-class p99 latency and to microsecond-grade jitter under deterministic load. CHAMBER's URLLC degradation profiles (chamber.comm.URLLC_3GPP_R17) treat the IEEE Std 802.1Qbv-2015 jitter envelope as the floor of what a real 5G-TSN deployment can sustain — the urllc and factory profiles parameterise jitter against this floor; the wifi and saturation profiles deliberately violate it to model commodity wireless and overloaded fabrics.

IEEE Std 802.1CB-2017 [ieee8021cb_2017]

IEEE Std 802.1CB-2017 defines frame replication and elimination for reliability (FRER). Each protected stream is sequence-numbered at ingress, replicated onto multiple disjoint paths through the network, and de-duplicated at the egress so that a single-path failure causes no frame loss as observed by the receiver. The standard normatively requires per-stream sequence-recovery state, a bounded sequence-recovery window, and consistent stream identification across replication points. CHAMBER references IEEE Std 802.1CB-2017 as the formal template for a redundancy variant of the comm-degradation wrapper: the wrapper can compose any saturation or drop profile with a configurable FRER overlay that masks single-path losses up to a bound, letting the Stage 2 CM spike isolate whether observed violations are intrinsic to the stream or recoverable by redundancy.

3GPP TS 23.501 v17 §5.27 [threegpp_ts23501_r17]

3GPP TS 23.501 v17 (Release 17) is the canonical 5G-system-architecture specification; §5.27 covers the integration of the 5G system with deterministic networking. The 5G system is exposed to the TSN domain as a single virtual TSN bridge fronted by two translator functions: a device-side TT (DS-TT) co-located with the UE and a network-side TT (NW-TT) co-located with the UPF. The clause normatively requires that QoS, scheduling, and time-synchronisation primitives be mapped transparently between TSN streams and 5G QoS flows, so that an end-to-end deterministic stream can cross the 5G radio segment without losing its TSN guarantees. In CHAMBER, the URLLC profiles represent the over-the-air segment within this end-to-end model; the AoI timestamp seen at the receiver is the composition of the DS-TT-side enqueue time, the radio segment, and the NW-TT-side release.

5G-ACIA 5G-TSN integration white paper (2019) [fiveg_acia_2019_5gtsn]

The 5G Alliance for Connected Industries and Automation (5G-ACIA) publishes the industry-side white papers that translate the 3GPP TS 23.501 v17 §5.27 integration model and the IEEE 802.1 TSN family into deployable factory-floor configurations. The 2019 white paper "Integration of 5G with Time-Sensitive Networking for Industrial Communications" is the load-bearing reference for the comm stack's factory parameterisation. The white papers specify deployment patterns (single-DS-TT, multi-DS-TT, redundancy combinations), parameter ranges measured from industrial trials, and reference architectures for time-sensitive control loops over 5G. They are not binding standards in the IEEE / 3GPP sense, but they are the de facto industry reference for what numeric envelopes are actually achievable on the factory floor. CHAMBER's six URLLC profiles cross-check their numeric envelopes against the 5G-ACIA factory parameterisation; deviations are documented in the profile-level comments in chamber.comm.URLLC_3GPP_R17.

3. Standards-and-measurement stack diagram

The flowchart below ties each ADR-007 heterogeneity axis to its governing standard (where one exists), to the measurable benchmark variable CHAMBER exposes, and to the ADR-014 report-table column it populates. Where no binding standard applies, the column points to the project's own anchoring document (ADR-007 implementation staging or ADR-009 partner-zoo construction).

flowchart LR
    classDef axis    fill:#eef,stroke:#446,stroke-width:1px;
    classDef std     fill:#efe,stroke:#464,stroke-width:1px;
    classDef metric  fill:#fef,stroke:#646,stroke-width:1px;
    classDef report  fill:#ffe,stroke:#664,stroke-width:1px;

    AS[Action space - AS]:::axis
    OM[Observation modality - OM]:::axis
    CR[Control rate - CR]:::axis
    CM[Communication - CM]:::axis
    PF[Partner familiarity - PF]:::axis
    SA[Safety - SA]:::axis

    AS_S[ADR-007 §Stage 1<br/>no binding standard]:::std
    OM_S[ADR-007 §Stage 1<br/>no binding standard]:::std
    CR_S[ADR-007 Open Q #3<br/>chunk-async; no binding standard]:::std
    CM_S[IEEE 802.1AS + 802.1Qbv + 802.1CB<br/>+ 3GPP R17 URLLC + TS 23.501 §5.27]:::std
    PF_S[ADR-009 §partner-zoo<br/>no binding standard]:::std
    SA_S[ISO 10218-2:2025<br/>+ ISO 13849-1:2023<br/>+ IEC 62061:2021<br/>+ IEC 61508-1:2010]:::std

    AS_M[action-space dimensionality;<br/>embodiment-class pair]:::metric
    OM_M[sensor-suite vector;<br/>vision / +F-T / +tactile fusion]:::metric
    CR_M[control-rate ratio;<br/>chunk size]:::metric
    CM_M[latency p99; AoI;<br/>drop rate p99]:::metric
    PF_M[trained-with vs frozen-novel;<br/>partner-swap transient]:::metric
    SA_M[force/torque peak per joint;<br/>SIL / PL pair;<br/>contact-pressure ISO/TS 15066 table]:::metric

    AS_R[Table 2 per-condition row]:::report
    OM_R[Table 2 per-condition row]:::report
    CR_R[Table 2 per-condition row]:::report
    CM_R[Table 2 per-condition row]:::report
    PF_R[Table 2 per-condition row<br/>+ λ re-init transient]:::report
    SA_R[ADR-014 Table 2 per-condition row<br/>+ Table 3 conservativeness-gap row]:::report

    AS --> AS_S --> AS_M --> AS_R
    OM --> OM_S --> OM_M --> OM_R
    CR --> CR_S --> CR_M --> CR_R
    CM --> CM_S --> CM_M --> CM_R
    PF --> PF_S --> PF_M --> PF_R
    SA --> SA_S --> SA_M --> SA_R

Reading the diagram column by column:

  • Column (i): heterogeneity axis — the six axes locked at ADR-007 revision 3.
  • Column (ii): governing standard or protocol — the binding reference, where one exists. AS, OM, CR, and PF have no binding standard yet; their reference is the ADR section that pins the spike protocol. CM is fully covered by IEEE TSN + 3GPP. SA is covered by ISO 10218-2:2025 (which absorbs ISO/TS 15066:2016), ISO 13849-1:2023, IEC 62061:2021, and IEC 61508-1:2010 — the three machine-functional-safety standards underneath ISO 10218-2 that fix the PL and SIL semantics of the per-vendor compliance pair.
  • Column (iii): measurable benchmark variable — what CHAMBER records and reports. These are the variables the spike pre-registration YAMLs commit to before launch.
  • Column (iv): report-table column — the ADR-014 three-table format. Table 1 is per-assumption violation rates; Table 2 is per-condition (predictor × conformal mode) rates; Table 3 is conservativeness gap vs. oracle CBF. The SA row populates the Table 2 per-condition row and the Table 3 conservativeness-gap row; if the Stage 3 SA spike confirms the safety-axis decomposition deferred to ADR-007 Open Question #4, Table 1 additionally gains an A4 row labelled "ISO 10218-2:2025 SIL/PL precondition satisfied" whose SIL/PL semantics are fixed by IEC 61508-1:2010 and claimed under ISO 13849-1:2023 / IEC 62061:2021.