For Certification & Compliance Reviewers

AUTHREX Standards Mapping

The author's mapping of AUTHREX governance components to standards and frameworks used in defense and aviation review: DO-178C, DO-333 Formal Methods Supplement, MIL-STD-882E System Safety, MIL-HDBK-516C Airworthiness, ISO 26262:2018, and NIST AI RMF. These mappings are the author's own interpretation and are not certification evidence; no certification authority has applied or assessed any of these standards for this work.

DO-178C: Software Considerations in Airborne Systems & Equipment Certification

RTCA / EUROCAE ED-12C · Software levels A through E, assigned by the system safety assessment · Recognized by FAA Advisory Circular (AC) 20-115D as an acceptable means, but not the only means, of compliance for airborne software

DO-178C provides guidance for software in airborne systems. FAA AC 20-115D recognizes it as an acceptable means, but not the only means, of compliance for airborne software; the software level for a system comes from that system's safety assessment. In the author's interpretation, authority-governed autonomy software for UAVs and crewed aircraft would fall within its scope. No software level has been assigned to any AUTHREX component, and no DO-178C objective has been satisfied or assessed.

AUTHREX ComponentDO-178C SectionMapping (author interpretation)
HMAA Authority State Machine §6.3.3 Reviews and Analyses of Software Architecture Four-level authority state machine (A0-A3) with transitions specified in TLA+; its architecture could be the subject of the reviews and analyses in this section.
SATA Sensor Trust Evaluation §6.3.1 Reviews and Analyses of High-Level Requirements; Annex A Table A-3, Verification of Outputs of Software Requirements Process Continuous trust scalar with cryptographic attestation, as specified in the SATA design; its sensor-input integrity requirements could be reviewed and traced under this section.
CARA Recovery Phases (GREP) §6.4.2.2 Robustness Test Cases Deterministic phased recovery (Guard, Reduce, Evaluate, Promote) defines off-nominal sensor and authority conditions from which robustness test cases could be derived.
FLAME Deliberation Window §6.3.4 Reviews and Analyses of Source Code (its accuracy and consistency item covers worst-case execution timing) Latency-injection layer for human-in-the-loop confirmation uses a bounded delay, which a worst-case execution timing analysis of the software would need to include.
MAIVA Byzantine Consensus §2.4.2 Multiple-Version Dissimilar Software; §12.3.2 Considerations for Multiple-Version Dissimilar Software Verification Multi-agent consensus with the Byzantine bound as its design basis is designed to provide redundancy related to multiple-version dissimilar software strategies (formal guarantees are not asserted while the MAIVA CriticalSafe invariant remains open).

DO-333: Formal Methods Supplement to DO-178C

RTCA / EUROCAE ED-216 · Formal analysis may replace some, not all, reviews, analyses and tests · Used together with DO-178C, not as a standalone document (FAA AC 20-115D, paragraph 8.a)

DO-333 is the formal methods supplement to DO-178C. It describes how formal analysis, such as model checking and theorem proving, may be used in place of some reviews, analyses and tests within a DO-178C process. The AUTHREX TLA+ work consists of bounded TLC model checks, not proofs. The DO-333 identifiers below have not been checked against a copy of DO-333.

AUTHREX ComponentDO-333 Identifier (unverified)Mapping (author interpretation)
HMAA TLA+ Specification FM.6.3.1: Formal Analysis (Property Verification); identifier unresolved (DO-333 not inspected) A TLC model check of the HMAA discrete authority model (HMAA_Authority_FSM, archived at Zenodo 10.5281/zenodo.21300977) at small constants (MaxTick = 8, MaxDwell = 8, OscMaxTransitions = 3, OscLockoutTicks = 4) explored 23,748 distinct states to depth 9 (26,397,356 states generated) with no counterexample. Of the eight properties checked, four are exercised (TypeInvariant, LevelBounded, DwellNonNegative, CompoundFailureReachesA0) and four hold only vacuously at these bounds because no upgrade and no lockout is reachable (UpgradeIsStepwise, TrustAuthorityWeakConsistency, LockoutConsistent, OscillationLockoutTerminates). The run at operational constants did not complete. This is a bounded check of the model, not a proof for other bounds, for the other architectures, or for the code. The model assumes that instantaneous authority equals its target, so continuous behavior between decision instants is not covered, and the archived spec numbers authority levels inversely (level 0 is full authority, level 3 is lowest). Any DO-333 credit would depend on an applicant's plan, tool qualification and acceptance by the certification authority; FAA AC 20-115D states that supplements are not used as standalone documents.
Authority Transition Design Properties FM.6.3.6: Property-Based Decomposition; identifier unresolved (DO-333 not inspected) Authority-transition design properties (monotonic downgrade, bounded upgrade delay, hysteresis) are stated as requirements, not as checked results: in the bounded HMAA check, the upgrade and lockout properties hold only vacuously because no upgrade and no lockout is reachable at those bounds. Their use as decomposable formal requirements is a proposal.
Dempster-Shafer Trust Fusion FM.6.3.4: Soundness of Formal Method; identifier unresolved (DO-333 not inspected) Dempster-Shafer evidence combination is a mathematically defined fusion rule, not a formal analysis method; soundness in the DO-333 sense concerns the formal analysis method and tool used, so the relevance of this row is unresolved. The SATA TLA+ work is bounded model checking, not proof.
Reachability Checks (TLC Output) FM.6.7: Verification of Verification; identifier unresolved (DO-333 not inspected) TLC model checker output (states explored, depth, and no counterexample at the stated bounds) is an auditable record of a bounded check, not a proof; whether it could support any DO-333 verification-of-verification objective is unresolved.

MIL-STD-882E: Standard Practice for System Safety

DoD-wide system safety standard practice (MIL-STD-882E w/Change 1, September 27, 2023) · DoD system-safety standard; applicable requirements and tasks depend on governing policy and solicitation or contract tailoring · Risk classification matrix used by all military services

MIL-STD-882E governs system safety analysis across the DoD. Its hazard analysis tasks, severity and probability categories, and risk assessment matrix define how hazards, including those of autonomous systems, are assessed and accepted; failure mode, effects and criticality analysis (FMECA) and fault tree analysis appear in it only as example analytical techniques, not as tasks. In the author's interpretation, AUTHREX authority reduction is a candidate hazard mitigation (an engineered feature under the design order of precedence in section 4.3.4) whose effect on risk would need system-specific hazard analysis.

AUTHREX ComponentMIL-STD-882E Task or SectionMapping (author interpretation)
SATA Sensor Trust Task 204: Subsystem Hazard Analysis; Task 205: System Hazard Analysis Per-sensor failure modes (jamming, spoofing, drift) are modeled in the SATA simulation with detection coverage and authority response, and could be an input to subsystem and system hazard analyses. The simulation's 10-mode FMEA table follows MIL-STD-1629A, which was cancelled by Notice 3 on August 4, 1998.
HMAA Authority Computation Section 4.3.4: Identify and document risk mitigation measures (design order of precedence) Authority reduction is a candidate hazard mitigation, an engineered feature under the design order of precedence; whether it reduces the risk of a given hazard would need system-specific hazard analysis. Authority levels are not MIL-STD-882E severity categories.
CARA Recovery Protocol Task 205: System Hazard Analysis CARA's deterministic recovery transitions are designed to move the system to a safe state on trust collapse; any reduction in mishap probability would have to be shown by system-specific hazard analysis and has not been measured.
ERAM Escalation Risk Model Task 209: System-of-Systems Hazard Analysis ERAM computes model-based escalation-risk metrics for command-and-control environments operating under degraded conditions; these could inform the identification of hazards unique to a system of systems.

MIL-HDBK-516C: Airworthiness Certification Criteria

DoD airworthiness handbook, December 12, 2014 · Marked "THIS HANDBOOK IS FOR GUIDANCE ONLY. DO NOT CITE THIS DOCUMENT AS A REQUIREMENT." · Used by service airworthiness authorities to define an air system's certification basis

MIL-HDBK-516C sets out airworthiness certification criteria, standards and methods of compliance for manned and unmanned military air systems; it is guidance, not a requirement. In the author's interpretation, the AUTHREX authority-governance design is relevant to criteria in Section 15 (Computer Systems and Software) and Section 9 (Crew Systems, including 9.4 Human performance) for autonomous and optionally-piloted military aircraft. No airworthiness assessment has been made.

AUTHREX ComponentMIL-HDBK-516C CriterionMapping (author interpretation)
HMAA Authority Layer §15.2.1: Functional coupling Authority gating is designed to limit which flight-safety-critical actions an autonomy function can command; hardware-anchored gating appears only in reference designs that have not been built. This limit bears on the §15.2.1 criterion on avoiding unnecessary coupling between safety-critical elements.
FLAME Deliberation Window §9.4: Human performance (Section 9, Crew systems) Bounded latency injection is designed to give the operator a minimum decision time before an irreversible action, which is relevant to the §9.4 criteria on operator tasks and workload for optionally-piloted aircraft.
SATA Trust Attestation §15.2.4: Critical discrepancies Cryptographically attested sensor trust, as specified in the SATA design, is intended to flag possible tampering with, or faults in, flight-control inputs; such flags could feed the discrepancy reporting that §15.2.4 covers. They are not V&V evidence and cannot show that inputs were not tampered with.

ISO 26262 (2018 edition): Road Vehicles, Functional Safety

ISO 26262:2018, the international standard for road vehicle functional safety · ASIL D is the highest automotive safety integrity level, assigned to safety goals through hazard analysis and risk assessment for a specific item

ISO 26262:2018 sets out the functional safety lifecycle for road vehicles; an ASIL is determined for each hazardous event of a specific item through hazard analysis and risk assessment, not assigned to a design in advance. BLADE-AV, the automotive AUTHREX reference design (design package and bill of materials; not built), has not been assessed against ISO 26262, and no ASIL has been determined for it.

AUTHREX / BLADE-AV ComponentISO 26262:2018 Part / ClauseMapping (author interpretation)
HMAA Authority State Machine Part 3 §6: Hazard Analysis and Risk Assessment (safety goals, §6.4.4) Authority levels are not ASILs or HARA categories. Once a HARA for a specific vehicle function has determined its safety goals, authority reduction could be considered as a candidate safety measure; no HARA or ASIL decomposition has been performed.
SATA Sensor Trust + MAIVA Consensus Part 6 §7: Software Architectural Design (Redundancy) Heterogeneous sensor reasoning with consensus, with the Byzantine bound as its design basis, is designed to provide redundancy that could be considered in software architectural design; no ASIL has been determined, and formal guarantees are not asserted while the MAIVA CriticalSafe invariant remains open.
CARA Safe-State Transitions Part 4 §6: Technical Safety Concept (safety mechanisms, §6.4.2) Deterministic safe-state transitions on trust collapse are a candidate safety mechanism that a technical safety concept could specify; they do not show safety goal compliance, and no item or safety goal has been defined.
TLA+ Bounded Model Checking Part 6 §9: Software Unit Verification Formal verification is listed in ISO 26262 Part 6 §9 as a software unit verification method. The AUTHREX TLA+ work checks design-level models, not software units (HMAA: 23,748 distinct states, depth 9, no counterexample; 4 of 8 properties exercised, 4 vacuous at these bounds), so it is not unit verification evidence.

NIST AI Risk Management Framework (AI RMF 1.0)

NIST AI 100-1 (AI RMF 1.0, January 2023), a voluntary framework · Defines GOVERN, MAP, MEASURE, MANAGE functions

NIST AI RMF defines four functions for AI risk management. In the author's interpretation, AUTHREX mechanisms may support each function for autonomous-systems use cases; GOVERN also covers organizational policies, processes and culture, which a technical mechanism does not address.

AI RMF FunctionAUTHREX MechanismMapping (author interpretation)
GOVERN HMAA authority computation + audit log Authority decisions with a cryptographic audit trail, as specified in the design, may support governance evidence at runtime as well as at design time; hardware anchoring is specified only in reference designs that have not been built.
MAP Domain-specific scenario libraries (drone, AV, ship, UGV, ICS) Operational scenarios mapped per domain in the BLADE reference designs (design packages and bills of materials; not built) and their browser simulations, with failure modes cataloged with detection coverage, may support MAP.
MEASURE SATA sensor trust scalars + ADARA adversarial probability Continuous quantitative measures of sensor trust and adversarial-deception probability, computed in simulation, may support MEASURE as AI risk metrics.
MANAGE FLAME deliberation + CARA recovery Latency injection (human review windows) and deterministic recovery protocols on trust degradation may support active risk management under MANAGE.

DoD Directive 3000.09: Autonomy in Weapon Systems

Effective January 25, 2023 · Requires that autonomous and semi-autonomous weapon systems be designed to allow commanders and operators to exercise "appropriate levels of human judgment over the use of force" · Senior review applies to autonomous weapon systems (para 1.2.c); semi-autonomous systems with no autonomous mode are exempt (para 1.2.d(1))

DoDD 3000.09 requires that autonomous and semi-autonomous weapon systems be designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force. The directive establishes policy but does not prescribe technical mechanisms for meeting it. In the author's interpretation, AUTHREX mechanisms could contribute design features toward that requirement for a specific weapon system; this has not been reviewed under the directive.

DoDD 3000.09 RequirementAUTHREX MechanismMapping (author interpretation)
Appropriate human judgment FLAME deliberation window + HMAA authority gating Authority levels A0-A3 are designed to narrow autonomous authority as sensor trust falls, and FLAME is designed to insert a bounded delay before an irreversible action so that an operator can review it; these could contribute to a design that allows appropriate levels of human judgment over the use of force.
Reliable performance under realistic conditions SATA + MAIVA + ADARA pipeline Heterogeneous sensor reasoning with Byzantine consensus and adversarial-deception detection is intended to contribute reliability evidence under contested-domain conditions; all results to date are from simulation, and the MAIVA CriticalSafe invariant remains open.
Adequate operator training & HSI Authority-level transparency + audit trail The design shows operators the current authority level and the reason for any change (demonstrated in simulation), and a cryptographic audit trail specified in the design is intended to support post-event review.
Failure prevention CARA deterministic recovery + bounded model check of the HMAA state machine A TLC check of the HMAA authority model at small constants (23,748 distinct states, depth 9, no counterexample; 4 of 8 properties exercised, 4 vacuous at these bounds) is a bounded check of the model, not of the physical system or the code. CARA is designed to apply deterministic recovery on trust collapse.
Note on scope: DoDD 3000.09 mappings above describe how AUTHREX could serve as a technical compliance mechanism for the directive's policy requirements. AUTHREX has not been reviewed or endorsed by the DoDD 3000.09 Senior Review process. Compliance determination is the responsibility of program offices and the Senior Review.

NERC CIP: Critical Infrastructure Protection

North American Electric Reliability Corporation · Mandatory standards for bulk electric system cyber security · CIP-002 through CIP-015

BLADE-INFRA, the AUTHREX reference design for industrial control systems (design package and bill of materials; not built), targets NERC CIP-regulated environments where AI integration into ICS/SCADA must meet strict cyber security and reliability standards. No NERC CIP compliance has been assessed.

AUTHREX / BLADE-INFRA ComponentNERC CIP StandardMapping (author interpretation)
SATA Hardware Trust Anchor CIP-013: Supply Chain Risk Management TPM-anchored trust attestation, specified in the SATA design and the BLADE-INFRA reference design, is intended to provide device-integrity records that could support CIP-013 vendor and equipment provenance controls.
SATA Continuous Sensor Monitoring CIP-007: System Security Management Continuous integrity monitoring of cyber-asset sensor inputs could support some CIP-007 controls (security event monitoring, system status monitoring).
HMAA Authority Audit CIP-008: Incident Reporting & Response Planning A cryptographic audit trail, as specified in the design, could supply records for incident response under CIP-008.
ADARA Adversarial Detection CIP-005: Electronic Security Perimeters Continuous adversarial-deception probability is intended as a monitoring signal for AI-integrated control loops, which could complement electronic security perimeter monitoring.

NASA SBIR 2026: Autonomous Systems Subtopic Alignment

NASA Small Business Innovation Research Program · FY2026 Solicitation · INSITU.1.S26B and EXPAND.3.S26B

NASA's FY2026 SBIR solicitation identifies autonomous systems fault management and distributed spacecraft health management as funded research priorities. The AUTHREX governance architecture maps to both subtopics, with simulation framework at TRL 3 matching the published Phase I TRL expectation. EXPAND.3.S26B alignment is further illustrated by the BLADE-SPACE Governance Node reference design (design package and bill of materials; not built): a 91-component, 6U+ SmallSat payload module Preliminary Design Phase data package targeting LEO 400-1200 km, 30 krad TID and a 5-year mission life (design targets; not tested, certified or assessed), specifying hot-redundant RTG4 FPGA + Aitech SBC compute and an ECDSA-anchored audit chain. Subtopic technical specifications and Q&A are public records at the NASA SBIR/STTR 2026 Appendix A and B Q&A Hub.

AUTHREX ComponentNASA SubtopicMapping (author interpretation)
SATA Sensor Fault Detection INSITU.1.S26B Continuous trust scoring via Dempster-Shafer evidence fusion provides real-time sensor fault detection for autonomous-systems fault-management requirements.
HMAA Authority Degradation INSITU.1.S26B Four-level authority state machine (A0-A3) degrades autonomous authority proportionally in response to detected faults, satisfying the subtopic's fault-response requirements.
CARA Deterministic Recovery INSITU.1.S26B Governed Recovery Phases (Guard, Reduce, Evaluate, Promote) provide deterministic post-fault recovery sequences for autonomous-systems fault-management.
MAIVA Byzantine Consensus EXPAND.3.S26B Multi-agent consensus with the Byzantine bound as its design basis is intended to support distributed-spacecraft swarm continuity under degraded-node conditions.
ADARA Distributed Deception Detection EXPAND.3.S26B Adversarial deception detection across distributed nodes addresses cross-satellite trust requirements for distributed-system health management.
SATA Cross-Node Trust Monitoring EXPAND.3.S26B Continuous health and trust monitoring suitable for cross-satellite health exchange under intermittent-uplink conditions.
CARA Degraded-Communication Recovery EXPAND.3.S26B Autonomous recovery sequences operate under degraded-communication and intermittent-uplink conditions, addressing distributed-system autonomous-recovery requirements.

Reference: NASA SBIR/STTR 2026 Appendix A and B Q&A Hub. Per published Q&A, terrestrial Earth-analog and processor-in-the-loop testbeds satisfy validation requirements; expected TRL 3 (Phase I) and TRL 4 (Phase II).

Software Layer: Standards & Policy Alignment

AUTHREX Systems venture site · Reference architectures (TRL 3-4) · Single-author research, not deployed

The software authority-governance systems published on AUTHREX Systems are positioned, on a self-assessed basis, against the items below. Some are technical security standards (NIST SP 800-82, ISA/IEC 62443); others are Federal policy frameworks and legislative priorities (NDAA sections, Space Policy Directive 5, Five Eyes joint guidance) that the work is designed to support rather than engineering standards it is certified against. These are reference architectures, not certified or deployed systems.

Software SystemStandard / FrameworkSelf-assessed alignment
AUTHREX-AGENTFive Eyes joint guidance: "Careful Adoption of Agentic AI Services" (CISA / NSA et al.)Authority-lifecycle gating, per-tool authorization, and a hash-chained audit ledger for agentic AI support the joint-guidance principle of constraining autonomous tool use.
AUTHREX-ASSURE2026 National Cybersecurity Strategy; NDAA Section 1533Pre-deployment assurance gating addresses the priority of establishing that an autonomous system is authorized to act before it acts.
AUTHREX-ICS-GATENIST SP 800-82; ISA/IEC 62443; NERC CIPAuthority gating across the IT/OT boundary aligns with OT security controls and zone-and-conduit segmentation principles. Self-assessed; not an ISA/IEC 62443 certification.
AUTHREX-AGENT-CYBERFive Eyes agentic-AI guidance; DARPA AIxCC; NDAA Section 1513Governance-only authority constraints for autonomous cyber-defense agents. No offensive function.
AUTHREX-SPACECYBERNASA SBIR 2026 EXPAND.3.S26B; Space Policy Directive 5Onboard authority governance for orbital autonomy under intermittent ground contact aligns with space-system cybersecurity principles.
AUTHREX-SANDBOXNDAA Section 1534; NDAA Section 1533Test-and-evaluation authority governance aligns with the directed DoD activity on AI sandbox environments.

How to read these mappings

The mappings on this page are self-assessed technical alignment claims, not certification claims. AUTHREX has not been certified under any of these standards. The mappings describe how AUTHREX components could be used as evidence within a certification package, subject to:

A formal traceability matrix from each AUTHREX architecture paper (HMAA, SATA, CARA, FLAME, MAIVA, ADARA, ERAM) to specific clauses of these standards is part of ongoing work and will be released as Zenodo deposits as it matures.

Looking for the technical depth?

AUTHREX architecture papers and deposits on Zenodo include formal specifications, simulation material or bills of materials; contents vary by deposit, so check each deposit's file list. TLA+ source files are published where a deposit lists them.