Published March 4, 2026 | Version 2.4.2
Report Open

HMAA — Hierarchical Mission Authority Architecture: Technical Assurance Report v2.4.2

Authors/Creators

  • 1. Georgetown University, School of Continuing Studies

Description

HMAA (Hierarchical Mission Authority Architecture) is a deterministic authority-management framework for autonomous systems operating in safety-critical and contested environments.

This deposit contains the HMAA Technical Assurance Report v2.4.2, which specifies and evaluates an authority engine that computes a scalar authority value A ∈ [0,1] from four inputs:

  • Q: operator quality

  • C: context confidence

  • E: environmental threat

  • τ: sensor trust

The engine uses a multiplicative base × gate × damping × τ structure with an Environmental Warning (EW) hysteresis state machine and a five-tier authority classification (T0–T4). The report includes:

  • A complete, self-contained piecewise mathematical specification of the engine (eng()), sufficient for independent re-implementation

  • Explicit EW hysteresis rules and dynamic weighting under EW

  • A deterministic boundary test suite (10 test vectors) covering invariant boundaries, EW thresholds, and guard behavior

  • Monte Carlo simulations with reproducible CLI commands and a fixed canonical seed

  • A formal safety assurance case (hazards H1–H5b), plus an FMEA and threat model

  • An Evidence Register mapping claims to CLI runs and content-stable SHA-256 hashes (excluding timestamp/run_id)

Scope / limitations: This is a software model and simulation report (not a certified flight computer). Safety properties are empirically validated using Monte Carlo tests plus deterministic boundary vectors; formal proof and hardware-in-the-loop validation are listed as future work.

How to reproduce (high level): The report provides command-line invocations for the core scenarios (CI gate, adversarial Monte Carlo, EW hysteresis stress, nominal scenario) and expected summary statistics.

Files

HMAA-Technical-Assurance-Report-v2.4.2.pdf

Files (285.6 kB)

Name Size Download all
md5:236635399c0486ef446d68d4290ff8cf
285.6 kB Preview Download

Additional details

Dates

Other
2026-03-04

Software

Repository URL
https://github.com/burakoktenli-ai/hmaa
Programming language
JavaScript
Development Status
Active