This page logs every technical capability change across the FSG-A system — new SDK releases, new code modules, new Fischer 26/26E functions, new Lisa 26 algorithms, new interoperability with Swedish Armed Forces systems, and new external integrations. Textual edits to wiki pages are not listed here; only engineering changes that affect verification, procurement, or operational capability. Military and research readers can use this log to confirm the exact version of each system element and the date of its introduction.
Purpose and Scope
This changelog exists because Försvarsmakten, FMV, FOI, and allied NATO evaluators need an auditable history of technical capability changes — not a list of prose revisions. Text clarifications, translation updates, and minor stylistic edits are tracked internally in the build system but deliberately omitted from this page. The entries below cover only changes that a systems engineer or procurement officer would care about.
Every entry contains: the release date, the SDK or system affected, a short description of the new capability, and the validation status (TRL level or proof reference where applicable). Rollbacks or deprecations are marked explicitly.
OFFENSIVE-200 / RELAY-200 / SPLIT-100-100 reconfigurable per mission via MAVLink parameter. 200 ms mode transition.
Stand-off role
Tier-3 cruise at ~1000 m AGL, 3–5 km behind FLOT. Above practical FPV ceiling (~600 m). Protects forward tier-1 (Fischer 26 at 300 m AGL) and tier-2 (Fischer 26E at 500–700 m AGL) via jamming at 5–15 km radius.
TRL
TRL 2 — analytical concept, no physical prototype built, no flight hours. All performance figures derived from first-principles physics and published component datasheets.
Physical sealed wire fuse on Mast C PA-enable line, 2 per airframe (one per gimbal's Mast C)
Function
Blocks civilian LTE/5G band transmission in peacetime regardless of software command state. Broken only by authorised officer in declared wartime zone, logged with officer identifier, date, and zone reference.
Design basis
Permissive Action Link (PAL) principle from nuclear weapons safety engineering. No software command — from operator console, flight controller, ground station, or USB-loaded profile — can activate transmission while fuse is intact.
Rationale for hardware over software
Software policy cannot be audited at the PA chip level; a physical fuse can. Tamper-evident lacquer seal provides visible evidence of authorisation. Distributed 2-fuse architecture avoids single-point failure on binary authorisation decisions.
TRL
TRL 2 — design principle documented, no prototype hardware built
Plain-text YAML profile loaded via USB by ground crew pre-flight. Defines per-mast center frequencies, hop lists, modulation, PA drive, duty cycle, and allowed dwell windows. Pushed to all 7 masts over CAN-FD.
Doctrine basis
Ukrainian EW practice since 2023 — adversary signatures change faster than classified acquisition cycles, so profile rotation belongs to the operating brigade, not to a national tooling authority.
Mast C hardening
Schema assertion forces pa_drive_w: 0 on Mast C regardless of wartime state — profile cannot enable civilian-band TX, only the physical fuse can.
Classification
Unclassified. Format is open and reproducible in any brigade workshop with laptop and text editor.
Fischer 26 €13,300–20,800; cluster €10,000–18,000. Analog Devices, Amphenol MIL-DTL-38999, Mini-Circuits ZHL-10W, CubePilot Kore carrier board.
Volume-break factors
17 % reduction at 10u, 40 % at 50u, 60 % at 500u — uniform across all three tiers
Components resisting Chinese substitution
Three: Starlink Mini (SpaceX-only), Silvus StreamCaster MANET (ITAR-controlled), Jetson Orin Nano (67 TOPS, no Chinese clone at envelope). All other Fischer 26 components have functioning Chinese equivalents.
Validation
All prices Q1 2026 single-unit distributor list prices, verified via web search on Digikey, Mouser, iFlight Europe, Aeroboticshop, AliExpress, LCSC. No formal procurement quotes obtained.
2026-04-21 — TIER-AWARE CPFH MODEL + EXPENDABILITY EQUATION
Release type
NEW COST-EFFECTIVENESS MODEL
Component
fischer26_cpfh.py (tier- and lifetime-aware) in fischer26-whitepaper.html
Model change
CPFH = (C_airframe / N_flights) + C_energy + C_maintenance. Both C_airframe (procurement tier) and N_flights (threat-environment-dependent lifetime) now treated as independent variables. Previously only C_airframe was varied.
Quantifies battery impact of sequential 10-target engagement batches at the LE variant's 207 W EMCON-3 cluster surge vs 74 W EMCON-2 baseline. Outputs engagement energy in Wh and residual endurance in summer/winter Nordic conditions.
LE variant is not energy-limited at 10 targets. Worst case consumes 7.2 % of usable battery. Real constraints are gimbal slew (30 s/target transit), CAN-FD throughput during simultaneous triplet retasking, and operator cognitive load.
Sustained high-tempo capacity
3 full 10-target batches (30 engagements) per 3-hour winter mission with 110 Wh margin. Validates 2,500 Wh battery sizing on LE.
src/assets/.htaccess, src/assets/manifest.webmanifest, build.py updated to copy both
.htaccess directives
HTTPS redirect, gzip (9 content types), cache headers (1 year static, 1 hour HTML, 0 sw.js), HSTS, X-Frame-Options SAMEORIGIN, X-Content-Type-Options nosniff, Referrer-Policy strict-origin-when-cross-origin, Permissions-Policy lockdown of unused capabilities, CORS for fonts, filesystem blocking of dotfiles/py/yaml/md
PWA manifest
Mobile home-screen install capability with FSG-A branding
Mobile-first audit results
15 media queries at 7 breakpoints (600/640/768/900/901/1200/1440 px); status-table overflow-x at ≤640 px; img/svg max-width; pre/code overflow-x; touch targets ≥44 px
AI-search optimisation
Title + meta description per page (all under length caps), canonical URLs, hreflang en/sv/uk/x-default, OpenGraph tags, Twitter Card, JSON-LD schema, sitemap.xml with 39 language-specific URL sets, robots.txt
Validation
🟢 27 GREEN / 0 RED session protocol, SEO 32/32 rules pass, 43,286 internal links resolve across 356 HTML files
2026-04-19 — FORMAL DOCUMENTATION SUITE
Release type
NEW DELIVERABLE CATEGORY
Component
formal-docs/*.py (generators) + 8 PDF deliverables in formal-docs/
Description
Eight reference-design PDF documents showing the structure a funded programme might use. All marked REFERENCE DESIGN — PUBLIC with concept disclaimer on cover: FMEA, TEMP, ILS, Risk Register, Safety Case, SBOM, Threat Assessment, Interop Matrix. Each document sketches the structure per its governing MIL-STD or STANAG standard. All dates, scores, and role names are illustrative placeholders for a real programme to replace. Regenerated deterministically from Python source.
Standards covered
MIL-STD-1629A, DoD 5000.02, MIL-STD-1388-1A, NATO AEP-7.3, MIL-STD-882E, ISO/IEC 5962:2021, NIST SP 800-30, STANAG 4586/4660/4609/4607/5525/4671/4795
Asymmetric OpenPGP encryption toolkit for secure delivery of source archives to Försvarsmakten, FMV, FOI, and NATO partners. Three operator scripts (setup, encrypt, decrypt) plus a 20-test stress suite on isolated throwaway keys.
Automatic target-handoff to FMV's GUTE II C-UAS system (Tridon Mk2 40mm, Trackfire 30mm, EW countermeasures, Giraffe 1X). Includes effector-recommendation logic based on target type and range.
EW-hardened tier-2 variant operating at 500-700 m AGL. SDR (AD9361) replaces Starlink for survivable anti-jam. €3,900 unit cost (10:8:1 ratio vs F26 and FPV).
TRL
TRL 3 — component analysis complete, no prototype built
provable_claims.py (distributed in fsga-code-v2.zip)
Description
63 numerical claims across Fischer 26/26E, Lisa 26, STANAG compliance, and cost analysis. Each proof has an identifier, derivation, and Python verification code. All 63 pass on every build.
Initial fixed-wing ISR/relay platform. 2.4 m wingspan, 8.5 kg MTOW, 710 Wh battery, 1.6-2.0 h endurance, Starlink Mini relay. All ArduPlane parameters published and SITL-verified.
TRL
TRL 3 — simulation verification complete, no physical prototype
This section is a cost-and-scope estimate for an organisation that wants to take the Fischer 26 reference design further — not a product roadmap for FSG-A. FSG-A's own work stops at TRL 3 (analytical verification via SITL and provable claims). Anything beyond TRL 3 requires physical hardware, test facilities, and a certification authority — resources that an open-design contributor by definition does not own. The estimates below help a potential adopter (FMV, FOI, Chalmers, Saab, RISE, Milrem, or an equivalent) judge what concretising the design would involve.
Airworthy prototype + at least 50 flight hours for airframe characterisation
Estimated effort
€50,000 – €100,000 for STANAG 4671 certification testing (illustrative order-of-magnitude based on published certification costs for comparable UAS)
Who could run it
Försvarsmaterielverket (FMV), FOI, a research or industry partner with flight-test infrastructure
ADOPTER STEP — GUTE II LIVE ENDPOINT INTEGRATION
Scope
Live interoperability testing against FMV's GUTE II C2 endpoint when available for partner connection
Dependencies
FMV GUTE II endpoint available for partner connection; formal interoperability MoU between FMV and any adopter
Reference implementation
libfischer26e.gute_ii in the SDK is ready for integration, tested in SITL against a mock endpoint
Important distinction: FSG-A is not seeking funding to build this ourselves. The reference design is published under CC BY-SA 4.0 so that any qualified organisation can adopt it. If you represent such an organisation — see Secure Delivery for how to reach us via PGP-encrypted email.
Versioning Policy
FSG-A follows semantic versioning for the SDK and for the wiki itself. Major releases change public API signatures; minor releases add capabilities; patch releases fix bugs without changing behaviour. The wiki edition (currently 2026-Q2) is bumped at each major release. Every release tagged here corresponds to a specific commit in the source archive, and the archive is rebuildable from the BATCH-LOG reference numbers.
How To Subscribe to Changes
Research and military consumers should watch this page directly and compare the latest dated entry against their local copy. There is no push-notification channel because FSG-A does not maintain user accounts; all changes are public under CC BY-SA 4.0 and propagate through the open-source build system.
For a cryptographic audit trail, the complete BATCH-LOG is distributed inside fsga-wiki-src.zip. Every entry in this technical changelog has a corresponding BATCH-LOG line with commit hash and file-level diff summary.