PWS - AI-Enabled Support.docx

PERFORMANCE WORK STATEMENT (PWS)

AI-Enabled Maintenance Decision Support and Edge Computing Platform RAF Lakenheath & RAF Mildenhall, United Kingdom

1. DESCRIPTION OF SERVICES

1.1 Background

Aircraft and ground vehicle maintenance operations within the United States Air Forces in Europe (USAFE), specifically at RAF Mildenhall supporting the 100th Maintenance Group (100 MXG) and 100th Logistics Readiness Squadron, Vehicle Management Flight (100 LRS/LGRV), require rapid, accurate access to complex technical data and diagnostic troubleshooting procedures.

Currently, technical order (TO) documents and commercial/military vehicle technical manuals are managed across separate digital systems (e.g., Enhanced Technical Information Management System [ETIMS], disparate PDF repositories, and commercial fleet diagnostic databases). Navigating multi-volume TO libraries and fault isolation trees on the flightline and in vehicle bays is labor-intensive, relies on keyword searches, and can lead to procedural omissions or extended troubleshooting timelines.

To enhance mission readiness, reduce Mean Time to Repair (MTTR), and streamline technician workflows, the Government requires a commercial, production-ready Artificial Intelligence (AI) decision-support software platform. This system must integrate Technical Orders, maintenance records, controller area network (CANbus) telematics, sensor inputs, and technician knowledge to guide personnel through step-by-step diagnostic and repair procedures on ruggedized mobile edge devices in both connected and disconnected environments.

1.2 Scope

The Contractor shall provide non-personal services, including all personnel, equipment, tools, software licenses, supervision, and services necessary to deploy, integrate, secure, and sustain a commercial AI-enabled maintenance decision-support platform at RAF Mildenhall, UK.

The software platform shall integrate aircraft technical orders, ground fleet service manuals, historical maintenance data, and vehicle diagnostic telemetry (CANbus) into modular, guided maintenance workflows. The system must operate securely on mobile edge devices (e.g., ruggedized tablets) with full operational capability during disconnected operations. The Contractor shall support the entire capability lifecycle, including operational testing, cybersecurity engineering under the DoD Risk Management Framework (RMF) leading to an Authority to Operate (ATO), user onboarding and training, hardware integration, and ongoing system sustainment.

1.4 Specific Support Services

1.4.1 Objective 1 – Initial System Demonstration and Operational Assessment The Contractor shall deploy a preliminary capability on designated Government-furnished or contractor-provided evaluation edge devices to conduct a baseline operational demonstration.

The system shall be demonstrated against representative maintenance and troubleshooting workflows (e.g., aircraft subsystem fault isolation, ground vehicle powertrain/hydraulic diagnostics).

The Contractor shall capture end-user evaluation metrics, including user experience, procedural accuracy, troubleshooting latency, and system usability, delivering an Operational Assessment Report to support baseline operational decisions.

1.4.2 Objective 2 – Technical Data Ingestion and System Architecture Data Ingestion: The platform shall directly ingest, parse, index, and structure technical publications from diverse formats, including Air Force ETIMS-exported TOs, Standard Generalized Markup Language (SGML), Extensible Markup Language (XML), commercial vehicle technical service manuals, PDF technical documentation, and relevant instructions (e.g., DAFMAN 21-101, DAFI 91-203, and local operating instructions).

Telemetry and Sensor Integration: For vehicle maintenance operations, the platform shall interface with commercial-standard vehicle diagnostic interfaces (e.g., J1939, OBD-II/CANbus) and vehicle-mounted edge compute telemetry modules. The system must ingest live and recorded vehicle fault codes and operational parameters to automate initial diagnostic path selection.

Maintenance Record Context: The software shall support correlation with historical maintenance logs and maintenance information systems (e.g., G081, DPAS, or local documentation) to identify repeat discrepancies and common failure modes.

1.4.3 Objective 3 – Edge Computing and Disconnected Operations The Contractor shall deliver an AI inference engine and technical data repository capable of executing fully on edge mobile devices (e.g., ruggedized tablets) without reliance on continuous cloud or local area network connectivity.

The edge solution must maintain full core operational capability (including natural language technical queries, fault trees, and step-by-step procedure guidance) with zero network connectivity for sustained operational deployments lasting up to 180 calendar days.

The system must execute local AI queries and return verified technical results with sub-second to low-second latency on standard mobile edge hardware.

1.4.4 Objective 4 – AI Accuracy, Determinism, and Source Verification (Zero Hallucination) To preserve flightline and operational safety, all generative or retrieved AI outputs must be strictly deterministic and derived solely from the ingested and authorized Government/technical data corpus. The system is strictly prohibited from generating external assumptions, unverifiable technical steps, or uncited operational procedures.

The AI engine must provide concurrent, interactive source citations (document title, chapter, section, and page/figure reference) for every technical instruction and troubleshooting recommendation, enabling technicians to visually corroborate procedures against authorized data sources.

1.4.5 Objective 5 – Data Synchronization and Software Patching The platform shall feature an automated, resilient synchronization mechanism that reconciles data whenever edge devices connect to an authorized network (wired or wireless).

Synchronization shall update local technical data repositories, apply security patches, incorporate model updates, and transmit anonymized maintenance metrics without requiring end-user administrative intervention.

1.4.6 Objective 6 – Cybersecurity Engineering, RMF, and ATO Support The Contractor shall execute cybersecurity engineering required to support system accreditation under the DoD Risk Management Framework (RMF).

The Contractor shall develop, maintain, and provide all necessary RMF artifacts, Security Technical Implementation Guide (STIG) compliance checklists, software vulnerability scans, and system architecture diagrams necessary to achieve and sustain an Interim Authority to Test (IATT), Interim Authority to Operate (IATO), and full Authority to Operate (ATO) on DoD networks.

The solution shall comply with FAR 52.204-21, NIST SP 800-171, and applicable DoD policies for safeguarding Controlled Unclassified Information (CUI).

1.4.7 Objective 7 – Deployment, Training, and User Onboarding Upon successful validation and security authorization, the Contractor shall support fleet-wide deployment across all designated devices within the 100 MXG and 100 LRS operational footprints.

The Contractor shall provide comprehensive New Equipment Training (NET), including flightline maintainer onboarding, vehicle technician instructions, and system administrator training.

The Contractor shall deliver modular user guides, quick-reference handbooks, and digital training aids.

1.4.8 Objective 8 – System Sustainment, Model Tuning, and Technical Support The Contractor shall provide tier-2 and tier-3 help desk technical support, issue resolution, and system maintenance.

The Contractor shall deliver continuous AI model maintenance, refining semantic search relevance, procedural mapping, and diagnostic accuracy based on user feedback and updated technical order publications.