A real-time, visual and auditory AI-powered aircraft factory-level, periodic, line maintenance, service, securing, and fault repair verification system.

Abstract

The invention is an offline, AI-powered monitoring and assistance system that uses wearable cameras and computer vision (OCR) to live-monitor control steps during factory-level, periodic, pre-flight / inter-flight / post-flight line maintenance, service / supply, ground securing, and fault repair operations on aircraft; digitally locks out a step if torque, pressure, fluid level, or safety pin checks do not fall within target limits; and provides audible / visual guidance to personnel with AMM / SRM instructions.

Description

1 TARIFF REAL-TIME AIRCRAFT WITH VISUAL AND AUDIO EQUIPMENT. FACTORY LEVEL, PERIODIC, LINE MAINTENANCE, SERVICE, SAFETY AND FAULT REPAIR VERIFICATION SYSTEM 5 1. SUMMARY OF THE INVENTION The invention is for heavy maintenance at depot level (overhaul) of civil and military aviation platforms. periodic maintenance (A, B, C, D maintenance), pre-flight (walkaround), in-flight / transit (transit / turnaround), post-flight line maintenance, servicing and maintenance (fuel, oil, 10 hydraulic, oxygen, nitrogen liquid and gas supplies), ground securing operations (safety pin, pitot-static (protective casing, grounding line, wedge checks) and unplanned fault repair and troubleshooting operations during; all inspection, assembly, repair and maintenance carried out by technicians, mechanics or flight crew. artificial measurement steps are determined through visual and auditory sensors without relying on human input. 15 that enables real-time verification with intelligence and monitors the process with locked digital approval layers. It is a dynamic assistant system. The system monitors the technician's work area, critical external fuselage / engine components of the aircraft, and the cabin / cockpit. their equipment, factory-level disassembly / assembly areas, and all smart / analog measurement and service systems used. gauges (torque meter, pressure gauge, hydraulic and engine oil level sight glass - sight 20) (glass, tire tread depth gauge, brake wear pin, etc.) live signals are transmitted via visual and auditory sensors. It analyzes it as follows: Computer vision and optical character recognition (OCR) All physical values, liquid levels, and fault repairs that need to be seen and read with algorithms It verifies the steps and part conditions. In case of values ​​detected outside of tolerances, in case of limit exceedances, Locked Digital Confirmation and Blocking Layer 25 on floor safety equipment that is forgotten to be removed or installed By activating the Lockout Layer, it prevents the task from being registered as "completed"; exceeding the limits. By automatically recording the deviation, reports and values / controls are limited by artificial intelligence. It keeps the process locked until it is brought under control. Also, structural damage, deformation, corrosion, cracks, or Aircraft maintenance and repair manuals (AMM) are used for situations requiring repair and troubleshooting, such as leaks. A multi-channel decision support system offering audio and visual solution procedures with reference to SRM, CMM, MEL. It is the mechanism. 2. THE TECHNICAL FIELD TO WHICH THE INVENTION RELATES This invention is used in aviation for Aircraft Maintenance Manuals (AMM), Structural... Repair Manual (Structural Repair Manual - SRM), Component Maintenance Manual (Component Maintenance 35) Manual (CMM), Master Minimum Equipment List (MMEL / MEL), Flight Operation Manuals (FCOM), Factory / Warehouse Level Maintenance Standards, Line Maintenance Digitalization of Maintenance and Ground Services Service Standards; wearable or mobile border Edge AI devices, computer vision, optical character recognition (OCR), borderline Small Language Models (SLM) running on devices, augmented reality (AR) interfaces, locked digital authentication 40 The invention relates to (Checklist Lockout) architectures and hardware measurement equipment interfaces. Specifically, the entire lifecycle of the aircraft, from production to factory overhaul and daily flight lines eliminating the risk of improper approval (checking despite faulty / incomplete processing) throughout, human error in assembly, torque, pressure, fluid supply, troubleshooting and securing. It is related to blocking systems. 45 50 2 3. STATE OF THE ART AND THE SOLUTIONS PROVIDED BY THE INVENTION In the known state of the technology, checklists are used in aviation operations (task cards / checklists / (walkaround list / work order) Personnel must check the relevant box on paper or a digital tablet. It is based on this principle. However, time pressure, adverse environmental conditions, fatigue, and inattention can cause problems. Therefore, personnel may leave insufficient torque during factory-level assembly, and during periodic maintenance, 5 This allows them to bypass corrosion steps without removing the pitot tube sleeve or landing gear before flight. Check the oil level during end-of-flight and transit maintenance without removing the ground lock pins. Using non-standard parts / methods in verification or troubleshooting processes and marking the process as "completed" It can be registered as such. Objective, visual, and reliable methods are needed to prevent these omissions in existing systems. There is no hardware-based blocking mechanism. In aviation, insufficient torque, sub-limit hydraulics 10 Pressure, an unremoved safety pin, or a faultyly executed repair can lead to catastrophic flight accidents. This leads to operational downtime. Furthermore, manual scanning of manuals during unplanned breakdown repairs causes disruptions. This significantly increases their AOG (Aircraft On Ground) times. The technical solutions offered by the invention are summarized below: 15 • Scope of All Maintenance and Service Phases: System factory-level major overhauls, planned periodic maintenance, pre-flight external checks (walkaround), quick return trip between flights (turnaround / transit) maintenance, end-of-flight checks, service / supply services, and fault repair / correction operations It covers everything completely. • Visual Indicator, Pressure Gauge and Level Verification (Computer Vision & OCR): In addition to torque meters, 20 fuel / oil / hydraulic sight glasses, tire pressure gauges, oxygen tank Indicators and brake wear pins are read live using computer vision. • Locked Digital Checklist Lockout: Measured value, assembly torque, or detection Unless the reported physical condition is within the target range, personnel must give approval via the screen or button. Blocked, step locked. 25 • Automatic Deviation Reporting and Limit Integration Cycle: Every out-of-limit check (e.g., low tire pressure) (pressure, insufficient torque, faulty repair) are logged immediately, and corrective actions to address the missing parameter are sent to the personnel. The procedure is dictated, and the lock is not released until the status is visually brought within the limits. • Critical Ground Securing Equipment Confirmation (Safety Lockout): Landing with "Remove Before Flight" flag. Kit includes locking pins, pitot / static port covers, engine intake covers, and grounding wires (see image 30). Maintenance and flight approvals will not be granted without verification. • Dynamic Fault Detection and SLM Assistance: Cracks, corrosion, or leaks in the housing are detected by the camera. When displayed, the system scales the extent of the damage pixel by pixel, AMM / SRM / CMM It finds the official repair procedure in its library and communicates it to the technician both audibly and visually. • Offline (Edge) and Deterministic Security (Guardrail): The system is internet-independent, local 35 It runs on hardware. It prevents Small Language Models (SLMs) from generating free text. It contains deterministic safety barriers. 4. BRIEF DESCRIPTIONS OF THE FIGURES Figure 1: The subject of the invention is factory-level, periodic, line maintenance, service, safety and troubleshooting 40 data flow between the hardware units, measuring instruments and maintenance databases of the verification system It is a block diagram showing the general architecture. Figure 2: Measurement verification, out-of-limit deviation reporting, locked confirmation blockage, confirmation of bringing within limits. It is a master operational workflow diagram that includes AI-assisted fault repair steps. Figure 3: Locked control appearing on the wearable / mobile screen interface used by the technician / personnel 45 panel showing OCR display readout window, securing status and live voice dictation area It is a wireframe arrangement. 50 3 5. EXPLANATION OF REFERENCES IN THE FIGURES 1. Visual and auditory sensor unit (Wearable camera and microphone set) 2. Interface for measurement and service equipment (Torque meter, pressure gauge, level indicator) 3. Central processing and artificial intelligence analysis engine. 31. Computer vision and OCR analysis module 5 32. AI-based maintenance assistant (SLM / SRM analysis module) 4. Dynamic aircraft maintenance database (AMM, SRM, CMM, MEL and checklists) 5. Locked digital authentication and blocking layer 6. Technician voice and visual notification interface 100. Step 10: Assigning a maintenance, service, or safety task to the system. 101. Visual sensor and OCR for work areas, indicators or ground safety equipment. scanning step 102. Whether the scanned parameter or condition is within the target tolerance limits. questioning step 103. Step locking and automatic deviation reporting for parameters detected outside of tolerance 15 step 104. Step: Administering the corrective action to personnel verbally and visually. Step 105: Automatic confirmation and registration of the step when the target parameter is reached. 106. Step: Investigation of unplanned structural damage or leakage at the work site. 107. Step 20: The AI ​​assistant analyzes the damage and dictates the repair instructions. 108. Successful completion of maintenance and service operation step. 300. Intelligent maintenance, service and safety interface wireframe layout. 301. Active task and target parameter definition field 302. Real-time camera and OCR verification window 303. Locked approval and status reporting panel 25 304. Digital blockage and lock warning indicator. 305. Text field for voice dictation and guidance protocol. 6. DETAILED EXPLANATION OF THE INVENTION AND HOW IT WORKS The core structure of the invention is based on 30 processes in aircraft factories, maintenance hangars, aprons, and line maintenance areas. Completely local hardware (Wearable Smart Glasses, Body / Helmet) without needing an internet connection. An AI Maintenance, Service, Safety system running on a camera, tablet, or mobile field terminal. and Troubleshooting Engine (3). The system's dynamic aircraft maintenance database (4); AMM, SRM, CMM, MEL, FCOM, and Manufacturer Checklists are presented as isolated modules in a hierarchical JSON / SQLite format. It stores only the Small 35, which belongs to the aircraft being worked on, in order to optimize hardware resources. The Language Model (SLM) and computer vision models are run locally at runtime. How the System Works, Exception (Edge Case) Management and Steps: 1. Phase and Task Injection: Selected operational phase (Factory Level Maintenance, Periodic Maintenance, Pre-flight walkaround, transit / turnaround, post-flight maintenance, service / supply, or breakdown repair. 40 The task list for Repair / Removal) is uploaded to the system from the Dynamic Aircraft Maintenance Database (4) (100). 2. Real-time Monitoring with Camera and Sensor: Visual and Auditory Sensor Unit (1) carried by the technician, the technician's hands, the component he is inspecting, the safety pins, the repair area, and the measurement / service equipment he is using. He monitors his instruments (2) continuously. 3. Value / Status Reading with Computer Vision and OCR: Computer Vision and OCR 45 in the control step Analysis Module (31) scans the working area and indicators (101). Module (31) scans the torque meter indicator, tire pressure gauge, hydraulic / oil level glass, brake wear pin, removed / installed part It scans the serial number or the presence status of the safety case in real time. 4. Out-of-Limit Detection, Reporting, and Lockout Routine: Targeting the scanned parameter or status. It is checked whether it is within the tolerance limits (102). The reading (e.g., 72 Nm instead of 75 Nm 50 torque, tire pressure of 170 PSI instead of 200 PSI, insufficient engine oil, unremoved landing gear pin or limiter If the target boundaries (4) in the database are outside the crack; Locked Digital Approval and Blocking Layer (5) 4 It is activated. Personnel can complete their step via the smart maintenance, service and safety interface (300). Even if the input is made, the system will override this approval, lock the step, and automatically correct the deviation. It logs and reports (103). While the system shows the blockage status in the interface panel (303, 304), Technician Audio / visual notifications can be sent to staff via the Audio / Visual Notification Interface (6) and the audio dictation text field (305). It dictates the missing parameter or corrective step (104). 5 5. Limit Internal Verification and Automatic Approval: When the employee corrects the transaction and the value enters the target range, When the safety pin is removed or the part assembly is completed in accordance with the manual, the OCR module (31) confirms the correct parameter; unlocks the blocking layer (5), step objectively verified registers (105) and then proceeds to the next task. 6. Two-Way Voice Interaction and Dynamic Failure Repair Analysis: Unplanned structural damage during flow 10 or the presence of leaks is checked (106). Personnel find an unplanned fault during maintenance or troubleshooting. When it detects a defect (crack in the body, fuel leak, cable damage), it points the camera at the defect and announces it audibly. "There is a crack in the body, what is the repair procedure?" asks the AI-based Maintenance Assistant (32). It processes the image and scales the extent of the damage pixel by pixel, using local SRM / AMM / CMM. scans libraries (4) and 15 that should be implemented for staff via headset (6) and interface (305) dictates the official troubleshooting instructions verbally (107). If there is no damage or after the repair is approved The operation is completed successfully (108). 7. Deterministic Security Barrier (Guardrail): SLM models (32) running in the system, security barriers Isolated from free / productive text generation via (Guardrail); only approved aviation It provides limited responses with its manuals and the noise filter (SNR) detects erroneous commands in ambient noise. 20 It prevents it from being detected.

Claims

REQUESTS 1. The invention concerns factory-level maintenance, periodic maintenance, and pre-flight (walkaround) services for aircraft. Transit / turnaround maintenance, post-flight maintenance, servicing and maintenance, ground securing. and 5 that check the accuracy of the control steps in fault repair and troubleshooting operations. an AI-powered maintenance verification system that provides dynamic decision support to technical personnel Its characteristic is; • Physical interventions at the work site, measuring instruments, indicators, service / supply areas, at least one visual and audible sensor unit monitoring fault repair areas and ground safety equipment (1), 10 • Factory / heavy maintenance guidelines, periodic checklists, service tolerances, AMM, SRM, At least one dynamic aircraft maintenance database containing CMM, MEL and FCOM standards (4), • Obtaining indicator values ​​of measuring equipment, liquid levels, and physical parameters through visual data. damage, part assembly suitability and safety components (safety pins, protective sleeves, (grounding lines) presence status is read in real time by a computer vision and OCR 15 analysis module (3.1), • The reading corresponds to the target limits (4) in the database of fault repair or physical condition. When the discrepancy is not present and the control cannot be visually brought within the limits, the information provided by the technician... By overriding manual approval entries, it prevents the registration of the relevant step and exceeds the limits. A Central Transaction and 20 which includes a Locked Digital Confirmation and Blocking Layer (5) that automatically reports the status. It contains an Artificial Intelligence Analysis Engine (3).

2. It is a system that complies with Claim 1, and its feature is that the computer vision and OCR analysis module (3.1) is factory torque meters, pressure gauges, and liquid level gauges used in periodic or line maintenance. The numerical / visual value on the sight glass, tire tread depth caliper, or wear pin is 25. It instantly detects and reads the data, and as long as the read value does not reach the tolerance limits, the digital verification is locked. The blocking layer is physically and software locked until step (5) is confirmed as "completed". It is holding on.

3. A system that complies with claims 1 and 2, and whose characteristic is; the targeted measurement, torque, pressure or service 30 If the level cannot be reached, the technician will notify the technician via the audio and visual notification interface (6). The technician audibly indicates the amount of deviation in the auditory canal and the corrective action that needs to be applied. The method of transmission via dictation protocol is to maintain the locked state until the transaction limit is reached.

4. A system that complies with Claim 1, and its feature is: securing before, during or after flight. 35 landing gear safety pins (ground lock pins) that must be removed or installed during inspections, pitot-static protective covers, engine covers and grounding cables visual sensors (1) and artificial If not confirmed by intelligence (3.1), the relevant safety step of the locked confirmation layer (5) and It blocks the flight approval process. 40 5. A system conforming to Claim 1, featuring; engine oil, hydraulic, fuel, oxygen and nitrogen service. Level and pressure gauges are visually monitored using artificial intelligence during services and maintenance. personnel are instructed to detect under or overfilling situations and bring them within the limits. It provides real-time audio / visual feedback. 45 6. It is a system that complies with Claim 1, and its feature is that the visual and auditory sensor unit is connected by the technician (1) Physical damage (cracks, corrosion, etc.) on the aircraft being straightened requires repair and elimination of malfunctions. Deformation, leakage, tire / brake wear) dimensions are determined pixel by pixel using image processing algorithms. (4) Structural Repair Manuals in the dynamic aircraft maintenance database that scales through (SRM), by aligning with CMM and MEL standards, defines the official troubleshooting and repair guideline. It includes an AI-based maintenance assistant (3.2). 6 7. A system that complies with claims 1 and 6, characterized by its bidirectional natural language communication between the technician and the system. It provides voice communication and answers the technician's questions regarding maintenance, service, and troubleshooting steps. via a Small Language Model (SLM) that runs offline / edge on the device a company that responds solely by referencing registered aviation maintenance documents (AMM / SRM / CMM / MEL) It has a deterministic safety barrier (Guardrail). 5 8. A system compliant with claims 1 and 7, characterized by its Small Language Model (SLM) operating on the device. By preventing the generator / free text from generating responses, it only provides approved maintenance manuals. limiting sound entry in cases of high acoustic noise in the factory or apron area. It prevents erroneous command detection by filtering the signal-to-noise ratio (SNR) in the channel. 10 9. It is a system compliant with Claim 1, and its feature is; Small Tongue Model (SLM) for different aircraft types. their weights, factory-level revision data, and maintenance libraries are independent of each other. to be structured in modules and to conserve memory / processor resources of the local hardware runtime 15 of the isolated SLM module, which is specific to the target aircraft platform being studied. A dynamic package management system that enables local execution during runtime. It contains a mechanism.

10. A system compliant with Claim 1, characterized by: factory-level, periodic, line maintenance, service and troubleshooting. All measurement, level, torque, and safety checks that fall outside the limits during the troubleshooting operations, 20 Digital maintenance records detected deviation amounts, applied troubleshooting procedures, and approval times. It is the process of automatically reporting and registering it in the log with a timestamp.

11. A system compliant with Claims 1, 2, 4, 5 and 6, whose feature is; all measurement, assembly, service, troubleshooting and Computer vision and OCR 25 ensure that ground safety parameters fall within the targeted limits. After confirmation with module (3.1), the Locked Digital Confirmation and Blocking Layer (5) is automatically blocked. by removing it and objectively verifying my step without requiring manual input. It is about moving to the next stage.

Images