DO-178B 적용 201372235 김영승 201372237 이선아
목차 프로젝트설명 프로젝트요구사항분석 DO-178B 적용 CMMI 평가
프로젝트설명
프로젝트설명 소형무인항공기에서촬영한동영상은환경적요인의영향을많이 받고, 특히바람등에의한영상의흔들림이발생. 무인항공기에서촬영한동영상을대상으로영상안정화기를구현 및실험. 프로젝트의목적은흔들림 ( 잡음 ) 을포함하고있는동영상을안정화하 여, 부드러운동영상을얻고자함.
프로젝트설명
프로젝트요구사항분석 Safety Critical 측면 무인항공기에장착된카메라를통해항공기를조작. 이때, 화면이심하게흔들리거나, 사람이인식하기힘들다면무인항공기운행에큰문제가발생할수있음. 인증대상 카메라로부터촬영된영상을실시간으로안정화시키는소프트웨어.
Process Overview
DO-178B 적용
PSAC(Plan for Software Aspects of Certification) 인증계획문서, 어느시점에인증절차를진행할지상세히기록. 각 SOI 에서검토할문서와검토방법들을기록. SOI#1 : PSAC,SQAP,SCMP,SDP,SVP 5 가지계획문서검토 SOI#2 : Development Process 인증을위해검토 SOI#3 : Verification Process 인증을위해검토 SOI#4 : 전체프로세스에대한검토
PSAC(Plan for Software Aspects of 전체시스템개요 Certification)
PSAC(Plan for Software Aspects of Certification) 전체소프트웨어개요 Previous Frame Current Frame Optical Flow Computation Affine Motion Parameter Calculation Parallel processing Outlier Rejection Stabilized Frame Image Warping Accumulate Motion Parameter Affine Motion Parameter Smoothing
DO-178B 적용
Safety Assessment & Requirements Software Level Level A (Catastrophic) Failure Condition 비행기와승객전체의손실을초래 LevelB (Hazardous) Level C (Major) Level D (Minor) 모든사람이죽지않을수도있지만그렇게될수있는위험성이높음 사람들이상처를입거나항공기가통제불가능하게될수있음 어떤영향을끼치지만비행기에의해극복될수있고, 조종사는통제할수있음 Level B Level E (No Effect) 아무영향을주지않음 Safety assessment 촬영된영상이사람이식별하기힘들다면무인항공기운행의큰영향을미침. But, 사람의목숨의위험한정도는아님.
DO-178B 적용
Start QA(Quality Assurance) Start QA : Software Quality Assurance Plan & Start 모든프로젝트의프로세스, 계획그리고표준이 DO-178B를따르는것을확인하는것 Software Quality Assurance Plan 누가 QA를할것인가? 5가지핵심계획과 3가지표준이 DO-178B를준수하는가? 얼마나자주감사를실시하는가? 테스팅또는코드검토를증명할것인가?
DO-178B 적용
Systems Requirements System Requirements 무인항공기가운행될수있도록시야가확보되어야함 지상에서무인항공기가통제되어야함 일정시간이상비행이되어야함 비상착륙이되어야함 등등
DO-178B 적용
Develop Plans, Standard, Checklist 개발, 검증, 형상관리계획과표준및체크리스트를수립 SDP(Software Development Plan) SVP(Software Verification Plan) SCMP(Software Configuration Management Plan)
Software Development Plan Requirement System Testing General Design Specification Integration Testing Detailed Design Specification Unit Testing Source Code Software Development Life Cycle
Software Development Plan Software Development Environment Windows 7 Intel Quad Core Q6600 CPU(2.4GHz) 4GB Main Memory Visual Studio 2010 Professional OpenCV 2.1 Development Standard JSF Air Vehicle C++ Coding Standard 4. C++ Coding Standards 4.9.1 Naming Identifiers : 만족
Software Verification Plan Software Reviews and Analyses + Software Testing High-Level Requirements Low-Level Level Requirements Software Architecture Source Code the Outputs of the Integration Process Verification 방법 Testing Tool : VectorCAST Review and Analysis : 직접수행 Traceability : Traceability Tree
Software Verification Plan Requirement System Testing General Design Specification Integration Testing Detailed Design Specification Unit Testing Software Verification Life Cycle Source Code
Software Configuration Management Plan 소프트웨어라이프사이클동안산출된각종결과물들을체계적으로 관리하는기법 소프트웨어에가시성과추적가능성을부여하여품질보증을도모 모든단계의모든이슈를빠짐없이작성 (Problem Reports) 이슈가해결된문서는릴리즈번호로태깅하고따로보관 이때릴리즈형상을 SCI(software configuration index) 로보관. 릴리즈형상이란릴리즈시점의모든문서와코드들의목록을의미. SVN 이나 SCADE 를이용하면이와같은작업을쉽게적용가능
DO-178B 적용
DO-178B 적용
Process Overview
DO-178B 적용
High-Level Requirements System Requirements Software Requirements (High-Level Requirements) 프레임안정화 뚜렷한영상촬영 실시간성 무인항공기가운행될수있도록시야가확보되어야함
DO-178B 적용
Low-Level Level Requirements 옵티컬플로우 프레임안정화알고리즘어파인모델 뚜렷한영상 촬영 이상치제거 실시간성 병렬스레드 Low-Level Requirements
Timing Constraints Low-Level Level Requirements 지연시간 : 프레임이입력되는최대지연시간은 30ms Memory Size Constraints Frame Size : 352 * 288 Buffer : Frame Size * 3 개 (thread 개수 ) Shared List Hardware and Software Interfaces Format : CIF Frequency of Input : 300 frames/sec Frequency of Output : 30 frames/sec
DO-178B 적용
General Design Specification Frame Sequence Display Frame t Frame t+1 Frame t+2 Frame t+3 Frame t+4 Frame t+5 Image Warping Accumulate Affine Parameter Affine Parameter Smoothing Optical flow computatio n Affine Parameter Calculation Thread Optical flow computatio n Affine Parameter Calculation Thread Task Manager Optical flow computatio n Affine Parameter Calculation Thread Frame Data Affine Parameter Frame Data Affine Parameter Frame Data Affine Parameter Frame Data Affine Parameter Shared List Cii Critical lsection
Detail Design Specification Algorithm Optical Flow Gunnar Farneback 이제시한 Densen Optical Flow Algorithm Affine Model Savitzky-Golay 가제안한 Affine Motion Parameter Smoothing Algorithm Affine Parameter Calculation Outlier Rejection
Control Flow Detail Design Specification Frame t-3 Frame t-2 Frame t-1 Frame t Frame t+1 Computed Affine Parameter Computed Affine Parameter Computed Affine Parameter Computed Affine Parameter Computed Affine Parameter Parallel processing Accumulated Affine Parameter Accumulated Affine Parameter Accumulated Affine Parameter Accumulated Affine Parameter Accumulated Affine Parameter Image Warping Image Warping Image Warping Image Warping Image Warping Stabilized Frame t-3 Stabilized Frame t-2 Stabilized Frame t-1 Stabilized Frame t Stabilized Frame t+1
DO-178B 적용
Code & Logic
DO-178B 적용
DO-178B 적용
Verification & Validation Test VectorCAST 를이용해 testing 수행 Review and Analysis 각단계의결과물들이상위레벨의요구사항이나표준을만족하 는지확인 Traceability Tree 를이용해만족여부를확인
Verification & Validation (Traceability) Software Requirements (High-Level Requirements) 옵티컬플로우 calcopticalflowfarneback GetAffineParameter 프레임안정화 알고리즘 어파인모델 low_pass_filter_25points 뚜렷한영상 cvwarpaffine 촬영 이상치제거 OpticalFlowReject 실시간성병렬스레드 TaskManager Low-Level Requirements Source Code
DO-178B 적용
DO-178B 적용
DO-178B 철학 Planning Process 가굉장히중요한것같음 Plan 을굉장히꼼꼼하고구체적으로작성 엄중한검토및검증과정 천천히, 꼼꼼히, 그리고반복적으로검토및검증을수행 어떻게 부분이자유로운대신, 모든부분을작성 형상관리계획에따라, 모든이슈를작성
CMMI 평가 DO-178B 적용을한조직이라면, 적어도 CMMI Level 2 의평가조건을충분히만족할것이라판단됨.