MIT ICAT National Workshop on Aviation Software Systems: Design for Certifiably Dependable Systems Civil Aviation Context Prof. R. John Hansman MIT International Center for Air Transportation rjhans@mit.edu 617-253-2271 MIT ICAT Objectives Define Current State of the Art Identify Key Issues and Needs Identify Promising Research Approaches Define Educational Needs and Approaches MIT ICAT System Scope Software Hardware Dependability problem can be fairly well defined with good specifications Tractable with current methods Hard MIT ICAT System Scope Environment Software Hardware Environment MIT ICAT System Scope Human Human Software Hardware Environment My Bias MIT ICAT System Scope Interactions Human Software Hardware Environment MIT ICAT What is High Dependability ? Civil Aviation Context Target Level of Safety Equivalent Level of Safety MIT ICAT MIT ICAT MIT ICAT Probability vs. Consequences Graph AC 25.1309-1A Catastrophic Accident Adverse Effect On Occupants Airplane Damage Emergency Procedures Abnormal Procedures Nuisance Normal Probable Improbable Extremely Improbable MIT ICAT Probability (per unit of exposure) Descriptive Probabilities FAR 1 JAR Frequent 10E-3 Probable Reasonably Probable 10E-5 10E-7 Improbable Remote Extremely Remote 10E-9 Extremely Improbable Extremely Improbable What is the correct unit of exposure : Flight hour, Departure, Failure MIT ICAT Level A - Anomalous behavior causes catastrophic failure Reduced capability of aircraft (safety margins, functionality) Reduced crew performance Injuries or discomfort to occupants Level D Anomalous behavior causes minor failure Large reduction in safety margins Inability of crew to perform Serious or fatal injuries to small number of occupants Level C - Anomalous behavior causes major failure Inability to continue safe flight and landing Level B - Anomalous behavior causes hazardous/sever-major failure Software Criticality Levels No significant reduction in aircraft safety Level E - Anomalous behavior causes no-effect on aircraft operational capability DO-178B “Software considerations in Airborne Systems and Equipment Certification” MIT ICAT Civil Aviation Applications Commercial Aircraft Fly by Wire/Light Flight Management Systems General Aviation Aircraft Very Light Jets Unmanned Air Vehicles Air Traffic Management Communication Navigation Surveillance Decision Support Integrated Air-Ground Systems MIT ICAT Civil Aviation Applications Commercial Aircraft Fly by Wire/Light Flight Management Systems General Aviation Aircraft Very Light Jets Unmanned Air Vehicles Air Traffic Management Communication Navigation Surveillance Decision Support Integrated Air-Ground Systems MIT ICAT Cockpit Evolution to Higher Criticality Boeing 747-200 Electro - Mechanical “Steam Gauge” Boeing 747-400 CRT - LCD Displays “Glass Cockpit” Boeing 777 Fly by Wire/Light MIT ICAT Vehicle Control Loops Inner Loops More Critical Pilot Displays CDU MCP Controls Trajectory Commands State Commands Manual Control FMS Autopilot Autothrust Flight Control Rate Navigation State Sensors MIT ICAT Fly-by-wire Systems A-320 Example Electrically controlled, hydraulically actuated * Rudder Slats Aileron Elevator Flaps * Trimmable . horizontal stabilizer Speed brakes Roll spoilers Ground spoilers * Rudder & stabilizer have back-up mechanical control Anomalies : eg Hard Over Failures, Redundancy Architectures, Software as Single Point of Failure MIT Typical Redundancy Architecture ICAT MIT ICAT Human Interaction - Manual Control Aircraft Pilot Coupling (aka PIO) Pilot Displays CDU MCP Controls Trajectory Commands State Commands Manual Control FMS Autopilot Autothrust Flight Control Rate Navigation State Sensors MIT ICAT QuickTime™ and a decompressor are needed to see this picture. MIT ICAT Mode Awareness Mode Awareness is becoming a serious issues in Complex Automation Systems automation executes an unexpected action (commission), or fails to execute an action (omission) that is anticipated or expected by one or more of the pilots Multiple accidents and incidents Strasbourg A320 crash: incorrect vertical mode selection Orly A310 violent pitchup: flap overspeed B757 speed violations: early leveloff conditions Pilot needs to Identify current state of automation Understand implications of current state Predict future states of automation MIT ICAT Complexity and Conditional Statements Used extensively in Pilot Guides “Through the FCU, an immediate climb/descent is initiated by selecting the desired altitude in the ALT SEL window and either pulling the set knob or pressing the LVL/CH P/B to engage the LVL CHANGE mode. Pressing the LVL/CH P/B also disengages PROFILE, however, if PROFILE is engaged, pulling the set knob does not disengage it, rather it initiates an immediate climb/descent to the altitude selected on the FCU. The exceptions are ...” MIT ICAT Evolution (code Reuse) leads to Lack of Underlying Model There does not appear to be a simple, consistent global model of current Autoflight Systems Not apparent in flight manuals Flight manuals focus on crew interface and procedures Manufacturer could not supply functional model or logic/control diagram Hybrid Automation Model created to allow analysis In absence of a simple consistent model, pilots develop their own ad-hoc models These models may not accurately represent AFS operation Concern in some (future) aircraft Individual pilot models may not be accurate Training/Design implications Models are created during nominal flight conditions and may not hold during abnormal or emergency situations Entropic Growth of Complexity MIT ICAT Functional Analysis Operator Directed Process Sanjay Vakill Thesis Autom atio n M odel is der ived fr om Fu nctional Analysis, ope rator and exper t us er inp ut. Iter ative Hum anCente re d Pr oto type Evaluation Stage Tr aining m ater ial is der ived fr om Autom atio n M odel. Tr ainin g Rep re sentation is cr eated. Automation Model Training Material Spe cification change s m us t be con siste nt w ith Autom atio n M odel. Softw ar e specification is der ived fro m Tr ainin g M ater ial. Software Specification Software System is cer tifie d against Autom atio n M odel. Configur ation M an age m ent ver ifies and m aintains con siste ncy w ith Au tom ation Mod el. Certification Configuration Management MIT ICAT New Functionality (eg Required Navigation Performance) MIT ICAT New Functionality Requires FMS and Memory Upgrades Cost Issues Maintenance MIT ICAT Maintenance and Capability Expansion (eg Memory) . Honeywell A320 Pegasus FMS Advanced Features Addition of LOC/VNAV autoflight capability GLS/MLS Precision approach FLS (ILS like) Non Precision approach Enhanced LOC capture Multiple same-type RNAV Runway Approaches Improved Offset entry and display Mixed QNH / QFE approach capability QNH range extended to 1100 HPA 2MB Navigation Database capability Expandable Through Software to 12+MB ARINC 615A Ethernet software and database loading …. http://www.honeywell.com/sites/aero/Flight_Management_Systems MIT ICAT Some Airline Comments However we have had lots of issues with the upgrades of the FMC software. 1. Magnetic course displays are a moving target. Each upgrade uses a different set of Magnetic variations so we have to revise our plates so that the FMC and the plates are in sync. 2. The vendor changes algorithms so the procedures that we or the FAA has designed are no longer flyable. Something like this is an unintended consequence of a "fix". Also the boxes are not regulated nor are specifications so this type of disconnect can occur. 3. The FAA is asking us and any other carrier approved to fly RNP SAAAR procedures to verify that the software is safe. I do not think that this is our job. But once again this goes back to lack of regulation. They have no way of assuring that the changes being made will be compatible with RNP SAAAR. Electronic Flight Bag MIT Information vs Navigation Requirements ICAT •Source: Brian Kelly, Boeing MIT ICAT Civil Aviation Applications Commercial Aircraft Fly by Wire/Light Flight Management Systems General Aviation Aircraft Very Light Jets Unmanned Air Vehicles Air Traffic Management Communication Navigation Surveillance Decision Support Integrated Air-Ground Systems MIT ICAT Very Light Jets Small turbofan aircraft Aircraft characteristics* Eclipse500 Mustang Adam700 Eclipse Aviation Cessna Adam Aircraft Phenom-100 ProJet D-Jet Embraer Avocet Aircraft Diamond Aircraft Epic LT HondaJet Safire26 Epic Honda Safire Aircraft Excel Sport Jet Spectrum 33 Eviation EV-20 Passengers: 4 to 8 Acquisition price: $m 1.4 to 3.6 Cruise speed: 340 to 390 kts Operating ceiling: 41,000ft to 45,000ft Range: 1100 to 1750 NM Take off field length: 2200ft to 3400ft Orders Eclipse: 2300 Adam: 75 Mustang: 330+ * for twin-engine VLJs (excludes D-Jet) MIT ICAT Eclipse 500 Cockpit Avio Avionics Suite MIT ICAT Next Generation ? George Jetson Car MIT ICAT Civil Aviation Applications Commercial Aircraft Fly by Wire/Light Flight Management Systems General Aviation Aircraft Very Light Jets Unmanned Air Vehicles Air Traffic Management Communication Navigation Surveillance Decision Support Integrated Air-Ground Systems MIT ICAT Spectrum of Current UAVs Sig Kadet II RC Trainer – 5 lb Aerovironment Black Widow – 2.12 oz. Boeing/ Insitu Scaneagle – 33 lb Gen. Atomics – Predator B – 7,000 lb BAE Systems Microstar – 3.0 oz. Boeing X-45A UCAV – 12,195 lb (est) AAI Shadow 200 – 328 lb NOAA Weather Balloon 2-6 lb Allied Aero. LADF – 3.8 lb Aerovironment Pointer – 9.6 lb Northrop-Grumman Global Hawk 25,600 lb Bell Eagle Eye – 2,250 lb 0 1 Micro 10 Mini 100 UAV Weight (lb) Tactical 1,000 10,000 Med Alt High Alt / UCAV 100,000 Heavy MIT Historical Comparison of Accident Rates ICAT 10000 UAVs (Average) General Aviation Yearly Accident Rate (Accidents / 100,000 hr) 1000 Air Force Aviation Commercial Airlines 100 10 1 0.1 1925 1935 1945 1955 1965 1975 1985 1995 2005 Year Notes: 1) UAV Accident Rates are averaged for Pioneer, Hunter, and Predator UAVs from OSD UAV Reliability Study. February 2003 2) General Aviation data from AOPA historical data, http://www.aopa.org/special/newsroom/stats/safety.html, 2006. Flight hours unavailable from 1943-1945. 3) Commercial Aviation Accident Rates from Air Transport Association aggregated CAB and NTSB statistics. Operating hours estimated from miles flown and average speed for 1927-1948. All air carriers operating under Part 121, including cargo. 4) Air Force Aviation Accident Rates from Air Force Safety Center – Includes UAV Accidents MIT ICAT Certification Considerations Catastrophic Accident Adverse Effect on Occupants Airplane Damage Emergency Procedures Abnormal Procedures Nuisance Normal AC 25.1309-1A Probable Improbable Extremely Improbable Consequences of Failure Change for Unmanned Operation MIT ICAT Predator Crash, Nogales, AZ From Steve Swartz, FAA UAS Program Office. 2006 CERICI Workshop. MIT ICAT Border Patrol Predator B Accident NTSB Accident #CHI06MA121 Nogales, AZ April 25, 2006, 03:41 MST Image © General Atomics Excerpts from Preliminary Report The flight was being flown from a ground control station (GCS) located at HFU. The GCS contains two nearly identical consoles, pilot payload operator (PPO)-1, and PPO-2. During a routine mission, a certified pilot controls the UAV from the PPO-1 console and the camera payload operator (typically a U.S. Border Patrol Agent) controls the camera from PPO-2. The aircraft controls (flaps, stop/feather, throttle, and speed lever) on PPO-1 and PPO-2 are identical. However, when control of the UAV is being accomplished from PPO-1, the controls at PPO-2 are used to control the camera. The pilot reported that during the flight the console at PPO-1 "locked up", prompting him to switch control of the UAV to PPO-2. Checklist procedures state that prior to switching operational control between the two consoles, the pilot must match the control positions on the new console to those on the console, which had been controlling the UAV. The pilot stated in an interview that he failed to do this. The result was that the stop/feather control in PPO-2 was in the fuel cutoff position when the switch over from PPO-1 to PPO-2 occurred. As a result, the fuel was cut off to the UAV when control was transferred to PPO-2. The pilot stated that after the switch to the other console, he noticed the UAV was not maintaining altitude but did not know why. As a result he decided to shut down the GCS so that the UAV would enter its lost link procedure, which called for the UAV to climb to 15,000 feet above mean sea level and to fly a predetermined course until contact could be established. With no engine power, the UAV continued to descend below line-of-site communications and further attempts to re-establish contact with the UAV were not successful. From: http://www.ntsb.gov/ntsb/brief.asp?ev_id=20060509X00531&key=1 MIT ICAT Predator Ground Control Station MIT ICAT Civil Aviation Applications Commercial Aircraft Fly by Wire/Light Flight Management Systems General Aviation Aircraft Very Light Jets Unmanned Air Vehicles Air Traffic Management Communication Navigation Surveillance Decision Support Integrated Air-Ground Systems ATM System Level Outer Loop Criticality MIT ICAT F lig ht S trip s Decision Aids ATC Displays Flight Plan Amendments Surveillance: Enroute: 12.0 s Terminal: 4.2 s ADS : 1s Vectors Voice AOC: Airline Operations Center Pilot ACARS (Datalink) CDU MCP Tra je cto ry C omm ands Initial Clearances Controls S tate C omm ands F lig ht M anagem ent C om p ute r Manual Control Autop ilo t Auto thrus t Aircraft S tate N av ig atio n Displays MIT ICAT Simplified Enroute Architecture Legacy Software Issues - JOVIAL Host Source: GAO/AIMD-97-30 Air Traffic Control MIT ICAT TCAS Emergent Backup MIT ICAT Bob Hilb UPS/Cargo Airline Association ADS-B GPS Backup Issue MIT ICAT 10 Year Plan FAA OEP and NGATS 20 Year Plan Multi-Agency FAA, DOD, Commerce DHS, NASA, DOT, OSTP MIT ICAT For more detail see Operational Improvement Roadmap in the Tech Hanger section of JPDO Website www.jpdo.aero Source: John Scardina JPDO MIT ICAT Software Criticality Exposure High Criticality Moderate Criticality Source: John Scardina JPDO MIT ICAT Civil Aviation Applications Commercial Aircraft Fly by Wire/Light Flight Management Systems General Aviation Aircraft Very Light Jets Unmanned Air Vehicles Air Traffic Management Communication Navigation Surveillance Decision Support Integrated Air-Ground Systems MIT ICAT GPS Wide Area Augmentation System (WAAS) •Increased Safety •Increased Efficiency and Capacity •Fuel and Time Savings •Cost Savings MIT ICAT Receiver WAAS Safety Architecture Corrections Processor Safety Processor Uplink/GEO User Satellite Signals Generates data (Level D) Monitors data (Level B) CRC protects data Weaknesses in Current System Monitor (Safety Processor) At Times Safety Processor Doesn’t Monitor Data Therefore, System Integrity Is Not Quantifiable Integrity Requirement Is No More Than One in 10 Million Chance of Hazardously Misleading Information (10 -7) MIT ICAT