Formal Methods Challenge: Efficient Reconfigurable Cockpit Design and Fleet Operations using Software Intensive, Networked, and Wireless-Enabled Architecture (ECON) Kristin Yvonne Rozier University of Cincinnati Dagstuhl Seminar on Formal Foundations for Networking February 10, 2015
Overview Net-enabled aircraft are being designed now! Goal: reduce the cost of aircraft by migrating towards software and net-enabled/cloud-based architecture and capabilities Method: reduce aircraft weight increase automation move from hardware to software move from aircraft-based systems to fleet-based systems reduce maintenance
Design Paradigm Increase: software networking digital communications automation for cockpit systems development aircraft avionics advances ground-to-air interaction Decrease cost: of operations of maintenance and inventory of mechanical parts of customizing and certifying new aircraft cockpit designs wiring control systems flight management algorithms How will we automatically verify these designs?
Overview Net-Enabled Aircraft Design Current Project Status Join the Team! Challenge 1: Cockpit Design Costs and Complexity Cockpit: highly complex hybrid system many heavy mechanical parts software hardware instrumentation control systems flight management system interfaces trip-switches other sub-systems customized for every aircraft type Temporal Logic Kristin Yvonne Rozier Boeing 787 Cockpit: wheel column alone weighs over 400 lbs Challenge: Net-Enabled Aircraft
Overview Net-Enabled Aircraft Design Current Project Status Join the Team! Challenge 1: Cockpit Design Costs and Complexity Cockpit: highly complex hybrid system many heavy mechanical parts software hardware instrumentation control systems flight management system interfaces trip-switches other sub-systems customized for every aircraft type Boeing 787 Cockpit: wheel column alone weighs over 400 lbs We have not solved the formal verification problem for current cockpits. Temporal Logic Kristin Yvonne Rozier Challenge: Net-Enabled Aircraft
Overview Net-Enabled Aircraft Design Current Project Status Join the Team! Challenge 1: Cockpit Design Costs and Complexity Cockpit: highly complex hybrid system many heavy mechanical parts software hardware instrumentation control systems flight management system interfaces trip-switches other sub-systems customized for every aircraft type Boeing 787 Cockpit: wheel column alone weighs over 400 lbs We have not solved the formal verification problem for current cockpits. How do we scale to net-enabled cockpits? Temporal Logic Kristin Yvonne Rozier Challenge: Net-Enabled Aircraft
Challenge 2: Electrical Wiring Weight and Complexity Replace wires with wireless fly-by-wire better than mechanical control wiring is both complex and heavy wiring is customized for every aircraft type wiring must be manually fitted for every aircraft How will we verify these hybrid wired/wireless networks?
A380-800 has about 100,000 wires, 470 km long, 5700kg of weight + additional 30% weight for wiring harnesses
Cost of Aircraft Weight
Challenge 3: Software Systems for Each Aircraft and Increasing Management Cost Now: flight controls, trajectory, and information management systems on every plane fleet management limited to resources on each aircraft software development, certification, and installation billed per aircraft Future: move hardware to software easier to upgrade/maintain move local software to cloud or network of same-type aircraft decrease fleet inefficiency/down-time per aircraft for upgrades software for a fleet, not for a plane easier fleet management
Challenge 3: Software Systems for Each Aircraft and Increasing Management Cost Now: flight controls, trajectory, and information management systems on every plane fleet management limited to resources on each aircraft software development, certification, and installation billed per aircraft Future: move hardware to software easier to upgrade/maintain move local software to cloud or network of same-type aircraft decrease fleet inefficiency/down-time per aircraft for upgrades software for a fleet, not for a plane easier fleet management Harder to formally verify...
What can we take off before we take off? A350/A380 cockpits on-board information management systems: how much can be moved?
Overview Net-Enabled Aircraft Design Current Project Status Challenge 4: Aircraft-Centric Operations Networked aircraft optimal planning more direct flights less fuel, crew fatigue, time Temporal Logic Kristin Yvonne Rozier Challenge: Net-Enabled Aircraft Join the Team!
Previous Work: V&V of Automated Air Traffic Control System Designs 1 2 3 4 1 A. Cimatti, M. Gario, C. Mattarei, K.Y. Rozier, and S. Tonetta. Comparing Automated Air Traffic Control Designs via Formal Safety Assessment. Under submission: 2015. 2 Zhao, Y. and Rozier, K.Y. Formal Specification and Verification of a Coordination Protocol for an Automated Air Traffic Control System. Science of Computer Programming, v-96 #3, 2014. 3 Zhao, Y. and Rozier, K.Y. Probabilistic Model Checking for Comparative Analysis of Automated Air Traffic Control Systems. IEEE/ACM ICCAD 2014. 4 Zhao, Y. and Rozier, K.Y. Formal Specification and Verification of a Coordination Protocol for an Automated Air Traffic Control System. AVoCS-2012.
Challenge 5: Must be safer! Real-time and Redundant improved intra-aircraft networks improved inter-aircraft/ground network technology Need real-time information about critical parts! Need redundant/back-up systems!
Current Project Status Join the Team! Previous Work: Real-time System Health Management for Intelligent, Autonomous UAS 5 Swift UAS Flight Computer... Laser Altimeter rtr2u2 Common Bus Interface Health Model (BN) Specification (ϕ) Higher Level Reasoning en {ϕ1,.., ϕn } Runtime Observers 6 7 health estimation Net-Enabled Aircraft Design system status Overview event updates Baro Altimeter... Radio Link IMU & GPS Event Capture & RTC 5 J. Geist, K.Y. Rozier, and J. Schumann. Runtime Observer Pairs and Bayesian Network Reasoners On-board FPGAs: Flight-Certifiable System Health Management for Embedded Systems. RV-2014. 6 T. Reinbacher, K.Y. Rozier, and J. Schumann. Temporal-Logic Based Runtime Observer Pairs for System Health Management of Real-Time Systems. TACAS-2014. 7 J. Schumann, K.Y. Rozier, T. Reinbacher, O.J. Mengshoel, T. Mbaya, and C. Ippolito. Towards Real-time, On-board, Hardware-supported Sensor and Software Health Management for Unmanned Aerial Systems. PHM-2013. Temporal Logic Kristin Yvonne Rozier Challenge: Net-Enabled Aircraft
ARMD Seedling Phase I designing for increased software operations networked/cloud-based systems for individual aircraft + fleet operations wired wireless modularity real-time connectivity: faster upgrades, lower maintenance
Need to Develop: New Cockpit Design wireless software-enabled controls digital, continuous, reconfigurable displays interaction with runtime monitors? cloud-controlled operations? NASA and partners are looking to design initial prototypes this year! Ex. Goal: reduce at least 1 ton of weight/aircraft
New Cockpit: Can We Move These to Software? Wheel column, yoke, & back drive (Boeing); joystick (Airbus) Thrust levers for propulsion control Brake pedals/levers Rudder pedals Display and flight management system interactions Trip switches, knobs, controls Flap setting, spoiler levers Landing gear controls
Overview Net-Enabled Aircraft Design Current Project Status Need to Develop: New Fleet Operations Design aircraft as networks network of aircraft cloud architecture supporting fleet operations Temporal Logic Kristin Yvonne Rozier Challenge: Net-Enabled Aircraft Join the Team!
New Fleet Architecture: Can We Move These to the Cloud? trajectory planning/re-planning fuel optimality weather traffic scheduling connections: passengers, crew, aircraft emergency assistance What are the constraints? What flight management systems cannot be moved to the cloud?
Need to Formally Reason About: Runtime monitoring/real-time system health management Integrity, reliability, latency of communications Security, encryption, trustworthiness of data Network mobility, software-defined networking Fault-tolerant networking Cloud resource availability, security, aircraft synchronization Redundancy & back-up systems
ECON Design Team, by group 4 NASA Aeronautics Centers & JPL 14 Industry Partners 6 Academic Institutions 1 Formal Methods expert so far...
ECON Design Team, by group NASA Aeronautics Centers (ARC, AFRC, GRC, LaRC), & JPL PI: Parimal Kopardekar, NASA Ames Research Center 14 Industry Partners Aurora, Boeing, CAFÉ Foundation, FedEx, GE, Gulfstream, Harris Corp, M2C Aerospace, NextGen AeroSciences, Nissan, Rockwell Collins, Sensurion/United, Terrafugia, Verizon, outside SMEs Additional interest: CISCO, Northrup Grumman 6 Academic Institutions U. Cincinnati, MIT, Georgia Tech, Penn State, U. Massachusetts, U. Colorado 1 Formal Methods expert so far...
Conclusion ECON is happening now Formal methods involved from initial design time How do we meet this challenge? What restrictions do we need to make to enable FM? How do we rise to the design-analysis challenge? Runtime verification and integration into the cockpit too Future UAS applications Join the team! rozierky@uc.edu; parimal.h.kopardekar@nasa.gov