Upload
phamkhue
View
213
Download
0
Embed Size (px)
Citation preview
Andrew Weinert
8 December 2016
Small UAS Well Clear
EXCOM/SARP Multi-agency collaborative sponsorship – FAAPOC: Sabrina Saunders-Hodge, UAS Integration Office Research Division (AUS-300)
DISTRIBUTION STATEMENT A. Approved for public release: distribution unlimited.
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 2AJW 8 December 2016
Legal Notices
This material is based upon work supported by the Federal Aviation Administration under Air Force Contract No. FA8721-05-C-0002 and/or FA8702-15-D-0001. Any opinions, findings, conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Federal Aviation Administration.
© 2016 Massachusetts Institute of Technology.
Delivered to the U.S. Government with Unlimited Rights, as defined in DFARS Part 252.227-7013 or 7014 (Feb 2014). Notwithstanding any copyright notice, U.S. Government rights in this work are defined by DFARS 252.227-7013 or DFARS 252.227-7014 as detailed above. Use of this work other than as specifically authorized by the U.S. Government may violate any copyrights that exist in this work.
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 3AJW 8 December 2016
• Unmanned Aircraft Systems are required to maintain well clear and avoid collisions, in particular with manned aircraft
• Quantitative well clear definition needed for the design and testing of separation systems for small UAS beyond line of sight operations
• This briefing outlines research towards a definition of well clear for small UAS for mid-term concepts of operations at low altitudes– Beyond line of sight in general use airspace– sUAS vs manned aircraft encounters– Out of scope: sUAS vs sUAS encounters and airspaces such as airport
terminals, over heliports or sporting events
Need for Quantitative“Well Clear” Definition for UAS
FAR 91.111: ...not operate so close to another aircraft as to create a collision hazard
FAR 91.113: Vigilance shall be maintained … so as to see and avoid other aircraft … pilots shall alter course to pass well clear of other air traffic
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 4AJW 8 December 2016
• Detect and Avoid (DAA) Science and Research Panel (SARP) created in 2011to coordinate DAA research– Supports multi-agency UAS Executive
Committee (ExCom)
• SARP well clear working group formedto rapidly deliver recommendation toRTCA SC-228– Lack of well clear definition identified as
highest priority research gap– R&D: August 2013 – August 2014– Multi-agency collaboration: FAA, DOD,
NASA, MITRE, MIT LL,subject matter experts
Previous Effort for Large UASQuantifying Safe Operations
+450 ft
4000 ft
Unmitigated risk threshold: P(NMAC|WCV) = 5%
Tau (time to 4000 ft) = 35 sec
Large UAS well clear adopted by FAA
NMAC = Near Mid-Air CollisionWCV = Well Clear Violation
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 5AJW 8 December 2016
• Separation standard, based oncollision risk, informed by operational acceptability
• Defined as the relative state where a desired risk threshold is achieved– Unmitigated: regardless of ownship or intruder
avoidance maneuvering– Collision avoidance action likely not needed
• Approach: define risk threshold and map to other states (range, altitude, time, …)
Well Clear as a Separation Standardfor UAS*
Relative State
Collision Risk
Acceptable Risk Threshold
well clear
Aircraft 1
Aircraft 2Relative state between aircraft
Future trajectories
Established Airspace Separation Standard MethodologyApplied to Detect and Avoid
* Applies to unmanned aircraft only
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 6AJW 8 December 2016
Large UAS Well Clear Definition Process Overview
Define formsand risk
thresholdDefine encounter
modelsTune candidate
definitionsEvaluate
operational suitability SME forum to
down selectcandidates
Make recommendation
SARP review and approval at each step
Primary Roles:
• MITRE, NM State U – Process management
• MIT LL – Monte Carlo simulations and tuning
• NASA – Human in the loop (HITL) simulations
• USAF – Stressing case analysis
• FAA – Subject matter experts (SME)
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 7AJW 8 December 2016
• Leverage lessons learned from large UAS well clear research• Approach:
– Define desired unmitigated risk = P(NMAC|Well Clear Violation)
– Explore different definition forms• Existing (Large UAS) form: mod-tau*, horizontal and vertical thresholds• “Hockey Puck” based on horizontal and vertical thresholds
– Tune parameters to risk threshold• Based on M&S using sUAS CONOPS and dynamics in operational airspace
– Consider operational suitability for sUAS operators and airspace users
Defining Well Clear for sUAS
Community Objective: Develop a sUAS vs manned aircraft Well Clear definition based on risk and operational suitability
* Time to horizontal distance thresholdNMAC – Near mid-air collision (HMD ≤ 500 ft, VMD ≤ 100 ft)
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 8AJW 8 December 2016
• Well Clear Overview
• Modeling and Simulation
– M&S Components
– Representative sUAS
– Manned Aircraft
– Simulation Environment
• Results and Recommendation
Outline
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 9AJW 8 December 2016
Modeling & Simulation ComponentssUAS vs. Manned Aircraft
Buildings
Airspace Class
Weather
Birds
People
General Aviation(Part 91)
CommercialAviation
(Part 121)
Large UAS
Study Focus
Potential Considerations
sUAS Swarm
sUAS
Helicopter Air Ambulance
Airspace Structure
Addressing challenges associated with limited sUAS flight data and
limited manned flight data at altitudes below 1200 ft. AGL
Terrain & Vegetation
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 10AJW 8 December 2016
Risk Assessment Architecture:Monte Carlo Fast Time Simulations
Development Focus
Rawradar data
Tracking and fusion Featureextraction
Collision Avoidance and Self-separation
Algorithms
Collisions per encounter
Encountermodels
Relative Risk analysis
Fast-timesimulation
Surveillance models
Aircraftflight profilesand dynamics
Collision Avoidance Safety System Tool (CASSAT)
LL Grid Computing EnvironmentCompute Nodes 274
Compute Cores 8768Peak Performance
77.1(TFLOPs)RAM (TB) 17.5
Central Storage (TB) 1,200
Distributed Storage (TB) 2,466
Network Storage
Compute Nodes Service Nodes Cluster Switch
LAN Switch
Shared File System
Scheduler
Monitoring System
Login Node
Previous Assessments
• Developed to certify TCAS version 7.1 • SARP large UAS well clear definition• Evaluation and tuning of multiple UAS
algorithms
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 11AJW 8 December 2016
ID 1 2 3 4MGTOW (lb)* 0 - 4.4 0 - 20 0 - 20 20 - 55
Mean Cruise Airspeed (kt) 25 20 30 60
Max Airspeed (kt) 40 30 60 100
Descend Rate (fpm) -300 -500 -500 -1000
Climb Rate (fpm) 500 700 700 1000
Representative sUAS55 lbs. while flying below 1200 ft AGL
ScanEagleDJI Phantom 4
Trajectories
Vertical Transit Horizontal Transit SectorSpiralCreeping Line
Aeromapper 300GoPro Karma
Notional sUASExamples
AGL – Above Ground LevelMGTOW– Max Gross Take-Off Weight
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 12AJW 8 December 2016
• Wide range of manned aircraft behavior simulated using two different probabilistic Bayesian network encounter models
• Uncorrelated fixed-wing encounter model – Represents low-altitude general aviation behavior– Since 2008, has supported a wide-range of manned
and UAS safety studies
• Helicopter air ambulance encounter model – Represents at-risk low altitude helicopter
operations– Newly developed from anonymized FOQA records
of Boston-area Medevac flights from 2015 – 2016– First helicopter focused statistical encounter model
Manned Aircraft ModelingFixed-Wing and Helicopters
MIT LL Uncorrelated Fixed Wing Encounter Model
FOQA – Flight Operational Quality Assurance Data
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 13AJW 8 December 2016
sUAS• 4 Representative Platforms• 5 Trajectory Types
Uncorrelated Manned Aircraft• 1200 code Fixed-Wing• Helicopter Air Ambulance
Airspace• Airspace Class (B,C,D,E,G)• Region (CONUS, Offshore)
Well Clear Forms• Spatial (i.e. Vertical miss distance)• Temporal (i.e. Tau)
Summary of Sensitivity Parameters
6+ million encounters simulated to derive the unmitigated risk: Sensitivity of NMAC risk to various parameters analyzed.
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 14AJW 8 December 2016
• Well Clear Overview
• Modeling and Simulation
• Results and Recommendation
– Example encounter
– Sensitivity analysis
– Risk contours
Outline
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 15AJW 8 December 2016
• Encounter representative of DJI Phantom sUAS and Cessna 150
• Manned aircraft flies 5X faster
• sUAS has minimal ability to influence outcome
• Likely well clear violation
Influence of Aircraft Speed Difference Encounter Example
Start CPA Turn
Y (ft
)
X (ft)0
-5000
-4000
-3000
-2000
-1000
0
1000
2000
3000
4000
5000
CPATime = 49 sRange = 930 ft
Manned
sUAS
StartTime = 0 sRange = 5626 ft
sUAS MannedAirspeed 15 kt 81 kt
Distance traveled from start to CPA 1177 ft 6580 ft
CPA – Closest point of approach
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 16AJW 8 December 2016
• Adding temporal criteria to HMD and VMD filter has limited influence on P(NMAC| Well Clear Violation)– Attributed to sUAS’ relative slow airspeed
• Recommend sUAS well clear definition to only use (HMD, VMD)
Sensitivity Analysis Example:Inclusion of Temporal Variables
Spatial Only
HMD – Horizontal miss distance (ft) VMD – Vertical miss distance (ft)NMAC – Near mid-air collision (HMD ≤ 500 ft, VMD ≤ 100 ft)
Spatial + TemporalVM
D (f
t)
HMD (ft)1000 2000 3000 4000
100
200
300
400
450
VMD
(ft)
HMD (ft)1000 2000 3000 4000
100
200
300
400
450
Example: 10% chance of an NMAC if
HMD = 2000 ft &VMD = 250 ft
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 17AJW 8 December 2016
• NMAC risk has limited to no sensitivity to the sUAS’ trajectory– sUAS in vertical transit has no horizontal ability to reduce risk– At slow airspeeds, sUAS simply can’t travel far during an encounter
• Unmitigated risk is not sensitive to a sUAS trajectory
Sensitivity Analysis Example:sUAS Trajectories
Vertical Transit Horizontal Transit Pattern
VMD
(ft)
HMD (ft)5001000 2000 3000 4000
100
200
300
400
450
VMD
(ft)
HMD (ft)5001000 2000 3000 4000
100
200
300
400
450
HMD – Horizontal miss distance (ft) VMD – Vertical miss distance (ft)NMAC – Near mid-air collision (HMD ≤ 500 ft, VMD ≤ 100 ft)
VMD
(ft)
HMD (ft)5001000 2000 3000 4000
100
200
300
400
450
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 18AJW 8 December 2016
• “Hockey puck” well clear definition using HMD and VMD– 250 ft. VMD x 2000 ft. HMD– ~ Half height and distance of
large UAS well clear– No temporal conditions
• Risk consistent across CONOPS, sUAS and manned aircraft dynamics models
• Subject to additional validation
• Full technical paper in progress
sUAS vs Manned AircraftSARP’s Well Clear Recommendation
VMD
(ft)
HMD (ft)
P(NMAC|WCV) (%)
500 1000 1500 2000 2500 3000 3500 4000100
150
200
250
300
350
400
450
Average risk contours over all models
SARP’s Recommendation to FAA ExCom
HMD – Horizontal miss distance (ft) VMD – Vertical miss distance (ft)NMAC – Near mid-air collision (HMD ≤ 500 ft, VMD ≤ 100 ft)
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 19AJW 8 December 2016
• Well clear definition for sUAS vs manned aircraft needed to extend operations to beyond line of sight– Definition based on unmitigated risk and operational suitability
• Risk modeling required development of new low altitude encounter models– Representative sUAS platforms and CONOPS– New model for low altitude helicopter operations
• Risk determined to not be sensitive to assumptions– Lack of sensitivity attributed to generally slow sUAS airspeeds– Additional analyses recommended to evaluate niche cases such as consistently
fast sUAS operations
• Additional community vetting of operational suitability expected
Summary
SARP’s recommended well clear definition: “hockey puck” centered on sUASHorizontal 2,000 ft, Vertical 250 ft
Lincoln Laboratory Air Traffic Control Workshop 2016sUAS Well Clear - 20AJW 8 December 2016
Thank You
Andrew WeinertAssociate Technical StaffHumanitarian Assistance and Disaster ReliefEmail: [email protected]: (781) 981-0986
Questions?
Feedback?
Rodney ColeAssistant Group LeaderSurveillance SystemsEmail: [email protected]: (781) 981-7423