safe autonomous flight environment for the notional “first ......presented to the sae / nasa...

Presented to the SAE / NASA Autonomy and Next Generation Flight Deck Symposium. Moffett Field, CA, USA. April 18-19, 2017

Safe Autonomous Flight Environment for the Notional “First/Last 50 Feet” (SAFE50) Project

Toward UAS Operations in High-DensityLow-Altitude Urban Environments

Dr. Corey A. Ippolito

Intelligent Systems Division

NASA Ames Research Center

Moffett Field, CA 94035


UAS Operations in High-DensityLow-Altitude Urban Environments

2

Unmanned Aircraft Systems (UAS) Traffic Management (UTM) concepts are advancing toward flight over populated regions.

Significant technical challenges are imposed by these environments that makes traffic management difficult, particularly for low-altitude flight in high-density urban environments.

Studies anticipate high demand

and economic growth potential

in this market.

How do you facilitate routine,

safe, and fair access to this

high-demand airspace?


Motivating Scenarios3

Safe and Regular Access for sUAS to High-Density Low-Altitude Urban Airspaces


Challenges4

• Low-altitude autonomous flight is inherently higher risk

• Mixed-use airspace

• Highly-constrained spaces within urban canyons

• High-density environment

• Other manned and unmanned airborne vehicles

• Flight near and above high-valued assets

• Cluttered wireless environment

• Hazardous ambient conditions, precipitation, and adverse winds

• Dynamic environment with significant uncertainty

• Limited size, weight, and power (SWaP)

• Regulations must establish acceptable risk posture and safety margins

• Separation assurance (SA) and collision avoidance (CA) are difficult services to provide


Risk to the vehicle

Nominal Operations Risk

Off-nominal (failure) risks

Consideration of Risks5

Dynamic Ground

Objects (DGO)

Static Ground

Objects (SGO)

Other Aircraft

Bilateral RisksUAS

Nominal Risk (e.g., TO and Landing)Off-Nominal (Failure) Risk

Stakeholders

(e.g., general public, operators, commercial

entities, insurance companies, municipalities,

certifying authorities, regulatory agencies)

Risks include potential damage, litigation, insurance costs, effects of vehicle/payload loss to businesses, etc.


SAFE50 Vehicle Autonomy Requirements

6

Dynamic Ground

Objects (DGO)

Static Ground

Objects (SGO)

Other AircraftDetect, Operate-Near, and Avoid-

Endangering SGOs

Detect, Operate-Near, and Avoid-Endangering

DGOs

Hazard Footprint Awareness, Risk Minimization/Avoidance,

Health Monitoring

Detect, Operate-Near, Avoid-Endangering

Other Aircraft

UAS

Environment

Challenges

Atmospheric

Uncertainty

Failures and

Contingencies

Degraded RF, SAT-COM, GNSS

Winds and microbursts

Avoid endangering objects in environment.

Ground

Operators and

UTM System


Challenges for UAS in Urban Environments

• Low reliability of current small UAS (high failure rates)

• Significant variability in vehicle systems and technologies on the market

• Limitations in current guidance, navigation, and control technologies

• Inability to see-and-avoid

• Limited onboard autonomy

• Limited understanding of vehicle behavior and dynamics in this environ.

• Limited onboard failure accommodation

• Insufficient communications technologies for urban environments

• Vehicle to ground, vehicle to vehicle, satellite coms, GNSS derived PNT

• Surveillance technologies are difficult to apply to this environment

• There is no common set of vehicle-level and systems-level requirements yet available for UAS in low-altitude urban flight.

7


Vehicle Autonomy

• ‘Autonomy’ broadly generalized encompasses anything that allows systems to sense, think, communicate, and react with less human intervention.

• Research literature in UAS and vehicle autonomy is extensive, covering a broad range of disciplines and techniques, and touching on all of the challenges and limitations we have identified to some degree.

• Substantial levels of private/commercial R&D investments are targeted toward advancing vehicle autonomy technology.

• While the technology is rapidly advancing, there are still severe limitations in commercially available off-the-shelf (COTS) technologies and UAS vehicle systems.

8


SAFE50 Project Goals and Approach

• Conduct an advanced study focusing on onboard vehicle-centric autonomy requirements that will allow safe, autonomous and routine sUAS access to high-density low-altitude urban environments, and integrates into the emerging UTM framework.

• Advanced study will guide the next phase for a larger systems-level study

• Develop feasible point-designs for system-level and vehicle-level concept

• Develop prototypes and demonstrate feasibility of point-design

• Assemble and develop analysis tools

• Validated high-fidelity sims, software/hardware prototypes, flight vehicles

• Analyze effectiveness of the point-design in addressing technical challenges

• Leverage UTM partnerships to track emerging trends, technologies, gaps

• Work with academia and industry towards enabling urban area access

• Peer-reviewed and competed awards, encouraging academic/commercial partnerships, see announcements at https://nspires.nasaprs.com/

9


Research Highlights

10


Dynamics Modeling and Simulation11

Credit: Tim Sandstrom, NASA Ames Research Center

Using computational fluid dynamics and wind tunnel experiments to created higher-fidelity and validated flight dynamics models.

Vehicle Testing in 7x10 ft Wind Tunnel Courtesy of Carl Russel, UTM, NASA Ames Research Center

Simulation Models


Autonomous Sensor Fusion,Environment Mapping and Hazard Characterization

12

Powerline Identification and Reconstruction. Raw LiDAR point clouds (left), voxel processing (middle),

reconstructed powerlines at 75m (right).

Environment Mapping Evaluations (LiDAR and Vision)


GNSS/GPS Denied and Degraded Environments13

Investigating integrated GNSS, LiDAR and vision for robust simultaneous localization and mapping (SLAM)

Vision-Based SLAM –

NASA NUARC Test FacilityLiDAR SLAM in NASA Disaster Assistance and Rescue Team (DART) Training Facility

LiDAR SLAM in NASA RoverScape Test Facility (collaboration with Near-Earth Autonomy, Inc.)


Natural Terrain Multi-Species Wind Modeling andEstimation under Uncertainty

14

Swarming Dragon-Eye Volcanic Plume Monitoring Project -

CFD study investigated SO2, CO2, and water vapor plume transport at anticipated emission rates

over the Turrialba Volcano in Costa Rica.


Urban Environment Wind Uncertainties15

Urban Architecture and CFD Simulation of Wind Profiles.

Presented to the SAE / NASA Autonomy and Next Generation Flight Deck Symposium. Moffett Field, CA, USA. April 18-19, 2017 16

UrbanScape Wind Uncertainties


Autonomy Payload Architecture and Prototyping

Var: objJoyPanel

DLL: joypanel.dll

m_dt_sec

m_dataset1_..

m_dataset0_..(gui)

m_dataset2_..(a/p)

m_ouputdata_..

m_data_accelXYZ_fps2

m_data_airspeed_fps

m_data_attackAngle_rad

m_data_baroAltitude_ft

m_data_eulerRPY_rad

m_data_flightPathAngle_ra

m_data_gpsAltitude_ft

m_data_headingTrue_rad

m_data_hwstate

m_data_latitude_rad

m_data_longitude_rad

m_data_pos_ENUft

m_data_rateEulerRPY_radps

m_data_ratePQR_ba_radps

m_data_sideslip_rad

m_data_vel_bafps

m_data_vel_ENUfps

m_data_velocityMagnitude_

m_dt_sec

m_in_Ml2w

m_in_velocityAngular_ENU_

m_in_velocityLinear_ENU_f

m_in_windVelocity_ENUfps

m_Ms2l

m_data_latitude_rad

m_data_longitude_rad

m_data_gpsAltitude_ft

m_data_headingTrue_rad

m_data_eulerRPY_rad[0]

m_data_eulerRPY_rad[1]

m_motor0_ucmd_0p1

m_motor1_ucmd_0p1

m_motor2_ucmd_0p1

m_motor3_ucmd_0p1

m_motor4_ucmd_0p1

m_motor5_ucmd_0p1

m_motor6_ucmd_0p1

m_motor7_ucmd_0p1

m_motor0_Speed_radps

m_motor0_AngPos_rad

m_motor1_Speed_radps

...

output_position_north_ft

output_position_east_ft

output_altitude_ft

output_velocity_mag_fps

output_velocityW_wa_fps

output_acceleration_y_fps2

output_pitchEuler_rad

output_rollEuler_rad

output_heading_True_rad

CmdAileron

CmdElevator

CmdThrottle

CmdRudder

Ml2w

output_velocity_wa_fps

output_angVelocity_wa_radps

m_in_Ml2w

m_in_velocityLinear_ENU_fps

m_in_velocityAngular_ENU_radps

Var: objOctoDLL: octocopter.dll

Var: objOctoIMUDLL: imuemulator.dll

Var: objOctoRendObjDLL: RendObj (scenevis.dll)

Ml2w

Var: _octocopter_map_H01

DLL: ProcAnimNode

Node: Helice_H01"

m_inRotation_rad[0]

Var: objJoy

DLL: directplay.dll

m_dt_sec

m_analogBuffer[]

m_buttonBuffer[]

m_sensorInput_airspeed_mps

m_sensorInput_pos_east_ENU_m

m_sensorInput_pos_north_ENU_m

m_sensorInput_pos_up_ENU_m

m_sensorInput_roll_rad

m_sensorInput_pitch_rad

m_sensorInput_heading_rad

m_out_prevWpPos_ENU_m

m_out_nextWpPos_ENU_m

m_out_nextWpSpd_mps

m_status_nWaypoints

m_status_isEngaged

m_status_currentWaypoint

m_status_timeInCommand_se

m_status_WaypointCommand

Var: objFMSDLL: fms.dll

m_inputs_stick[.. ]

m_outputs_motorCmds_m1p1[0]

m_outputs_motorCmds_m1p1[.. ]

m_outputs_motorCmds_m1p1[7]

m_sense_roll_rad

m_sense_pitch_rad

m_sense_hdg_rad

m_intrm_LMNFcmd[3]

m_intrm_aglCmd_m

m_intrm_hdgCmd_rad

m_intrm_pitchCmd_rad

m_intrm_rollCmd_rad

m_intrm_Ucmds_m1p1[4]

m_in_prevWpPos_ENU_m

m_in_nextWpPos_ENU_m

m_in_nextWpSpd_mps

Var: objAPDLL: octocontrol.dll

Var: objOctoINSDLL: octoins.dll

Var: (sensor emulator(s))

Var: (environment model(s))

Var: (Integrated INS solution)

Var: (Taemin s combinator)










Mca

Lcmd

PID LcmdferrS

fest

frefRoll Cmd

Ref Filter

(2nd order)

Roll

Cmd

Limiter

Lateral Modes

LATMODE_STICKCMD

LATMODE_ROLLZERO

LATMODE_ROLLSTICKCMD

LATMODE_ROLLCMD

LATMODE_GNDSPDHOLD

LATMODE_TRACKTOWP

fcmd_hold

fcmd_gndspd

Mwa2lhv

Vg

ndre

f_lh

v[2

]

Vgndcmd_wp_wa[2]

S

Vgnd _lhv_est[2]

Verr[2]

Roll to Moment (L)

GndSpd to Roll

PID

Vgnd Cmd

Ref Filter

(2nd order)

VGnd

Cmd

Limiter

Vg

ndre

f_w

a[2

]

stickLat_m1p1K

L Cmd

Limiter

Mcm dPID McmdqerrS

qest

qrefPitch Cmd

Ref Filter

(2nd order)

Pitch

Cmd

Limiter

qcmd_hold

qcmd_gndspdS

Vgnd _lhv_est[2]

Verr[2]

Pitch to Moment (M)GndSpd to Pitch

PID

Stick Cmd Filter

(1st order)KMstickcmd

M Cmd

Limiter

Vgndref_lhv_y

Vgndref_lhv_x

Longitudinal Modes

LONMODE_STICKCMD

LONMODE_PITCHCMD

LONMODE_GNDSPDHOLD

Heading Modes

HDGMODE_STICK_N_CMD

HDGMODE_HDG_ZERO

HDGMODE_STICK_HDG_CMD

HDGMODE_HDG_CMD

Stick Cmd Filter

(1st order)KNcmd

Ncmd

N Cmd

Limiter

PIDyerr

yest

Yaw to Moment (N)

yrefHdg Cmd

Ref Filter

(2nd order)

Ncmd

Hdg

Cmd

Limiter

Yaw

Error

Control Allocation

(u=Mca*[LMNFz] )

Fzcmd

U[1..8]

Lref

Nref

Mref

stickLon_m1p1

StickRdr_m1p1

Stick to L

Kst icklat2L

Kst lon

Stick to Rate

Stick to Rate

KstickRdr2N

Lcmd

Thrust Modes

THRMODE_STICK_THR_CMD

THRMODE_VSPD_CMD

THRMODE_ALT_AGL_HOLD

THRMODE_ALT_MSL_HOLD

hcmd_agl

Alt Cmd Ref

Filter

(2nd order)

hcmd_msl

Alt.

Cmd

Limiter

href

Altitude Error

to Thrust

h_agl_est

S

MSL/AGL

h_msl_est

MSL/AGL

herr

PIDVert

Speed

Limiter

dvspdS

vspdref

vspdcmd

ALT/VSPD

PID

VSPD to FZ

vspdref

S

Fzcmd

StickThr_m1p1

Fz Cmd

Limiter

(0,+1)

m_motorCmds[8]

m_inputGearCmd (0,1) m_outGearServoCmd_pwmConvert

m1p1 to

pwm

Top: LATMODE_STICKCMD

Bot: Others

Rate

Limiter

w/(s+w);

M1p1

Limiter ;

PWM

convert

Normalize

(-1,+1)

Rate

Limiter

w/(s+w)

Kstthr

y=(x+1)/2

-1

Note: +stick fwd

means -pitch down

Note: +stick right

means +roll ri ght

Note: +stick is

rudder right,

+rotat ion about Z

U0

U1

U2

U3

U4

U5

U6

U7

=

-S L N T

-L S -N T

-L -S N T

-S -L -N T

S -L N T

L -S -N T

L S N T

S L -N T

L

M

N

Z

Control Allocation Matrix

fCmd=0

fCmd

rollCmdLimit_rad

qCmd=0

Lim it range from -

PI to +P I

safeHoldAlt_m_agl

AGL

LimiterAGLCmdLim itM in_m

AGLCmdLim itMax_m

PID

AGL to Fz

aglErrS

aglCmd fzcmd

fzCmdM in_0p1

Ksticklat2Roll

KfCmdStick

m_int rm .

hdgCmd_rad

StickRdr_m1p1

m_params.

Kst ickRdr2Hdg

Stick to HDG

K

m_inputs.headingCmd_rad

m_inputs.prevWP_ENU

m_inputs.nextWP_ENU

Calculate

Track-To

XtrackErr

xtrackerr Xtrack to

Roll

rollcmd

A/C Position

LATMODE_TRACKTOW PLATMODE_GNDSPD

Calculate

Track-To

HdgCmd

Experimental Multicopter Flight Management

and Flight Control System

Real-Time Embedded Software Architecture

Flight Testing - NASA SAFE50 and UTM Flight Test - August 2015

Autonomy Architecture

Systems Analysis, CAD and Hardware Design

Integrated Payload/Vehicle Test Platform


Flight System Interfaces andGround Control System Development

18

Hardware-In-The-Loop Vehicle Simulation Configuration

Ground Control Stations Custom GCS Control InterfacesMulti-Vehicle Simulation Integration -

Airspace Operations Laboratory (AOL)



Safe Autonomous Flight Environment for the Notional “First/Last 50 Feet” (SAFE50) Project

Toward UAS Operations in High-DensityLow-Altitude Urban Environments

Dr. Corey A. Ippolito

Intelligent Systems Division


Moffett Field, CA 94035

safe autonomous flight environment for the notional “first ......presented to the sae / nasa...

Documents