33
Urban Aerial Mobility: The Cyber‐Physical Systems Approach Yan Wan Associate Professor Electrical Engineering University of Texas at Arlington @NCTCOG UAS Safety and Integration Task Force Meeting October 30, 2019 1

Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

  • Upload
    others

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Urban Aerial Mobility: The Cyber‐Physical Systems Approach 

Yan WanAssociate ProfessorElectrical Engineering

University of Texas at Arlington

@NCTCOG UAS Safety and Integration Task Force MeetingOctober 30, 2019

1

Page 2: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

About Me• 2004‐2009, worked on NASA project on strategic air traffic management • 2010‐2015, worked on MITRE project (subcontract from FAA) on strategic air traffic management

Snapshot of FACET, NASA MITRE’s FCM Decision Support Tool

Page 3: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

About Me• 2010‐Present, research on unmanned aerial vehicles (UAVs) and unmanned systems funded by national agencies (NSF, NIST, ONR, ARO) and industries (Lockheed, Ford, MITRE) totaling over $6M• NSF Faculty Early Career Development (CAREER) Award

• Tech Titan of the Future – University Level award, Metroplex Technology Business Council

• Best Demonstration Award, US Ignite• Third Prize GENI Demonstration Award

Page 4: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

UTA Research Institute

• UTARI bridges the gap between academic research and product development to the point ofcommercialization

• UTARI collaborates with business and industry to add value to their enterprises by developingtechnology that benefits society

• UTARI provides hands‐on research experiences to undergraduate and graduate students, givingthem exposure to real‐world applications

Biomedical Technologies Institute for Predictive Performance Methodology Automation & Intelligent Systems

Dr. Eileen ClementsDirector of Research

Page 5: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

IoT and Smart City Applications of UAVs

Precision Agriculture

Aerial Taxi

Cargo Transport

Sports Coverage

Emergency Response

Personal Assistance

Traffic Surveillance

Land Survey

Infrastructure Health Monitoring

Many Others…

UAVs are becoming aerial robots with integrated sensing, communication, control, computing, and networking capabilities

Page 6: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Sample Efforts from My Group: Smart Emergency Response System (SERS)

• Smart Emergency Response System, our collaborative efforts with Boeing, MathWorks, National Instruments, NCSU, U of Washington, WPI, and MIT Lincoln lab, 2013‐2017

4

Smart Emergency Response System, our collaborative efforts with Boeing, MathWorks, National Instruments, NCSU, U of Washington, WPI, and MIT Lincoln lab, 2013‐2017https://www.youtube.com/watch?v=Yi_dK4iRCA4&t=34s

Page 7: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Sample Efforts from My Groups: ARFL Search and AI Challenge 

AFRL Search and AI Challenge, 1st place in one competition run, and among the only two teams who scored top five in all competition runs, 2019

https://www.youtube.com/watch?v=fOq56R7DxDk

5

Page 8: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Market of Urban Aerial Mobility

UAV market is expected to grow US$ 51.85 billion by 2025 from US$ 11.45 billion in 2016.

7

Page 9: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Market of Urban Aerial Mobility

UAV market is expected to grow US$ 51.85 billion by 2025 from US$ 11.45 billion in 2016.

Page 10: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

• Traditional air traffic management (ATM) is concerned with multiple spatiotemporal scales:  Individual aircraft GNC Air traffic control  air traffic flow management Airspace management

Ensure Safety and EffiencyUrban Aerial Mobility (UAM) & UAV Traffic Management (UTM)

10• Yan Wan, Christine Taylor, Sandip Roy, Craig Wanke, and Yi Zhou, "Dynamical queueing network model for flow contingency management", IEEE Transactions on Intelligent 

Transportation Systems, vol. 14, no. 3, pp. 1380‐1392, September 2013.• C. Taylor, C. Wanke, Y. Wan, and S. Roy "A Framework for Flow Contingency Management," AIAA Aviation Technology, Integration, and Operations (ATIO) Conference, 2011.

Snapshot from FACET, NASA

More uncertainty

More heterogeneity

More diverse service providers

Page 11: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Technical and Operational Challenges• Safety is difficult to achieve in a dense airspace of aerial vehicles

• Uncertain and flexible UAV trajectories• Heterogeneity

• Aerial vehicle types: types (weight, maneuvering), environment (urban, rural), missions, users• Incomplete information

• Constrained payload for UAS platforms• Limited capability of sensors to reach out for sense‐and‐avoid• With sensors on board, where does the computing take place?

• Complicated uncertain environment• Powerlines, people and properties on the ground• Uncertain weather spread• Intelligent crash?

• Unstable communication and control links, GPS• Communication and control co‐design

• Levels of responsibilities• On‐board pilot, drone operator, traffic managers, etc. 

• Stakeholder roles: profit‐driven industries vs. the safety of NAS• Should the market self‐adapt or should organization be provided?• Lack of certifications: levels of certifications, certification of components, need/risk‐based certifications

• Security• Jammer, secure communication link with respect to spectrum and performance requirement

• Societal acceptance• Environmental impact, e.g., noise and pollution• Privacy• Ownership of the low‐amplitude space

11

NSF: Airborne Computing Networkshttp://www.uta.edu/utari/research/robotics/a

irborne/index.php

NSF: Safe Task‐Aware Autonomous Resilient Systems (STAARS)https://nsf.gov/awardsearch/showAward?AWD_ID=1724248&HistoricalAwards=false

NSF CAREER: Co‐Design of Networking and Decentralized Control to Enable Aerial Networks in an Uncertain Airspacehttp://www.uta.edu/faculty/wanlab/career.html

Page 12: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

• FAA 107 Rule (released in August 2016)• UAS less than 55 lbs• Remote pilot certificate• Visual line of sight, daylight or civil wrilight operations• Yield right‐of‐way to manned aircraft• One UAS per remote pilot• No operations over people• 400 feet or below• Waivable provisions

• Six UAV test sites authorized in 2012 and reauthorized in 2018 till 2023

• Alaska, Nevada, New York, North Dakota, Texas (Lone Star UAS Test Center), Virginia

• 10 FAA UAS Integration Pilot Program aims to accelerate safe UAS integration launched in 2017

12

Current Advances of UAV Traffic Management (UTM)

• NASA Technology Capability Levels (TCL) https://utm.arc.nasa.gov/

• TCL1 (August 2015): geofencing, altitude "rules of the road" and scheduling of vehicle trajectories

• TCL 2 (October 2016): beyond visual line‐of‐sight operations in sparsely populated areas

• TCL 3 (January 2018): maintain safe spacing between cooperative (responsive) and non‐cooperative (non‐responsive) UAS over moderately populated areas

• TCL 4 (ongoing 2019): UAS operations in higher‐density urban areas for tasks such as news gathering and package delivery. It will also test technologies that could be used to manage large‐scale contingencies

• Lone Star 690 planned UAS stories over 10 days (August 12‐16 and 19‐23) 

Page 13: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Technology Development Opportunities:UAM and UTM Focus Areas

• Human‐Systems• Reduce the probability of human commanding errors• Improve the situation awareness of the operator • Increase automation on the vehicle such that the operators tasks are a manage‐by‐exception • Enable decision support tools for emergency situations and contingency management • Enable a single operator to command and control multiple vehicles 

• UAS vehicle and ground support automation • Provide onboard and/or ground‐based separation assurance from other airborne traffic, terrain and natural obstacles, man‐made obstacles, and people on the 

ground.• Fault tolerant systems to reduce the risk in emergency situations (lost‐link, hardware failure, etc.) • Path planning in complex environments (GPS‐denied environments, variable weather conditions and obstructions, man‐made structure and terrain avoidance, etc.) • vehicle health monitoring and diagnostics. 

• Capacity and Airspace Management • Spectrum allocation and management• Airspace management system health monitoring • Flight monitoring and conformance monitoring • Flight planning, scheduling and demand management and separation assurance • Contingency Management • Providing information to various communities that are connected to the airspace (other ATM systems, UAS operators, general aviation community, public, law 

enforcement, etc.) 

• Mission planning and contingency management • Risk‐based operational planning and contingency management • Using vehicle performance modeling to determine operation feasibility and contingencies 

13

Page 14: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Opportunities: New Businesses• Application end users

• Customized solution design

• Platform• Passenger Aerial Vehicle • Air Taxi • Cargo Aerial Vehicle • Air Ambulance • Air Metro • Air Shuttles • DroNav• Last Mile Delivery • Other Platforms

• Components• Power solutions• Sensors• Communication• Control• Autopilot

14

Page 15: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Opportunities: New Businesses• Infrastructure

• Charging Stations • Traffic Management • Maintenance Facilities • Vertiports

• Data and Service Provider• Testing• Certification• Weather data• Other awareness data• Collision and avoidance solution

• Training and Education• Licensing• Pilot Training

• Legal Issues• Insurance • Lawsuit 

15

Page 16: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

My Group’s Effort: Cyber‐Physical Systems Co‐Design for UAV Networks and UTM

Vehicle/V2V Communication• Challenge: UAVs are controlled wirelessly. Communication and control are tied together.• Highlight: Communication/control co‐design for long‐distance UAV communication

Vehicle Sensing and Computation• Challenge: UAVs have stringent physical constraints for resources such as sensing, power, communication, and computation.• Highlight: Balance between on‐ and off‐board sensing, and sensing/UAV mission co‐design

Multi‐Vehicle Coordination• Challenge: UAV types, trajectories, dynamics, and missions are heterogeneous.• Highlight: Graphical games, distributed reinforcement learning, and Bayesian beliefs

From Local‐Scale Autonomy to Large‐Scale Capacity Management• Challenge: UAVs provide flexible on‐demand services, different from traditional air traffic which are mostly flight plan based. • Highlight: CPS approach to connect local autonomy with global capacity

Contingency Management • Challenge: UAVs are safety critical systems sensitive to uncertain environmental impact.• Highlight: Online and offline decision co‐design driven by probabilistic spatiotemporal scenario 

Page 17: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Communication and Control Co‐design for Long‐distance and Broad‐band UAV Networking• UAVs to provide long‐distance broad‐band on‐demand 

emergency communication• The control of directional antennas facilitates communication• Received signal strength, the communication indicator, serves 

as measurement and goal function for control• Communication measurement data learns the environmental‐

specific communication model, and distributed reinforcement learning is used for adaptive optimal control.

• Flight tests, water‐proof design, and user‐friendly interface design for technology transfer in the safety‐critical emergency response application. 

S. Li, M. Liu, C. He, Y. Wan, Y. Gu, J. Xie, S. Fu, and K. Lu, “The Design and Implementation of Aerial Communication Using Directional Antennas: Learning Control in Unknown Communication Environment,” IET Control Theory and Application, accepted, October 2018.J. Chen, J. Xie, Y. Gu, S. Li, S. Fu, Y. Wan, and K. Lu, “Long-Range and Broadband Aerial Networking using Directional Antenna (ANDA): Design and Implementation, IEEE Transactions on Vehicular Technology, Vol. 66, No. 12, pp. 10793-10805, December 2017.

Page 18: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Disaster Drills

S. Li, M. Liu, C. He, Y. Wan, Y. Gu, J. Xie, S. Fu, and K. Lu, “The Design and Implementation of Aerial Communication Using Directional Antennas: Learning Control in Unknown Communication Environment,” IET Control Theory and Application, accepted, October 2018.J. Chen, J. Xie, Y. Gu, S. Li, S. Fu, Y. Wan, and K. Lu, “Long-Range and Broadband Aerial Networking using Directional Antenna (ANDA): Design and Implementation, IEEE Transactions on Vehicular Technology, Vol. 66, No. 12, pp. 10793-10805, December 2017.

(a) (b)

(c) (d)

• May 2014, With Austin Fire Department and WPI UAV in coordination with robot for S&R. https://www.youtube.com/watch?v=ToPw9bEObHA

• February 2016, with Tarrant County Fire Service Training Center, Filmed by Discovery Canada. https://youtu.be/Ujf1RWRsj_U

• May 2016, Denton Full‐scale Disaster Drill https://www.youtube.com/watch?v=dUXGG3iR6Wc&t=4s

• May 2017, Disaster drill at the UNT Campus

Click for video Click for video

Click for video Click for video

Page 19: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Application built on the Aerial Communication using Directional Antennas (ACDA) SystemBeyond Visual Line of Sight Control  Multi‐UAV Formation

S. Li, Y. Gu, B. Subedi, C. He, Y. Wan, Atsuko Miyaji, and Teruo Higashino, “Beyond Visual Line of Sight UAV Control for Remote Monitoring using Directional Antennas,” accepted byIEEE GLOBECOM 2019 Workshop on Computing-Centric Drone Networks, Waikoloa, Hawaii, December 2019.

19

Page 20: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

• UAV path planning can be significantly affected by wind disturbances.

• An integrated model of the communication model (802.18) that transmits wind map data, vehicle dynamics, and environmental impact to quantify the benefit of weather. service to UAV path planning and its configuration. 

2. UAV Weather Service and On/Off‐board Sensing

M. Pinhero, M. Liu, Y. Wan and A. Dogan,  “On the Analysis of On‐board Sensing and Off‐board Sensing through Wireless Communication for UAV Path Planning in Wind Fields,” in Proceedings of AIAA Schitech, January 2019. 

Dr. Matthias SteinerDirector Aviation Applications Program

Research Applications Laboratory (RAL)National Center for Atmospheric Research

(NCAR)

Wind sensing

On‐board wind sensors(Limited range)

Off‐board weather service(Constrained by communication 

bandwidth and channel performance)

Page 21: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

3. Differential Graphical Games for Multiple‐Vehicle Coordination• Consensus Control and Distributed Optimal Control

• Consensus control does not  generally produce optimal solution. • Distributed optimal control does not assume the existence of self interests. 

• Graphical Games: allows each  agent  to optimize its  own  performance  index• Standard games utilizes  static  and  immediate costs. • Differential multi‐agent games: Game theory integrated with optimal control theory to 

develop feedback controls.• Players in a system cannot obtain the complete system state information, and they have to 

make their decisions based on limited sensing capabilities. • In differential graphical games, players are connected by a communication graph, which 

represents the information flow, and each player aims to make their decisions based on its own and its neighbors’ state information.

• Our directions• Epistemic games• Stochastic games• New game with distributed Nash solutions

Agents’ dynamics𝑥 𝐴𝑥 𝐵𝑢

𝑥 𝐴𝑥Tracking error dynamics

𝛿 𝑎 𝑥 𝑥 𝑔 𝑥 𝑥∈

𝛿 𝑎 𝑥 𝑥 𝑔 𝑥 𝑥∈

𝐴𝛿 𝑑 𝑔 𝐵𝑢 𝑎 𝐵∈

𝑢

Cost function

𝐽 𝑟 𝛿 , 𝛿 , 𝑢 , 𝑢 𝑑𝑡

Value function

𝑉 𝑡 𝑟 𝛿 , 𝛿 , 𝑢 , 𝑢 𝑑𝜏

Where𝑢 is a function of 𝛿𝛿 𝛿 , 𝑗 ∈ 𝑁

𝑢 𝑢 , 𝑗 ∈ 𝑁 21

Page 22: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

• Reinforcement learning to achieve adaptive optimal solution• Multi‐agent graphical games to capture the cooperative and 

adversarial UAV networks.

• Reinforcement learning is to used to solve the BHJI equation.

• Dealing with Incomplete information • To capture with heterogeneous vehicle intentions, we use 

Bayesian Graphical games, where the agents are uncertain about their payoff and must employ the evidence that their environment provides to improve their estimation of the real setting of the game. 

• All agents play their best response strategies to achieve the so‐called Bayes‐Nash equilibrium. The agents also update their beliefs using information obtained from the graph topology.

Epistemic Game

SystemAgent

ActorControl Policy

CriticExpected Value

Belief Update

Neighborsstates

.V. G. Lopez, Y. Wan, and F. L. Lewis, “Bayesian Graphical Games for Synchronization in Dynamical Networks,” IEEE Transactions on Control of Network Systems, accepted, September 2018.

V. G. Lopez, F. L. Lewis, Y. Wan, E. N. Sanchez, and L. Fan, “Solutions for Multiagent Pursuit‐Evasion Games on Communication Graphs: Finite‐Time Capture and Asymptotic Behaviors,” IEEE Transactions on Automatic Control, accepted, January, 2019.

12( (t), , ) ( )

i

T T Ti i i i i ii i i ii i j ij j

j Nt

V u u Q u R u u R u dt

2 1 2 1 11 1 12 2 2( ) ( ) 0,

i

TT Tj jc T T Ti i i

i i ii i i i i ii i j j j jj ij jj ji i i j jj N

V VV V VA Q d g B R B d g B R R R B i N

( , , , ) ( | ) ( , , ),i i i i i i i i iiEJ u u p J u u

Page 23: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

• Stochastic Games• MAS often operate in uncertain environments.

• Uncertain intentions, if exploited, can assist tracking and predictive control.

• Uncertain intentions can be captured using random switching models.

• N‐Player Nonzero‐Sum Game

• Developed scalable uncertainty evaluation method (MPCM‐OFFD) to evaluate the cost and design optimal solutions

Stochastic Games

.M. Liu, Y. Wan, F. Lewis, V. G. Lopez Mejia, "Stochastic Two‐Player Zero‐Sum Learning Differential Games," in Proceedings of the 15th IEEE International Conference on Control & Automation (ICCA), Edinburgh, Scotland, June 2019.

M. Liu, Y. Wan, F. Lewis, and V. Lopez, "Adaptive Optimal Control for Stochastic Multi‐Agent Differential Games using On‐Policy and Off‐Policy Reinforcement Learning," submitted to IEEE TNNL, Jan, 2019.

System dynamics: 𝒙 𝑨 𝒂 𝒙 ∑ 𝑩 𝒖

𝒂 𝒂 𝑡 ∈ 𝓡 is 𝑚‐dimensional time‐varying uncertain vector.

Expected cost:  𝐽 𝒙 0 , 𝒖 , 𝒖 𝐸 𝑟 𝒙, 𝒖 , 𝒖 𝑑𝑡

𝐸 𝒙𝑻𝑸𝒊𝒙 𝒖 𝑹𝒊 𝒖 𝑑𝑡

Value function: 𝑉 𝒙 𝑡 𝐸 𝒙𝑻𝑸𝒊𝒙 ∑ 𝒖 𝑹𝒊 𝒖 𝑑𝜏

Optimal control policy:  𝒖 ∗ argm𝑖𝑛𝒖

𝐽 𝒙 0 , 𝒖 , 𝒖

23

Page 24: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

• Existing solutions

• New Game that guarantees distributed global Nash solution • We propose a game, with extra terms in the objective to decouple the HJ equation

𝐽 ∑ 𝛿 𝑄 𝛿 𝑢 𝑅 𝑢 𝑎 𝑢 𝑅 𝑢∈ 𝑑𝑡, where 𝛿 𝛿 , 𝛿 , and 𝑄𝑄 , 𝑄 ; 𝑄 , 𝑄 . 𝑅 and 𝑄 are symmetric matrices, and 𝑅 0, 𝑄 0.

New Game with Distributed Global Nash Solution

• Best Responses: The solution derived from best responses can constitute Nash, but is generally not distributed. 

• 𝛿 𝑄 𝛿 𝛻𝑉 𝐴𝛿 𝛻𝑉 𝐵𝑅 𝐵 𝛻𝑉∑ 𝑎 𝑑 𝑔 𝛻𝑉 𝐵𝑅 𝐵 𝛻𝑉∈

∑ 𝑎 𝑑 𝑔 𝛻𝑉 𝐵𝑅 𝑅 𝑅 𝐵 𝛻𝑉∈ 0• There does not generally exist a distributed 

quadratic value function 𝑉 𝛿 that solves the HJ equation to provide a global Nash equilibrium solution, where each agent only uses the state information of its own and its neighbors. 

• Minmax strategies: The solution derived from minmax strategies is distributed, but is generally not Nash.• Agent i prepares its policy by assuming

that the goals of its neighbors are to maximize 𝐽 to oppose it, i.e.,

𝑢∗ arg min max 𝐽 𝛿 , 𝑢 , 𝑢

• We prove that the solution found by min‐max strategies cannot constitute global Nash equilibrium, i.e., 𝑢 𝑢∗ or 𝑢 𝑢∗ .

Page 25: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

• To support flexible on‐demand UAV applications, UAS movement patterns can be highly random, flexible, and uncertain.

• Understand airspace capability bound under highly random UAV mobility.

• Random mobility equipped with the only safety constraint: local sense‐avoid protocol  

• Under flexible UAV operations, understand the relationship between local autonomy to maintain safety and global capacity. • S&S has the negative impact • S&R is the most effective, however the effectiveness is 

affected by the number of vehicles• Sensing distance: extending sensing distance results in 

less collision probability and larger airspace capacity.• Travel time: longer travel time leads to less collision 

probability, and larger airspace capacity 

4. The Connection Between Highly Flexible Local Autonomy and Global Capacity

Highway in the air                     Flexible mobility

coordinate in x axiscoordinate in y axis

0100 100

50 50

10-4

1

0 00

0.2

0.4

0.6

0.8

110-4

When a pair of UASs are within the sensing distance, i.e., 𝑫𝒊,𝒋𝒅𝒐, the vehicle to the relative right stops, and the other vehicle maintains the original RD RMM until 𝐷 , 𝑑 .

With S&SWitouth S&S

0 50 100 150

Distance (m)

0

0.005

0.01

0.015

0.02

0.025

0.03

pdf

simulated pdfupper bound for part (B)lower bound for part (B)

0 50 100 150Distance (m)

0

0.01

0.02

0.03

0.04

pdf

Simulated pdfTheoretical pdf

Page 26: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

5. Probabilistic Spatiotemporal Scenario Data Driven Online and Offline Decision

• Stochastic automaton for spatiotemporal weather spread

• Spatiotemporal scenario data similarity quantification and query

• Scalable uncertainty quantification and decision under uncertainty

• Path planning of multiple UAVs subject to collision avoidance 

Probabilistic weather ensembles

J. Xie, A. R. Kothapally, Y. Wan, C. He, C. Taylor, C. Wanke, and M. Steiner, “Similarity Search of Spatiotemporal Scenario Data for Strategic Air Traffic Management,” AIAA Journal of Aerospace Information Systems, accepted, Feb. 2019.

26

Page 27: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Key References: Published Results on UAV Communication

27

• M. Liu, Y. Wan, S. Li, F. Lewis, and S. Fu, “Learning and Uncertainty‐exploited Directional Antenna Control for Robust Long‐distance and Broad‐band Aerial Communication,” IEEE Transactions on Vehicular Technologies, accepted, 2019. 

• S. Li, M. Liu, C. He, Y. Wan, Y. Gu, J. Xie, S. Fu, and K. Lu, “Design and Implementation of Aerial Communication Using Directional Antennas: Learning Control in Unknown Communication Environment,” IET Control Theory and Application, in press, April 2019. 

• J. Chen, J. Xie, Y. Gu, S. Li, S. Fu, Y. Wan, and K. Lu, “Long‐Range and Broadband Aerial Networking using Directional Antenna (ANDA): Design and Implementation, IEEE Transactions on Vehicular Technology, Vol. 66, No. 12, pp. 10793‐10805, December 2017.

• J. Yan, Y. Wan, S. Fu, J. Xie, S. Li, and K. Lu, Received Signal Strength Indicator based Decentralized Control for Robust Long‐distance Aerial Networks using Directional Antennas, IET Control Theory and Applications, vol. 11, no. 11, pp. 1838‐1847, July 2017. 

• Y. Wan, S. Fu, J. Zander, and P. J. Mosterman, "Transforming On‐Demand Communications with Drones: The Needs, Analyses, and Solutions," Homeland Security Today Magazine, pp. 32‐35, April/May 2015.

Page 28: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Key References: Published Results on the Game‐theoretic UAV Coordinated Control

28

• V. G. Lopez, Y. Wan, and F. L. Lewis, “Bayesian Graphical Games for Synchronization in Dynamical Networks,” IEEE Transactions on Control of Network Systems, in Press, July 2019.

• V. G. Lopez, F. L. Lewis, Y. Wan, E. N. Sanchez, and L. Fan, “Solutions for Multiagent Pursuit‐Evasion Games on Communication Graphs: Finite‐Time Capture and Asymptotic Behaviors,” IEEE Transactions on Automatic Control, in press, July 2019.

• M. Liu, Y. Wan, and F. Lewis, “Adaptive Optimal Decision in Multi‐Agent Random Switching Systems,” IEEE Control Systems Letters, vol. 4, no. 2, pp.265‐270, April 2020. 

• M. Liu, Y. Wan, F. L. Lewis, V. G. Lopez Mejia, “Stochastic Two‐Player Zero‐Sum Learning Differential Games,” in Proceedings of the 15th IEEE International Conference on Control & Automation (ICCA), Edinburgh, Scotland, July 16‐19, 2019.

Page 29: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Key References: Published Results on UAV Weather and Path Planning

29

• C. He, Y. Wan, and F. Lewis, “On the Identifiability of the Influence Model for Stochastic Spatiotemporal Spread Processes,” IEEE Transactions on Systems, Man, and Cybernetics, accepted, April 2019. 

• M. Pinhero, M. Liu, Y. Wan and A. Dogan,  “On the Analysis of On‐board Sensing and Off‐board Sensing through Wireless Communication for UAV Path Planning in Wind Fields,” in Proceedings of AIAA Schitech, January 2019. 

• C. He and Y. Wan, Clustering Stochastic Weather Scenarios using Influence Model‐based Distance Measures, in Proceedings of AIAA Aviation Conference, Dallas, TX, June 17‐21, 2019. 

• C. He, Y. Wan, and J. Xie, “Spatiotemporal Scenario Data‐Driven Decision For the Path Planning of Multiple UASs,” in Proceedings of the 4th Workshop on Science of Smart City Operations and Platforms Engineering (SCOPE 2019), CPS Week, Montreal, Canada, April 15‐18, 2019. 

Page 30: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Key References: Published Results on Airspace Safety and Detect and Avoid Protocols

30

• M. Liu and Y. Wan, “Analysis of Random Mobility Models with Sense and Avoid Protocols for UAS Capacity Management,” in Proceedings of AIAA SciTech Conference, Florida, January, 2018.  

• M. Liu, Y. Wan, and F. Lewis, “Analysis of the Random Direction Random Mobility Model with A Sense‐and‐Avoid Protocol, in Proceedings of Wi‐UAV Workshop, Globecom, December 2017.

• T. Li, Y. Wan, M. Liu, and F. L. Lewis, “Estimation of Random Mobility Models using the Expectation‐Maximization Method,” in Proceedings of the 14th IEEE International Conference on Control & Automation, ICCA 2018, Anchorage, Alaska, June 12‐15, 2018. 

• J. Xie, Y. Wan, B. Wang, S. Fu, and J. H. Kim, “A Comprehensive 3‐Dimensional Random Mobility Modeling Framework for Airborne Networks,” IEEE Access, vol. 6, pp. 22849 – 22862, March 2018. 

Page 31: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Key References: Published Results on Contingency Management and Capacity Management

31

• J. Xie, A. R. Kothapally, Y. Wan, C. He, C. Taylor, C. Wanke, and M. Steiner, “Similarity Search of Spatiotemporal Scenario Data for Strategic Air Traffic Management,” AIAA Journal of Aerospace Information Systems, vol. 16, no. 5, May 2019.

• J. Xie, Y. Wan, Y. Zhou, S.‐L. Tien, E. P. Vargo, C. Taylor, and C. Wanke, "Distance Measure to Cluster Spatiotemporal Scenarios for Strategic Air Traffic Management," Journal of Aerospace Information Systems, vol. 12, no. 8, pp. 545‐563 (19 pages), 2015. 

• W. Zhang, J. Xie, and Y. Wan, “Spatiotemporal Scenario Data‐Driven Decision‐Making Framework for Strategic Air Traffic Flow Management,” in Proceedings of the 15th IEEE International Conference on Control & Automation (ICCA), Edinburgh, Scotland, July 16‐19, 2019.

• C. He and Y. Wan, Clustering Stochastic Weather Scenarios using Influence Model‐based Distance Measures, in Proceedings of AIAA Aviation Conference, Dallas, TX, June 17‐21, 2019. 

Page 32: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Community Engagement

32

• Gave talks at NSF, Southeast Tarrant Transportation (SETT) Partnership Breakfast, and many universities• Organized workshops at international conference and UTA• Work with Bell on the Student Competition 

Page 33: Urban Aerial Mobility: The Cyber‐Physical Systems Approach · •2004‐2009, worked on NASA project on strategic air traffic management •2010‐2015, worked on MITRE project

Thank You!!! Acknowledgement: 

Main collaborators: Dr. Frank Lewis (UTA), Dr. Shengli Fu (UNT), Dr. Jeff Lei (UTA),  Dr. Jae Kim (Boeing), Dr. Dr. Zongli Lin (U. of Virginia), Dr. Kevin Mills (NIST), Dr. Christine Taylor (MITRE), Dr. Craig Wanke (MITRE). 

Students and Postdocs: Dr. Songwei Li (Ph.D., now with Samsung), Dr. Junfei Xie (Ph.D., now faculty at SDSU), Dr. Yi Zhou (Southwest Airlines), Ms. Chenyuan He, Ms. Mushuang Liu, Dr. Yixin Gu (Postdoc), and many others.