Research and Teaching Plans

Hen-I Yang, PhD Candidate

Computer and Information Science and Engineering Department

University of Florida

May 1, 2008

Pervasive computing

Source: Mobile Com

puting “Introduction to Pervasive Computing”

lectureExcerpted and m

odified from ‘U

biquitous City Plan’, NCA, Korea, 2005.

Pervasive ComputingApplications

Users

Computers

ResearchTeaching

House of Horror:Corridor of Doom

Shard resourcesNo reliable monitoring

and arbitrationServices unaware of

others’ operations on share resources

Proper coordination needed Energy Saving Svc

Lighting Service

ResearchTeaching

House of Horror:Steak Incinerator

Predefined range and conditions of operations for each device

The cues, hints and visual feedbacks fall to deaf ears when operated by systems

Invalid operations need to be monitored and prevented

3500°F !!

ResearchTeaching

House of Horror:Duel of Stove and Fridge

Side effects exist Impossible to account for all possible

interferences from environmental effects Non-determinant behaviors

Hey, Who Turns Off The Light?

ResearchTeaching

GoalsTotal safety guarantee is impossible◦Accidents happen ◦People make mistakes◦Murphy’s Law

Attainable goals:Do no harmHandle safety risks Proactively detect, prevent and manage

the inherent safety risks

ResearchTeaching

Challenge and ImpactPervasive computing is still in its infancy◦ Enhancement and assurance of safety a major block in real-

world adoption and deploymentSafety practice for existing software system does not work◦ General purpose versus specialized system◦ Financial feasibility◦ Explore viable safety practice for low-cost general-purpose

mission-critical systemsCurrent (limited) safety research on pervasive computing

only address very specific issues◦ Generic software practice and architecture required◦ Results can be extend to improve the reliability and security of

other dynamic and distributed systems such as web services

ResearchTeaching

Pervasive Computing Research

Other Research Aspects

System Support

Research Agenda

Safety

SecurityPrivacy

Software Engineering

Practice

Authorship Support

Evaluation Framework

Reliability/ Availability

ResearchTeaching

Safety Analysis, Protection and EvaluationTo enhance safety:Reduce probability of

bad things happening◦ System Modeling and

Analysis◦ Safety Protection◦ Risk Avoidance

Reduce damage caused by potential failure◦ Risk detection and

recognition◦ Emergency

termination and handling

◦ Emergency Recovery

ResearchTeaching

Risk Reduction Ambient calculus based modeling

◦ Function, Interaction and Mobility Description logic based modeling

◦ Context, Interpretation and Decisioning Component-based parameter analysis

◦ Operation and Invocation Protection of single component

◦ Casting, validation and event-driven programming paradigm Protection of component Interaction

◦ Choreography verification, concurrency System-wise Protection

◦ Ontology and active context reasoning Risk warning system

◦ Context-driven preventive behavior specification Service recomposition

◦ FPQSPN based service monitoring and composition Data compensation

◦ Virtual sensor technique, in-network data aggregation

Risk Avoidance

System Modeling

Safety Protection

ResearchTeaching

Damage Containment and Reduction

Risk detection and recognition Optimized context definition and reasoning Behavioral pattern recognition

Emergency termination and handling Middleware support for emergency stop (preemption,

prioritization and garbage collection) Emergency handler vector and mandatory emergency shutdown

routine Exception propagation mechanism

Emergency Recovery Highest priority emergency recovery services Idempotent service implementation

ResearchTeaching

Sustainable Research Program

Resources

Funding Facility Talent

Innovation Leadership

Research Outcome

Publication Implementation Theory SoftwarePractice

Real-worldApplications

Organization SupportCollaborationService

ResearchTeaching

Sustainable Research Program: Funding National Science Foundation (NSF) grants

◦ Networking Technology and Systems (NeTS),◦ Cyber-Enabled Discovery and Innovation (CDI) programs, ◦ Information Technology Research for National Priorities (ITR)

grants,◦ CAREER award

Administration on Aging (AoA) grants, National Institute on Disability and Rehabilitation Research

(NIDRR) grants National Institute of Health (NIH) grants Industry sponsorship with matching state funds.

◦ Microsoft Research, Ubiquitous Computing Group◦ IBM Research, Emerging Software Standards◦ GE Global Research, Computing and Decisioning System◦ Intel Research, Sensor Network Technologies and Applications

ResearchTeaching

Sustainable Research Program Facilities

◦ Lab space for intelligent environment◦ 3 Servers and 6 workstations◦ 3 Pocket PCs◦ A couple Dozens of sensor platforms◦ Various sensors and actuators◦ RFID tags and readers

Student Recruitment◦ Intra-departmental recruiting through

undergraduate network/HCI classes and senior projects

◦ Inter-departmental recruiting through interdisciplinary projects and collaboration

◦ External recruiting through Network of collaborators from other

universities and Publications in conferences Well-designed web presence Publicity of research outcome

◦ Initial funding from starting package

Collaboration◦ Interdisciplinary Nature of PerCom◦ @ UF

William Mann (Gator tech smart house) Benjamin Lok (HCI) Prabhat Mishra (embedded system) Michael Fang (sensor network)

◦ Academic Carl Chang (software engineering, Iowa State) Srini Seshan (network, CMU) Dave O’Hallaron (performance evaluation,

CMU)

◦ International Bessam Abdulrazak (HCI and robotics, Canada) Hani Hagras (intelligent agents, UK) Hyun Kim (robotics, Korea)

◦ Industry Randy Carroll (emerging software standards,

IBM) Ming Lu (information systems and service

computing, IBM)

Services

ResearchTeaching

Vision 2013Establish the position as the leading institute in research on

safe pervasive computing systemsDesign and implement the cutting-edge software

engineering practice and platform viable for safe mass real-world deployments

Create a mid-size sustainable research program (8-10 scholars) supported by a mix of funding from both the government and industry

Collaborate with a network of renowned scholars in the U.S. and international on an open platform

Lead in inter-departmental and inter-collegian joint exploration of safe pervasive computing deployment in various application domains

ResearchTeaching

Teaching Experience

CEN 4500C

Computer Network Fundamentals

Teaching Assistant

Spring 2007Spring 2008

CIS 6930

Special Topics on Sensor Networks

Guest Lecturer &MentorSpring 2007

CEN 5531

Mobile Computing

Guest Lecturer

Fall 2006

CGS 3460

Computer Programming Using C

Instructor

Spring 2006

ResearchTeaching

STEP Program

Chris Toth, HS junior(summer ‘07)

“Social Navigator for the Visually Impaired”

Graduate Research

Chao Chen, PhD student(Jan. ‘07 – present)

“Performance Evaluation of Pervasive Computing Systems”

Rohin Sethi, Masters student(Apr. ‘07 – present)

“Universal Remote Control for Smart Houses”

Mentoring ExperienceTerm Project

Sushant Gupta, Ankit Hirdesh, Masters student

(Jan. - July ‘07)“Remote GUI Client for Atlas”

ResearchTeaching

Teaching PhilosophyDeliver the fundamentals, but encourage

students to exploreTeaching is an interactive process and a

two-way streetPush available resources forward (80-20

rule)Teaching and research are synergetic, not

competing activitiesLearning is all that matters

ResearchTeaching

Courses of InterestUndergraduate courses◦CGS 2414 Computer Programming Using Java◦CGS 3460 Computer Programming Using C◦COP 4600 Operating Systems◦CEN 4500 Computer Network Fundamentals

Graduate courses◦CEN 5531 Mobile Computing◦COP 5615 Operating System Principle (Distributed systems)◦CEN 6505 Advance Computer Networks◦CIS 6930 Sensor Networks

ResearchTeaching

Sample Undergraduate CourseCGS 3460 Computer Programming Using C

Course Activities – 6 Individual Projects, 6 Quizzes, Midterm and Final

Teaching◦3 Hours of lectures, 10 Hours of computing lab, and

another 3 office hours per week◦ Interleaved use of slides, board and demo code

• Online forum and designated rotating email response TA

• Fun projects with backstory• Mid-term evaluation

ResearchTeaching

Sample Graduate CourseCEN 5531 Mobile Computing

Course Activities – 1 Group Term Project, 1 Individual Survey Paper, short paper summaries, 2 Paper Presentations and Discussions, Midterm and Final

Teaching◦ 3 Hours of lectures, including opening introductions, student

presentations, and concluding discussion◦ Meetings with each group at checkpoints throughout the semester

• Student-designed and managed projects• Town hall meetings• Interdisciplinary Guest panelists and judges• Offer existing research outputs from the lab to students, and

draw inspiration from discussions and projects of students

ResearchTeaching

Course Development 2-Course set for Pervasive Computing Upper-level undergraduate course “Introduction to Pervasive

Computing”◦ Introductory lectures on context-awareness, ad-hoc protocols and self-

organization, service-oriented architecture, reliability and availability, etc◦ Panel/brainstorming session with domain experts from occupational therapy,

civil engineering, agriculture, and education, etc Graduate course “Advance Pervasive Computing System Design,

Implementation and Analysis”◦ Collaborative offering with other department/college◦ Targeted for students with specialties in systems, networks, and HCI, as well

as other colleges with domain expertise◦ Create interdisciplinary teams to create interoperable projects◦ Experts give guest lectures on visions, requirements and limitations on

applying these new technologies in their respective domains◦ Employ lab research output as the backbone of course offering

ResearchTeaching

Questions?Thank you for your attention.

Further information on is available @http://www.cise.ufl.edu/~hyangOr contact me @hyang@cise.ufl.edu