1 / 27

Human mobility: taking a fresh look at its form and goals

Human mobility: taking a fresh look at its form and goals. Vincent Borrel, Franck Legendre, Marcelo Dias de Amorim Laboratoire LIP6 – CNRS Université Pierre et Marie Curie – Paris 6. Who’s this guy ?. 3rd year Ph.D in LIP6 - Paris Mobility modeling Algorithms for sensor networks

qabil
Download Presentation

Human mobility: taking a fresh look at its form and goals

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Human mobility: taking a fresh look at its form and goals • Vincent Borrel, Franck Legendre, Marcelo Dias de Amorim • Laboratoire LIP6 – CNRS • Université Pierre et Marie Curie – Paris 6

  2. Who’s this guy ? • 3rd year Ph.D in LIP6 - Paris • Mobility modeling • Algorithms for sensor networks • Internship in CoC - Atlanta • Mobility for the DTN group • Fresh air: we won’t agree ;D

  3. Mobility ? • Large-scale testsbeds are still lacking • Mobility models are required • For performance evaluation (analytical/simu) • As a cognitive tool for protocol design • Mobility is not well understood yet… • How to express it ? What mobility ? • What about realism ???! • How can it help ?

  4. The research shift (gladly stolen from Prof.Ammar)

  5. So what ? • GHOST: unifying mobility framework • SIMPS: Social trait in mobility

  6. GHOST

  7. The expression problem • Dozens of mobility models • Brownian, Vehicular, Pedestrian, Workspace, Campus, City Section, Calendar oriented, ... • Each one for a particular mobility case • Reality is more complex • Various people and behaviors coexist • One’s mobility varies throughout time • Persons react and adapt mobility to their surrounding • Infinite combinations of possible mobility models

  8. Mobility is a complex interaction

  9. The main aspect • Instead of mobility models, let's consider mobility traits • A particular mobility of a given individual at a given time is the result of the influence of several traits (e.g. calendar following, social interaction, obstacle avoidance, map following...) instead of one all encompassing model. • A component in the GHOST framework is the instantiation of a mobility trait, once formalized. It results in one or more interacting behavioral rules.

  10. GHOST: the idea • GHOST, a Mobility Meta-Modeling approach • Relying on the formalism of behavioral rules (from biological physics and AI) • Defining mobility primitives: chase, join, leave, … • GHOST is • Flexible: it allows to combine, add, delete new components • Expressive: it allows to define new models using trait composition • Interactive: TCL script interface (scenario definition, live interference)

  11. GHOST inside Basic inputs for ghosts

  12. GHOST Inside (cont'd) • Behavioral Rules: output acceleration requests • Accumulator: combines rules • Motion Core: • Physical limits check • Dynamic rules priority system

  13. GHOST Inside (cont'd) Mobility core: Behavioral rules are weighted in an acceleration request Which is checked against physical limits

  14. GHOST outside outdoor mobility indoor mobility

  15. SIMPS

  16. SIMPS: Where are we ? • Exploring a cause of mobility: the social trait in human motion • Typical predominance in crowd motion: mall, conference, protest, party, park, cafeteria… (did I tell you…)

  17. SIMPS: Origins in network sociology • Sociability: the number (volume) and classification (int.-ext.) of relationship with others • Fact 1: each individual has his own fixed sociability need (mostly dependent of social class and age) • Fact 2: individuals try to meet their needs by their actions (sociostating)

  18. Sociability evolution

  19. SIMPS • Is a mobility trait • Translates sociostation in the mobility domain • Concerns the volume aspect of sociability • Simplest set: two behavioral rules • Implemented using GHOST ;-)

  20. SIMPS: the twin behaviors • Socialize: When under-socialized (lonely), an individual is attracted toward each of his acquaintances • Isolate: When over-socialized (bored), the individual is repulsed by each stranger

  21. SIMPS: details 1 • Each individual has his own sociability: preferred number of others hanging around • One’s socialization feeling given by proxemics: number of others closer than in one’s social distance (~12ft in US, cf. E.Hall) • One’s socialization > his sociability: he’s oversocialized • Socialization < sociability: undersocialized

  22. SIMPS: details 2 • Attractive/repulsive forces diminish with distance between individuals • Direction of one’s acceleration request given by the sum of his attractions/repulsions • Force of one’s acceleration request given by his over/undersocialization amount

  23. SIMPS: The big picture

  24. SIMPS: results on contact and inter-contact durations • Simulated pure SIMPS motion (no other influence) • In-contact condition: node under a certain distance (here 6m for BT-like connectivity) • Main result: scale-free (with cutoff) contact/inter-contact distributions (Not aimed at !!!) • Robust feature through parameter change ! • Seems dependant on Socialize/Isolate assymmetry only. • Independent to changes in R.V. distributions (uniform or gaussian)

  25. SIMPS: things to take home • Mobility based on causes, not on consequences • Social trait: maintain one’s sociability • Renders Power-law contact and inter-contact distributions • No power-law at input • Robust • Not aimed for !

  26. Thanks ! (and now the demo…)

More Related