1. Cost Effective Mobile and Static Road Side Unit Deployment for Vehicular Adhoc Networks Presenter: Yesenia Velasco (Senior in Computer Science) Department of Mathematics and Physics North Carolina Central University Reference: Donghyun Kim (NCCU), Yesenia Velasco (NCCU), Zishen Yang (Xi’an Jiaotong Univ., China), Wei Wang (Xi’an Jiaotong Univ., China), Rasheed Hussain (Hanyang University, South Korea), and R.N. Uma (NCCU), "Cost Effective Mobile and Static Road Side Unit Deployment for Vehicular Adhoc Networks," Proceedings of International Workshop on Computing, Networking and Communications (CNC) in conjunction with International Conference on Computing, Networking and Communications (ICNC 2016), February 15-18, 2016, Kauai, Hawaii, USA.Cost Effective Mobile and Static Road Side Unit Deployment for Vehicular Adhoc NetworksInternational Conference on Computing, Networking and Communications (ICNC 2016),

2. Vehicular Adhoc Networks (VANETs)  Uses motor vehicles to create a wireless mobile network  Vehicles are:  producers of data  services consumers  VANET infrastructure:  VANET service providers  Authorities  Road Side Units (RSUs) o Wireless relay nodes o Maximizing coverage is crucial Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

3. Motivation Traffic monitoring Vehicle safety Inter-vehicle communication Routing Location-based services Real time data Unbounded network size Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

4. RSU Deployment  Goal: Deploy RSUs such that total deployment cost is minimal and below the allotted budget  Chance of a VANET node having access to an RSU is maximal  We assume RSUs will be deployed over:  Static locations o Difficult to relocate o Constant coverage at a particular location  Mobile busses or light rail o Location changes over time o Route in known in advance  Fully controllable vehicles o Able to be easily relocated o Strategic route trajectory must be calculated  RSU location calculated using complex city road map Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

5. Assumptions  Budget  Fixed and known cost to deploy RSUs of all types  Mobile RSUs  Routes  Uniform speed  No traffic jams o Prevented in light rails and cities without much traffic  Fully controllable RSUs  Uniform speed  No traffic jams o Prevented by constructing their travel schedules under very low speeds  Known significance of each region  Determined by collecting relevant statistical information over time Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

6. City Map Abstraction  Communication range of VANET nodes and RSUs is r e.g. Fig. 1  Assume shape of metropolitan area map is rectangular e.g. Fig. 2  Partition map into squares or radius r  Each square is a candidate location for an RSU e.g. Fig. 3  Two adjacent locations (dots) have a path connecting them e.g. Fig.4 o Vehicles can travel one path per unit of time  Add weight (importance) to each location  Note that weight changes over time Fig. 2 Fig.1 Fig.2 Fig.3 Fig.4 Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

7. Directed Acyclic Graph Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

8. Overview of Proposed Approach Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

9. Budgeted Maximum Coverage Algorithm With Cardinality Constraints Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

10. Budgeted Maximum Coverage Problem – Cont’ Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

11. Maximum k Coverage Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

12. Maximum k Coverage Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

13. Performance Analysis Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016..

14. Performance Analysis – Cont’ Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016..

15. Performance Analysis – Cont’ Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016..

16. Budgeted Maximum Coverage Example Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 60 55 10 15 50 45 50 40 30 20 30 45  Budget: 70  Static RSU Cost: 10  Mobile RSU Cost: 10  Fully Controllable RSU Cost: 30

17. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 60 55 10 15 50 45 50 40 30 20 30 45 Budgeted Maximum Coverage Example

18. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 10 15 50 45 40 30 20 30 45 Budgeted Maximum Coverage Example

19. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 10 15 30 20 30 45 Budgeted Maximum Coverage Example

20. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 10 15 30 20 30 Budgeted Maximum Coverage Example

21. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 15 20 30 Budgeted Maximum Coverage Example

22. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 15 20 30 Maximum k Coverage Example

23. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. 15 20 30 Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0

24. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0 15

25. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0 15

26. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0 15 Two Possible Routes

27. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0 15

28. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0 15

29. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0 0 15

30. Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.. Maximum k Coverage Example 0 0 0 0 0 0 0 0 0 0 0 0 0 15

31. Simulation Results Fig.5 Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

32. Simulation Results - Cont’ Fig.6a Initial Graph Fig.6b Six fully controllable RSUs selected Fig.6c Ten mobile RSUs selected Fig.6d Two static RSUs selected Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

33. Spatial Temporal Coverage  Weighted coverage fluctuates as:  mobile RSUs travel to higher or lower weighted areas  Bus routes are not continually being offered at all times  Hours from 2 A.M to 4 A.M have the lowest coverage e.g. Fig. 7  Most mobile bus routes are not offered at that time  Peak coverage: 0.0165  Lowest coverage: 0.0124  Mean: 0.0143  Range: 0.0041  Higher stability could be achieved with a larger number of static RSUs or fully controllable RSUs such that routes are continually offered Fig.7 Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

34. Weighted Coverage Change Over Cost  Effects an increasing cost of deploying of an RSU type has on the overall weighted coverage given an initial condition. e.g. Fig. 8  Initial Condition:  Budget: 100  Static RSU cost: 10  Mobile RSU cost: 10  Fully controllable RSU cost: 30  Static RSUs  Few selected initially  Steady but low coverage area  Relatively inexpensive Fig.8 Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.  Fully Controllable RSUs  Few selected initially  Greatest coverage  Expensive  Mobile RSUs  Many selected initially  Good coverage  Relatively inexpensive

35. Future Work Investigate the tightness of our algorithm Investigate the tightness of our algorithm Investigate other possible solutions leading to better results Investigate other possible solutions leading to better results Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.

36. Questions? Yesenia Velasco, Senior, Computer Science @ Department of Mathematics and Physics 2016 ICNC, North Carolina Central University, February 15, 2016.