Context Assisted Routing Protocols for Inter-Vehicle Wireless

Document Sample
Context Assisted Routing Protocols for Inter-Vehicle Wireless Powered By Docstoc
					Context Assisted Routing Protocols for
Inter-Vehicle Wireless Communication


        V. Dumitrescu and J. Guo


        Presented by: Imranul Hoque
Contribution

   Authors’ Claim
       Network Partitioning Method
       Context Assisted Routing Protocol
Motivation


             Partitioning




             Destination

               Source
Partitioning Method!!
   Assumptions:
       Vehicles know their coordinate through GPS
       Static traffic info are stored on-board
       Real-time dynamic traffic info is gathered through IVC
   Push data:
       Non-critical data broadcast to all vehicles in the region
       Critical data broadcast to all vehicles approaching the
        region
   Pull data:
       Vehicles inquire for region specific data
   Messages: GET_LCL_CTX, GET_RMT_CTX,
    SENS_CTX, SND_LCL_CTX, RE_CTX
Context Assisted Routing

   Spatial Model
   Predictive Location Service
   CAR Algorithm
Spatial Model
Spatial Model Graph

                                    Fastest Route
                                    Highest Traffic Density
                                    Shortest Route




                                       Edges are weighted




            Layered format of the spatial model can be used
Predictive Location Service

   PLS keeps track of node locations
   Locations becomes obsolete
   Frequent broadcast is a solution: high
    overhead
   PLS is the solution: predicts node
    location
Predictive Location Service (2)




                            Broadcast
Context Assisted Routing
                     S wants to route a packet to D

                     1. S gets coordinate from GPS
                     2. S knows D’s coordinate from
                        PLS
                     3. S knows the Spatial Model
                        Graph
                     4. S locates itself and D on the
                        SMG
                     5. S calculates shortest path to
                        D on SMG
                     6. S does Greedy Forwarding
                        to A (A retention cache can
                        be used if a suitable node
                        is not found)

         Problems?
Evaluation

   NS2 simulation with CMU wireless
    extension
   Compare: CAR, CARC, GPSR
   Metrics:
       Packet Delivery Ratio
       Packet Delivery Delay
       Average Hop Count
       Packet Delivery Ration
Retention cache
improves performance



  GPSR can switch to
  perimeter mode
      Packet Delivery Delay

Retention cache




Perimeter mode in GPSR
Conclusion

   Drawbacks:
       Fixed destination
       Assumes that any intermediate vehicle
        will go to the destination
       Not very much novel