A Mobility-Centric Data Dissemination Algorithm for Vehicular Networks (MDDV)

1
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks (MDDV)

• Evgueni (Eugene) Khokhlov

2
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Hao Wu Ph.D.
• Richard Fujimoto Ph.D.
• Randall Guensler Ph.D. Civil Engineering
  (transportation department)
• Michael Hunter Ph.D.
• At Georgia Institute of Technology

3
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Purpose
• I.T. (mobile wireless communications) in
  Surface transportation Systems
• Enhance communication in existing intelligent
  transportation systems.
• The Intelligent Transportation Systems (ITS)
  program is a worldwide initiative to add I.T. to
  transport infrastructure and vehicles.
• Also known as part of Telematics in Europe the
  science of sending, receiving and storing
  information via telecommunication devices.

4
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Current ITS (Intelligent Transportation Sys)
• Infrastructure heavy
• Rely on roadside sensors, cameras, networking
  gt
• Hard for gov. agencies to get for ITS

5
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Trend
• Equip cars with computing and communication
  capabilities thus
• Reduce dependency on the government
• As cars are replaced so as the equipment
• How to avoid traffic

6
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• MDDV is designed to exploit vehicle mobility for
  data dissemination
• MDDV combines
• Opportunistic forwarding
• Trajectory based forwarding
• Geographical forwarding

7
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• V2V network (vehicle to vehicle ad-hoc) contains
  GPS, digital map, and sensors.
• Predictable (high mobility)
• Dynamic (rapidly changing topology)
• Constrained (one dimension movement)
• Potentially large scale
• No power constraints

8
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• V2V network
• Trees are undesirable,
• Localized algorithms based on vehicles
  interacting with neighbors are preferred.
• Problems
• Unreliable communication channels
• Vehicle failure
• High mobility

9
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Data dissemination transport of information to
  intended receivers (all vehicles going to UIC,
  police cars that are close by).
• Low delay
• High reliability
• Low memory occupancy
• Low message passing overhead.

10
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Data Delivery Mechanism
• Opportunistic forwarding targets networks when
  an end-to-end path cannot be assumed to exist.
  Messages are stored and forwarded to another
  node, a copy may remain with the original and be
  forwarded again later to improve reliability

11
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Data Delivery Mechanism
• Trajectory based forwarding directs messages
  along predefined trajectories. Messages are
  moving along the road graph.
• Helps limit data propagation along specific
  path and reduces message overhead.

12
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Data Delivery Mechanism
• Geographical forwarding attempts to move the
  message geographically closer to the destination.
  
• In V2V geographical distance has to be defined
  as graph distance.

13
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Assumptions
• The vehicle knows the road topology through a
  digital map and its own location in the road
  network via a GPS device.
• Source ID, source loc, generation time,
  destination region, expiration time, etc, are
  placed in the message header.

14
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Deliver a message to all vehicles in/entering
  region r before time t while the data source s is
  outside of r

15
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Propagation depends on vehicle traffic.
• High vehicle density fast information
  propagation
• Only use vehicles that travel in the same
  direction as where the message needs to go.
  Ignore vehicles traveling in the opposite
  direction.
• Use smallest sum of weights from the source to
  the destination region in the weighted road
  graph.

16
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Data Exchange
• With a peer-based scheme a node talks to
  encountered peers.
• Structure-base schemes (clusters, trees) wont
  work for this context.

17
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Two possible data exchange sequences.

18
(No Transcript)
19
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Conclusion
• MDDV is designed to address the data
  dissemination problem in a partitioned and highly
  mobile vehicular network. Messages are forwarded
  along a predefined trajectory geographically.
  Intermediate vehicles must buffer and forward
  messages opportunistically.

20
A Mobility-Centric Data Dissemination Algorithm
for Vehicular Networks

• Questions?