Topic: Vehicular Networks Team 6 R99922041 陳彥璋 R99922083 梁逸安 R99945051 洪晧瑜 CARS: Context-Aware Rate Selection for Vehicular Networks P. Shankar , T. Nadeem , J. Rosca , L. Iftode , Proc. IEEE ICNP , Oct. 2008 , pp. 1 – 12 Speaker: 陳彥璋 Outline Introduction Key challenges Context Aware Rate Selection Performance Conclusion Rate Selection IEEE 802.11 allows multiple transmission rate at the physical layer (PHY). Low bitrate High bitrate Low link quality Good High error rate High link quality Underutilization Good Link Quality High link quality. Low link quality due to long distance. Outline Introduction Key challenges Context Aware Rate Selection Performance Conclusion Key challenges 1. Rapid variations of the link quality. Mobility at vehicular speed. Key challenges 2. Few or no packets transmitted in estimation window during infrequent and bursty transmission. No past history to estimate link quality. Key challenges 3. Distinguish losses due to environment from hidden-station induced collision. Loss due to hidden station: rate↓ transmission time↑ contention↑ Outline Introduction Key challenges Context Aware Rate Selection Performance Conclusion Architecture Past transmission history. Positions and speeds of itself and its neighbors. Algorithm for all rate do Context information Transmission rate Packet length Past frame transmission statistics input input Ec α Eh Packet error rate Throughput end for (1-α) α is assigned based on the vehicle speed. Find the rate that maximize the throughput. Algorithm Ec Empirical model. Measurements from extensive outdoor vehicular experiments. Use multivariate linear regression as the learning approach. Eh Exponentially weighted moving average (EWMA) of past frame transmission statistics. http://en.wikipedia.org/wiki/File:Exponential_moving_average_weights_N%3D15.png Outline Introduction Key challenges Context Aware Rate Selection Performance Conclusion Experimental Result Simulation Result Outline Introduction Key challenges Context Aware Rate Selection Performance Conclusion Conclusion Context Aware Rate Selection Use context information to perform fast rate adaption in vehicular network. Connectivity-Aware Routing (CAR) in Vehicular Ad Hoc Networks Valery Naumov & Thomas R. Gross ETH Zurich, Switzerland IEEE INFOCOM 2007 speaker:梁逸安 Outline Introduction Related Works (GPSR) Connection-Aware Routing (CAR) Simulation Conclusion Introduction Vehicular ad hoc networks (VANETs) using 802.11-based WLAN technology have recently received considerable attention in many projects Several geographic routing (GR) protocols use an idealized mechanism such that for every originated data packet the true position of the destination is known Introduction Another problem is that, all of the GR protocols do not take into account if a path between source and destination is populated. This paper presents a novel position-based routing scheme called Connectivity-Aware Routing (CAR) to address these kind of problems Outline Introduction Related Works (GPSR) Connection-Aware Routing (CAR) Simulation Conclusion Greedy Perimeter Stateless Routing Greedy Perimeter Stateless Routing Perimeter Mode Greedy Perimeter Stateless Routing Outline Introduction Related Works (GPSR) Connection-Aware Routing (CAR) Simulation Conclusion Connection-Aware Routing (CAR) The CAR protocol consists of four main parts: ◦ (1) destination location and path discovery ◦ (2) data packet forwarding along the found path ◦ (3) path maintenance with the help of guards ◦ (4) error recovery Destination location discovery A source broadcast a path discovery (PD) Each node forwarding the PD updates some entries of PD packets If two velocity vectors’angle > 18°, anchor is set. Greedy forwarding over the anchored path A neighbor that is closer to the next anchor point is chosen (greedy) , instead of destination. Path maintenance If an end node (source or destination) changes position or direction, standing guard will be activated to maintain the path. Path maintenance If end node changes direction against the direction of communication, traveling guard will be activated. ◦ A traveling guard runs as end node’s old direction and speed, and reroute the packets to the destination. Path maintenance Routing error recovery The reason for routing error ◦ A temporary gap between vehicles (1) Timeout algorithm When a node detects a gap – buffer the packets (2) Walk-around error recovery When Timeout algorithm fail , do location discovery Whether the location discovery is successful, the result will be reported to the source node. Outline Introduction Related Works (GPSR) Connection-Aware Routing (CAR) Simulation Conclusion Simulation Scenarios ◦ City ◦ Highway Traffic density ◦ Low – less than 15 vehicles/km ◦ Medium – 30-40 vehicles/km ◦ High – more then 50 vehicles/km SimulationPacket Delivery Ratio SimulationAverage data packet delay SimulationRouting overhead Outline Introduction Related Works (GPSR) Connection-Aware Routing (CAR) Simulation Conclusion Conclusion Address the populated problem about paths. Path discovery & Anchor points Path maintenance with guards Error recovery Higher performance and lower routing overhead than GPSR Delay-bounded Routing in Vehicular Ad-hoc Networks Antonios Skordylis, Niki Trigoni Oxford University Computing Laboratory ACM International Symposium on Mobile Ad hoc Networking and Computing, 2008 Speaker : R99945051 洪晧瑜 Outline Introduction ◦ VANETs ◦ Delay-bounded Routing Objective and Model Algorithm ◦ D-Greedy ◦ D-MinCost Evaluation and result Conclusion Introduction VANETs ◦ vehicles equipped with wireless transceivers that will enable them to communicate with each other form a special class of wireless network • Delay-bounded Routing ◦ timely and bandwidth efficient data dissemination from vehicles to an access point, given statistical information about road traffic ◦ tradeoff : timely data delivery v.s. low bandwidth utilization Outline Introduction ◦ VANETs ◦ Delay-bounded Routing Objective and Model Algorithm ◦ D-Greedy ◦ D-MinCost Evaluation and result Conclusion Objective Objective ◦ carry-and-forward algorithms leverage knowledge of traffic statistics in an urban setting ◦ enable timely delivery of messages from vehicles to APs ◦ minimizing wireless transmissions/optimizing bandwidth utilization Model Urban scenario Vehicles (mobile nodes ): ◦ geographical position(GPS receiver ) ◦ digital map(G(V,E)) : historical traffic statistics u: average speed, d: average vehicle density per road segment ◦ communication range: 250m APs(stationary access points ): ◦ infrastructure nodes whose absolute location in known to all vehicles Message informations: ◦ tg : message generation time ◦ λ : time-to-live value, message delay threshold Outline Introduction ◦ VANETs ◦ Delay-bounded Routing Objective and Model Algorithm ◦ D-Greedy ◦ D-MinCost Evaluation and result Conclusion Algorithm Forwarding a message ◦ minimize the number of transmissions ◦ within the message-specific delay threshold Alternate between two forwarding strategies: ◦ Multihop Forwarding ◦ Data Muling Algorithms ◦ D-Greedy ◦ D-MinCost D-Greedy Delay-bounded Greedy Forwarding ◦ No knowledge of global traffic conditions ◦ Available location information, ex. Node speed ◦ Best path: shortest path D-Greedy Each vehicle maintains a neighbor list by periodically broadcasting beacons ◦ id : unique vehicle identifier ◦ distToAP : the length of the shortest path between the vehicle’s current location and the location of the closest access point (Dijkstra) TTL : delay threshold value for message distToInt : the remaining length, until the next intersection, of the current street segment e D-Greedy:Formulation Data Muling strategy ◦ DelDM ≤ Del Multihop Forwarding strategy ◦ otherwise Del : available delay budget Del = TTL × distToInt/distToAP DelDM = distToInt/u D-MinCost Delay-Bounded Minimum Cost Forwarding ◦ Knowledge of global traffic statistics ◦ ex. estimated values of average vehicle speed u and density d for all edges of the street graph G. ◦ bandwidth-efficient delay-constrained paths D-MinCost Annotate each edge with two metrics ◦ C : cost , #message transmissions along the edge ◦ Del : delay , the time required to forward a message along the edge Graph extension D-MinCost Data Muling strategy Cost & Delay ◦ DelDM = l/u, CDM = 1 ◦ l : the length of the edge ◦ u : the average vehicle speed along that edge Multihop Forwarding strategy ◦ ◦ ◦ ◦ ◦ if l > R and d ≥ l/R CMH = l/R, DelMH = CMH × q R : communication range d : the average vehicle density for the edge q : the time required for the node to check its neighbor list and identify the best next hop D-MinCost:Path selection The minimum cost path The delay-constrained least-cost routing problem is known to be NP-complete Heuristics: the Delay Scaling Algorithm(DSA) [7] ◦ The AP that can be reached with the least cost. ◦ The exact min-cost path to that AP. ◦ The strategy that should be followed at each edge of the path in order to adhere to the message’s remaining delay budget. AP min-cost path strategy Outline Introduction ◦ VANETs ◦ Delay-bounded Routing Objective and Model Algorithm ◦ D-Greedy ◦ D-MinCost Evaluation and result Conclusion Evaluation and result Compared with ◦ Epidemic routing : achieves optimal delay and delivery ratio under our scenario ◦ MinDelay : a greedy delay minimizing scheme inspired by [16] Delivery Ratio Transmitted Bytes Message Delay Strategy chosen Outline Introduction ◦ VANETs ◦ Delay-bounded Routing Objective and Model Algorithm ◦ D-Greedy ◦ D-MinCost Evaluation and result Conclusion Conclusion Two novel packet forwarding schemes : ◦ D-Greedy ◦ D-MinCost Achievement ◦ Forwarding strategy ◦ Minimize communication cost ◦ Adhere to defined delay threshold Outperform ◦ communication cost Similar delivery ratio to Epidemic Compared with Epidemic &MinDelay