Information Dissemination in VANETs using Zone Based Forwarding Rayman Preet Singh Arobinda Gupta Dept. of Computer Sc. & Engg. Indian Institute of Technology, Kharagpur Dept. of Computer Sc. & Engg. Indian Institute of Technology, Kharagpur Abstract—Several VANET-based safety applications require information dissemination to all vehicles within a certain area for a certain time. Among the existing information dissemination protocols for VANETs, only Stored Geocast provides retention of information within a pre-determined area of the road for a duration of time. However, the overhead incurred by it is large. In this paper, we propose and evaluate ZBF, a zone based forwarding scheme for information dissemination which provides both a spatial and temporal retention of information and incurs less overhead than Stored Geocast. I. I NTRODUCTION A vehicular ad hoc network (VANET) is an ad hoc wireless communication system setup between multiple vehicles (V2V) or between a vehicle and some roadside infrastructures (V2I). Avoiding accidents and traffic congestions are two primary goals of building VANETs. A VANET can enable drivers to exchange information about current road conditions and potentially dangerous situations to warn other drivers about upcoming dangers, reducing the chance of accidents and traffic congestions. Any vehicle can initiate dissemination of such warning information. Usually, any such information has a specified area of interest and a fixed time period within which it is considered valid and usable. We will refer to these as the effect-area and the effect-time of the information respectively. For example, an information about an accident blocking a lane of a highway is of importance to vehicles only over a few kms from the accident site, for the duration of persistence of the lane blockage. An information dissemination protocol for VANETs that propagate such warning information should aim to deliver the message to maximum number of vehicles that pass through the effect-area during the effect-time, while incurring low message overhead. Using a flooding based approach in which all vehicles that receive the message broadcast it again causes too many messages to be sent. Several strategies for information dissemination in VANETs have been suggested that improve on simple flooding [1][2][3]. However, all these approaches are vulnerable to network partition during propagation. Fathy et al. in [4] propose a scheme which can handle network partitions. However, none of these works retain the message in the effectarea for the duration of the effect-time. The only existing information dissemination protocol in VANET that addresses message delivery within an effect area c 2011 IEEE 978-1-4577-2028-4/11/$26.00 β and time is Stored Geocast [5]. In this protocol, a leader vehicle is elected in the region of interest to store messages and handover of messages is done when this vehicle leaves the region. The elected vehicle broadcasts periodic GATE messages that also carry the warning information to be disseminated. A grid based approach is used for leader election which elects the node closest to the center of a region (grid box) as the leader. BID messages are used within each grid to exchange position information for electing a leader. In this paper, we present ZBF (Zone Based Forwarding), an information dissemination protocol for VANETs that delivers a message to a large number of vehicles passing a predefined effect-area during a predefined effect-time, while incurring very little message overhead compared to other protocols. Detailed simulation results are provided on realistic scenarios to show that ZBF outperforms Stored Geocast and other information dissmination protocols. An extension of ZBF is also proposed that can tolerate high network congestion. II. ZBF: Z ONE BASED F ORWARDING We assume that the transmission range of all vehicles are the same and is equal to π . The roads considered have twoway traffic with lower and upper speed limits as 0 and π£πππ₯ respectively. Also, each vehicle π knows its current position (ππ ), velocity (π£π ), acceleration (ππ ), and direction (πˆπ ) at any instant of time. ZBF divides the entire effect-area of the warning information into segments of length π , each of which is referred to as a zone. One vehicle in each zone is assigned the task of periodically broadcasting the warning information to notify other vehicles in that zone. This vehicle is referred to as a forwarder for that zone. Note that any broadcast by the forwarder will be received by all vehicles in a zone as long as the frequency π of periodic broadcast is sufficiently large (≥ π£πππ₯ π ). Whenever a forwarder exits a zone, a new forwarder is elected from amongst the vehicles present in the zone and the task of periodic broadcasting is then handed over to the new forwarder. At the end of the effect-time duration, all forwarders stop their periodic broadcast. The protocol uses three types of messages. An Info message is sent by a forwarder vehicle π to broadcast the actual warning information, and contains, other than a type field identifying it as an Info message, the fields <ID, Warning-Info, EffectTime, Effect-Area, π , πˆ > where πΌπ· is the sender’s id, and Note that ππ and πˆπ are available at π as the contents of Info message are also piggybacked in the Query message sent by π. Using π£π , ππ , and πˆπ , vehicle π computes the approximate time π‘π it will take to travel the distance π as π‘π = π£π π . Time π‘π √ π’2 + ππ .π − π’ can also be computed as for more accuracy. ππ Since our aim is just to approximately estimate which vehicle will be present in a zone for the longest duration, we prefer π to use ( ) for simplicity. Vehicle π then waits for a time π£π π π‘π€πππ‘ , which is calculated as π‘π€πππ‘ = max ( , π ) where π is π‘π a suitable normalizing factor. Thus, vehicles which will stay in the zone for a longer time will wait for a smaller time and vice-versa. If the π‘π€πππ‘ period expires without π receiving any other Reply message, π broadcasts a Reply message, changes its mode to Forward, and starts functioning as a forwarder. However, if a Reply message, broadcast from another vehicle π, is received during the wait period π‘π€πππ‘ , vehicle π cancels its transmission of the Reply message if both π and π are on the same side of the current forwarder π. If π and π are on different sides of π, π ignores the Reply message and continues to wait. On becoming the forwarder, vehicle π records its current position ππ as π πΈπΆππ in order to compute the zone boundaries. The forwarder π uses π πΈπΆππ to detect when it reaches the zone border, at which point it broadcasts a Query message. As described earlier, of all the vehicles in the zone, the one whose π‘π€πππ‘ expires first sends a Reply message and gets 200 No. of vehicles π and πˆ are the sender’s position and direction respectively. A Query and a Reply message contain a type field identifying them, plus all fields of the Info message (i.e., the contents of the Info message are also piggybacked on all Query and Reply messages sent). At any point of time, a vehicle can be in one of three modes, Receive, Forward, or Relay. The Forward mode signifies that the vehicle is a forwarder. Whenever a forwarder is about to leave a zone it changes its mode to Relay. During the Relay mode, a vehicle initiates election of a new forwarder while still continuing to function as a forwarder. All other vehicles, which are also the recipients of the periodic Info messages are in Receive mode. Before start of dissemination, all vehicles are in Receive mode. A vehicle π initiating the dissemination changes its mode to Relay, broadcasts a Query message, and waits for a predefined time π for receiving a Reply message. Any vehicle π at position ππ , on receiving a Query message from a forwarder π, computes its distance π from its current position ππ to the current position ππ of the sender of the Query message if vehicle π is headed towards ππ . In case π is headed away from ππ , π is calculated as the distance vehicle π needs to travel to reach the position ππ − π , i.e., the boundary of π’s transmission range from π’s current position. More formally, the distance π can be computed as follows: { if πˆπ = πˆπ β£ππ − ππ β£ π = π − β£ππ − ππ β£ if πˆπ = −πˆπ . 150 100 50 Total no. of vehicles in zone ZBF Torrent-Moreno Stored Geocast 0 0 0.015 0.03 0.045 0.06 0.075 0.09 0.105 0.12 f Fig. 1. Total no. of vehicles informed v/s π . elected as the new forwarder for the zone. The task of periodic rebroadcast of Info messages is hence, handed over to the new forwarder. The frequency of periodic re-transmissions of Info messages π£πππ₯ . The wait time π is empirically π is upper bounded by π selected such that π > 2 × (πππ π πππ ππππππππ‘πππ πππππ¦). III. S IMULATION R ESULTS The performance of ZBF is evaluated through simulation using an integrated traffic-cum-network simulator that couples ns-2 with the traffic sumulator VanetMobiSim for realistic VANET simulation. We measure (i) Coverage, The total number of vehicles passing through the effect-area during the entire effect-time which receive the information, (ii) total number of message broadcasts, and (iii) the distance a vehicle travels inside the effect-area before it gets the warning information for the first time. The performance of ZBF is compared with that of Stored Geocast [5], Torrent-Moreno’s algorithm [2], and Fathy et al.’s algorithm [4]. [2] and [4] are chosen after a detailed evaluation by simulation (not shown here) of several information dissemination protocols, in which these two protocols were found to outperform all others. A highway scenario of length 10 Km is used, with two lanes per driving direction, and a total of 500 vehicles with speed raging between 0 and 30 m/sec. The values of π and π are 0.1 and 1 second respectively. The zone length and the grid length respectively in the two approaches are set to the transmission range (=250 m). The performance is measured by varying the frequency π of the periodic Info message broadcasts in ZBF and GATE message broadcasts in Stored Geocast respectively π£πππ₯ = 0.12). The results are sampled (π is upper bound by π at 5 random instances for a period of 5 minutes for each of the three cases of one, two, and three zones or grids existing in the scenario, and then averaged over all these cases. Figure 1 shows the coverage achieved by both the algorithms with varying frequency π . The coverage of Stored Geocast increases as π increases. However it is surpassed by ZBF when π attains its maximum value of 0.12. At that value of π , ZBF ensures that the warning information is delivered to all the vehicles which are present in the zone during the No. of vehicles/messages No. of messages ZBF Torrent-Moreno Stored Geocast 200 150 100 50 0 0 0.015 0.03 0.045 0.06 0.075 0.09 0.105 0.12 Fig. 4. effect-time. Note that Torrent-Moreno’s algorithm delivers the warning information to a far less number of vehicles as it does not retain the information within the area of interest. Figure 2 shows the total number of message broadcasts by each of the strategies. As π increases the number of GATE and Info messages increases. We observe that the number of message broadcasts by Stored Geocast far exceeds those by ZBF. A further analysis shows that the increased message broadcasts in Stored Geocast is due to the large number of BID messages broadcast by it. Distance travelled 80 70 60 50 40 30 20 ZBF Stored Geocast 0 0 0.015 0.03 0.045 0.06 0.075 0.09 0.105 0.12 f Fig. 3. 300 200 100 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 Time from start of dissemination (sec) Total no. of messages broadcast v/s π . 10 400 0 f Fig. 2. No. of messages No. of nodes informed 500 Distance travelled inside a zone before being informed v/s π . Figure 3 shows the average distance any vehicle which is inbound to a zone travels before it gets the warning information. In case of Stored Geocast, this is the distance travelled before receiving the first GATE message broadcast from the leader vehicle. For both these approaches it is seen that as π is increased, the distance decreases. Stored Geocast outperforms ZBF at lower π values, but at higher values of π , ZBF performs better. Figure 4 shows the performance of Fathy et al.’s algorithm [4] on the same highway scenario. It is seen that as time progresses, information is eventually delivered to all the vehicles in the scenario. However, it takes 20-25 seconds to deliver the information to the same number of vehicles as in the case of ZBF. However, ZBF delivers the warning information only to the vehicles for which it is relevant within a distance of 40-80m from the zone boundary. At an average speed of π£πππ₯ = 15π/π ππ (π£πππ₯ = 30π/π ππ), this results in a delay 2 Result of the Fathy algorithm [4] on a sample highway scenario. of only 2.6 − 5.3 seconds. Thus, the worst case delay is much less for ZBF compared to that in [4]. In heavy traffic, Query messages may be lost due to collision etc., causing the warning information to be lost permanently in that zone. We extend ZBF as follows to address this problem. All vehicles in the Receive mode anticipate the receipt of a Query message in advance. On receipt of an Info message, a receiver vehicle obtains the forwarder’s position and direction, and uses that to compute the approximate distance π ′ which the forwarder will travel before reaching the zone boundary. If π ′ is less than a threshold π ′π π» , the receiver vehicle schedules the handling of the receipt of a Query message at an instant after π‘π π» seconds, at which the forwarder is supposed to reach the boundary. If a Query message is actually received by then, this scheduled execution is cancelled. However, if no Query message is received, the receiver vehicle sends a Reply message anyway after handling the perceived receipt of the Query message. All other steps of the algorithm remain the same as shown earlier. The extended protocol is simulated over the same highway scenario but with the number of vehicles increased to 1000 to induce heavy bumper-to-bumper traffic. It is seen that in heavy traffic, some Query messages are dropped. However, the proposed extension still retains the information in a zone by electing a new forwarder and achieves a greater vehicle coverage than ZBF. However, because Reply messages may be sent by more than one vehicle that fail to receive Query messages, the number of messages broadcast increases somewhat. R EFERENCES [1] M. Sun, W. Feng, T. Lai, K. Yamada, and H. Okada, “GPS-Based Message Broadcasting for Inter-vehicle Communication”, International Conf. on Parallel Processing (ICPP), 2000. [2] M. Torrent-Moreno, “Inter-Vehicle Communications: Assessing Information Dissemination under Safety Constraints”, 4th Annual IEEE/IFIP Conf. on Wireless On Demand Network Systems and Services, 2007. [3] R. Preet and A. Gupta, “Traffic Congestion Estimation in VANETs and Its Application to Infomation Dissemination”, 12th Intl. Conf. on Distributed Computing and Networking (ICDCN), 2011. [4] M. Fathy and S. Khakbaz,“A Reliable Method for Disseminating Safety Information in Vehicular Ad hoc Networks Considering Fragmentation problem”, 4th Intl. Conf. on Wireless and Mobile Communications, 2008. [5] C. Maihofer, W. Franz and R Eberhardt, “Stored Geocast”, Kommunikation in Verteilten Systemen (KiVS), Leipzig, Germany, 2003.