eterminiti Ae for SRC/802.11p Vehiular Safety Communiation Jihene Rezgui, Soumaya Cheraoui, Omar Charoun INTERLAB Reearh Laboratory Univerité de Sherbrooe, Canada {jihene.rezgui, oumaya.heraoui, omar.haroun }@uherbrooe.a Abtrat In thi wor, we preent the deign of an effiient eterminiti medium Ae (A for ediated Short-Range Communiation (SRC vehiular afety ommuniation over IEEE 802.11p, alled Vehiular A (VA. VA upport two type of afety ervie (emergeny and routine afety meage with different prioritie and trit requirement on delay, epeially for emergeny afety meage. VA proee both type of afety meage to maintain a balane between two onfliting requirement: reduing hane of paet olliion and lowering the tranmiion delay. To avoid long delay and high paet olliion, VA allow vehile to ae the wirele medium at eleted time with a lower ontention than would otherwie be poible within two-hop neighborhood by the laial 802.11p ECA or CF heme. Partiularly, our heme provide an effiient adaptive adjutment of the Contention Free Period (CFP duration to etablih a priority between emergeny and routine meage. Simulation how that the propoed heme learly outperform the laial CF heme ued by 802.11p in high-offered load ondition while bounding the tranmiion delay of afety meage. Inde Term Vehiular ad Ho networ, ontention-free, afety meage, determiniti ae. I. INTROUCTION Vehiular Ad ho Networ (VANET i urrently onidered an eential tehnology for future road afety and telemati appliation. The Federal Communiation Commiion (FCC of the U.S. approved the 75MHz bandwidth at 5.850-5.925 GHz band for Intelligent Tranportation Sytem (ITS. Thi wirele petrum i ommonly nown a the ediated Short-Range Communiation (SRC petrum alloated by to be ued eluively for Vehile-Vehile (V2V and Vehile-Road (V2R ommuniation. evie operating in SRC petrum will be uing IEEE 802.11p by following the WAVE operation mode [2]. SRC petrum i made up of even 10 MHz wide hannel a hown in Fig.1. Channel 178 i the ontrol hannel (CCH, whih i the default hannel for ommon afety ommuniation. The two hannel at the end of the petrum band are reerved for peial ue. The ret are ervie hannel (SCH available for both afety and non-afety ue. There ha been a vat literature [2, 8-12] on the deription and evaluation of SRC and VANET tehnologie. A thorough urvey an be found in [8]. Eiting wor that ue SRC/802.11p, tre the importane of meeting the trit delay and low paet olliion requirement of afety appliation, epeially in high offered-load ondition and try to find adequate olution to thee iue. Thee wor an roughly be divided into three ategorie: broadat enhanement heme [9], MAC layer olution for baoff algorithm improvement [2] and ommuniation rate and/or power adjutment trategie [12]. Fig.1 SRC petrum and hannel in USA Wherea, a few wor ontribute on etablihing delay bound to guarantee a hort delay in IEEE 802.11p [13], our novel propoed approah for Vehiular eterminiti Ae, alled VA, omplement previou olution in term of tringent delay bound for afety meage. To reah thi goal, we propoe a determiniti ae for afety appliation and we etablih a priority between routine and emergeny meage. To the bet of our nowledge, we are the firt to onider A over IEEE 802.11p. Some of the fator that affet mot IEEE 802.11p performane and reliability, epeially at high vehiular denitie, are it hannel ae priority mehanim and it CSMA baoff proe. Emergeny afety meage requirement of low delay and low paet olliion are diffiult to guarantee in dene vehiular enario, beaue of the random ontention ued by the traditional CSMA/CA MAC in IEEE 802.11p. Some tudie tried to olve thi problem by enforing a ontention-baed MAC with omple heme or by propoing modifiation to the baoff algorithm [2]. The development of a robut and effiient MAC protool will be eential to the apability of SRC devie in enabling reliable afety appliation. To ahieve uh a protool, we propoe enhanement to the 802.11p medium ae that are inpired from the optional Meh eterminiti Ae (MA mehanim propoed for IEEE 802.11 [3]. Thi mehanim 978-1-4577-9538-2/11/$26.00 2011 IEEE 595
allow determiniti ae to the medium at eleted time to redue the poibilitie of olliion. MA aim to provide tringent MAC delay guarantee for real-time ervie uh a voie over IP (VoIP, whih i a ondition that an hardly be atified in laial IEEE 802.11 tandard. The MA heme [3-5] etend the IEEE 802.11 medium intantaneou reervation proedure, alo nown a the virtual arrier ening (V-CS, to a more advaned reervation proedure uing heduled MA OPportunitie (MAOP within a two-hop neighborhood. MAOP are firt negotiated between neighboring node by ehanging broadat etup meage, then, MAOP reervation are performed in multiple of a time-lot unit, during the elivery Traffi Indiation Meage (TIM periodi interval. To limit the meage broadat ignaling overhead, MA-related meage are ent only within two-hop neighborhood. It i worth noting though that while the MA heme i nown to redue to a ertain etent the delay bound, it la the onept of differentiating frame with different prioritie. Baially, MA provide a hannel ae with equal probabilitie for all tation ontending for the determiniti ae in a ditributed manner. However, equal ae probabilitie are not deirable among afety meage with different priorie. Some reent tudie a in [7] how that enforing MA with an effiient adjutment of the Contention Free Period (CFP allowing differentiation between different lae of ervie outperform the Enhaned itributed Channel Ae (ECA in term of delay and paet lo probability for IEEE 802.11. Our ontribution, in thi paper, an be ummarized a follow: (1 we firt introdue and jutify the adaptation of the meh determiniti medium ae named MA to redue paet olliion in IEEE 802.11p; (2 we improve and adapt MA in the ontet of vehiular afety ommuniation with two level of afety ervie overing mot of the poible afety appliation; we all the new heme VA; (3 we derive analytially the orreponding epreion of the periodiity and VA opportunity (VAOP duration in order to guarantee tringent delay bound for afety meage; (4 we tae into aount the vehile in the Carrier Sening Range (CSR to guarantee that none of thee vehile tranmit/ontend with the ender in order to enure a high paet reeption rate and a low olliion a poible; and (5 we evaluate our model ompared to tandard 802.11p/CF in term of delay, throughput and paet reeption rate for both routine and emergeny afety meage. The remainder of the paper i organized a follow. Setion II preent the motivation behind the integration of determiniti ae in IEEE 802.11p. Setion V propoe our heme named VA and preent a mathematial formulation of the ey parameter. Setion IV evaluate the propoed olution via imulation. Finally, Setion V onlude the paper. II. MOTIVATION FOR THE USE OF ETERMINISTIC ACCESSS FOR IEEE 802.11P When upporting afety appliation over SRC/802.11p we have to tae into aount trit requirement on low olliion and delay, epeially for emergeny meage uh a Forward Colliion Warning (FCW or Eletroni Emergeny Brea Light (EEBL whih require trit delay bound; otherwie many enviioned future afety ytem would be uele to help the driver deal with emergeny ituation, avoid aident and ave live. The main point that motivate u to onider/adapt a determiniti ae uh MA in IEEE 802.11p are a follow: (1 Mot of afety meage are baed on diret or ingle hop broadat ommuniation among vehile within the tranmiion range of one another. Thi i jutified by the fat that if an emergeny meage happen, the vehile potentially affeted are thoe whih are loe to the ender. Therefore, diret ommuniation i enough to reah potentially affeted vehile. MA i proven [3] to be more effiient within two hop range than laial CF/ECF, and to guarantee a hort delay. (2 In a low-load ondition, where olliion are very rare, CSMA provide lower delay than MA ine the former tranmit almot intantaneouly in a random time lot. In a low-load ondition, MA ha a lightly higher delay than CSMA primarily due to the problem of non-ontiguoune of the reerved time-lot. MA wait longer period before being able to tranmit in peifi reerved ontiguou time lot. However, in high-load ondition, the delay with MA i bounded by *TIM [6]; being the imum number of hop in a path (=1 for broadat meage. The delay provided by CSMA inreae without any bound with the inreae of the offered load. Thi i beaue many more node are ontending for the ame hannel, auing many more olliion and reulting in both longer binary eponential baoff and more frequent MAC retranmiion. Therefore, it i intereting to invetigate/adapt a determiniti ae uh a MA over IEEE 802.11p to tae advantage of the bounded delay guarantie it offer. (3 Vehile afety ommuniation networ are entirely ditributed ad ho wirele networ, and MA i a ditributed determiniti medium ae. III. PROPOSE VEHICULAR ETERMINISTIC ACCESS SCHEME: VA A. Current IEEE 802.11p Communiation Sheme IEEE 802.11p adopt IEEE 802.11a layer peifiation with minor modifiation. Thi i a random ae heme for all vehile loated in the tranmiion range of the ender baed on CSMA/CA. IEEE 802.11p ue CSMA/CA with ECA a in IEEE 802.11e or CF a in IEEE 802.11a and alo ue four prioritie queue with different Baoff and AIFS parameter. Neverthele, the Baoff proe with ECA involve high probabilitie of olliion, epeially in high offered-load ondition. There are two type of afety meage: emergeny afety meage (M e and periodi beaoning (or routine: M r afety meage. While emergeny meage happen only oaionally and require very high reliability, le olliion and hort delay, routine meage are broadated by all vehile at a frequeny of 10-20 time per eond. Routine meage ontain the tate of a vehile uh a it poition and 596
diretion and they require low reliability and long lateny ompared to M e [2]. In fat, one of the main onern about 802.11p, i how it will perform when SRC devie will be largely adopted, maing high-offered-load ondition very liely in dene vehiular traffi ituation, while having ontinuou routine meage beaoning haring the medium with more urgent life-ritial event-driven emergeny meage. B. Introduing VA Proe in IEEE 802.11p VA heduling i baed on MA onept; therefore, we tart by introduing MA before going into detailing our propoed heme VA to how what we added and modified in bai MA. In bai MA [3], the time between oneutive TIM beaon frame i divided into time lot of length 32 µ. The periodi broadat of beaon frame to all radio in the ame tranmiion range allow the ynhronization of thee TIM interval. Initially, node reerve the wirele medium for MAOP, whih are reerved a multiple of time-lot during a given Contention Free Period (CFP of a imum ae fration (MAF= α T of the TIM interval T (ee Fig.2. The remaining part of the TIM interval, a illutrated in Fig. 3, i the ontention period (CP ued for throughput-enitive rather than delay enitive data appliation (it ould be ued in the ontet of VANET for eample for private ervie meage, M p. Note that MA doe not upport different ervie with different prioritie and ha the ame behavior for all ervie meage in the networ. The meage type illutrated in Fig 2 rather refer to VA heme. We haraterize eah MAOP (in MA /VAOP (in VA reervation requet for meage by the triplet < O, π,, δ > where O i the VAOP offet from the N TIM tart period, TIM period, and Π δ i the VAOP periodiity within the i the VAOP duration in number of time-lot. Π i the number of time the peified VAOP repeat themelve equiditantly within a TIM interval (T. In fat, all vehile in the ame tranmiion range are aware of the reervation hedule due to the broadat of VA advertiement meage by the VAOP requeter node and the granter node [3]. In VA heduling, δ i the number of time-lot reerved for afety meage of type (ee Eq. 1 in eah of the Π (ee Eq. 2 ub-interval that atifie a hard ontraint on a imal delay for a imum number of hop m in a path. We aume that M { M e, M r } where M repreent the afety meage of type ; being equal to e if it i an emergeny meage, r otherwie (i.e. routine meage. We note that thi tranmiion our after duration AIFS. To prevent eeeding the one-hop delay, the periodiity Π in the VA reervation requet ha to be uffiiently lower bounded by: Π T. For the / ae of impliity, we onider a uniform ditribution of over interfering lin even though a better repartition may tae into aount the non-uniformity of traffi load over thee lin. Thu, the VAOP duration (Eq. 1 and periodiity (Eq.2 are epreed a follow: δ AIFS = τ L + C N N and Where τ i the time-lot duration, L i the paet ize (inluding PHY and above, tranmiion rate, N (1 C i the IEEE 802.11 i the number of meage of type and i a imal delay for meage omputed in Eq.3. TIM T Π M = = Fig.2 how the detail of VA funtionality in the preene of M e and M r in the CFP. VA etablihe priority between both afety meage and partiularly, VA prioritize M e over M r. VA alo erve private meage in the CP period beaue uh meage are not delay-enitive. It i worth noting that the tandard multi-hannel withing operation in WAVE allow the CCH and SCH interval to be different, a long a their total length i the TIM interval. We then define the dwell-time ratio a the time-perentage between CCH and SCH interval (e.g., we ould have 75% CCH well and 25% SCH well. Fig.2 VAOP hedule for emergeny (M e and routine (M r meage in VA C. well Time-Ratio in VA In VA, we ue CCH=CFP Interval (ICCH=ICFP and SCH=CP Interval (ISCH=ICP. A mentioned before, WAVE allow CCH and SCH to be different, jut a long a the length of the Synhronization Interval (ISynhroniation = ICCH+ISCH whih i in our ae equal to TIM interval. We aume that the TIM i a divior of 1e. The ICFP and ICP an be dynamially adaptable in VA heme.. Paet Tranmiion elay in VA We define the delay a the um of the ervie and queuing delay. The ervie delay i the um of the VA heduling delay, the AIFS and the tranmiion delay of the paet. We define VA heduling delay a the waiting time of the net (2 597
paet to be ent, for it reerved VAOP during whih it an tranmit without ontention. We aume that the baoff delay i negligible over a long period of time ine we aume that a ontention with other node i very rare during the reerved VAOP. And we define the queuing delay a the time a paet wait in the tranmiion queue. For emergeny meage, we are in the ontet of 1-hop broadat; eah broadat ha π1 paet to tranmit in every TIM interval. Then the ervie rate ould be epreed by π S 1 π m rate = for one-hop, otherwie for m-hop Srate =. TIM TIM M M = m The imal delay i denoted by (3, i.e., the hard ontraint on imal delay for a imum number of hop m in a path and M. i the required delay by the afety meage E. Probability of reeption rate in VA The paet reeption rate i defined a the ratio of the number of paet uefully reeived to the number of paet tranmitted. The paet reeption rate an be een a the probability that all vehile within the tranmiion range of the ender vehile reeive the broadat afety meage M uefully. We denote thi probability P. Fig.3 Senario We aume that vehile are plaed on the line (ee Fig.3 aording to Poion proe with networ denity β (vehile/m [2]. We an epre the probability to have v vehile per tranmiion range R a follow: v 2β ( R (2βR e P( v, R = v! (4 And the probability to have N vehile per tranmiion range R -R denoted by C a follow where R i the arrier ening range: N 2 ( C (2 C C β β e P( N, C = N! (5 Where N i the number of vehile that ould ontend for the ame time-lot with the ender in it range R. The Probability of Reeption Rate ( P RR in tranmiion range R an be epreed a follow: PRR = P( X, R = PRR ( S, R PX R ( δ X, OX PX R ( δx, OX Therefore, we deribe two ae: X X RR (6 Cae1: N = 0 C PRR ( S, R = P( v, R P, R ( δ, O (7 Where P ( δ, O 1 ine the ender i the only owner of, R δ and offet O in it tranmiion range R. Sine we ue a determiniti ae in VA, we epet low olliion to be happening. The average number of vehile in R i equal to 2πR while in C it i N = 2πC. Cae2: N 0 PRR ( S, R = P( v, R P, R ( δ, O C ( δ, O (8 Where P S ( δ, O i the probability that none of the C vehile S in range C tranmit in the time-lot alloated to the ender vehile S in range R during the CFP period. Let u define firt P 0 a the probability that a vehile ha an event or a routine afety meage to tranmit. In order to ahieve determiniti ae for vehile S in it range R, we ompute the probability P S ( δ, O, that none of the vehile C in C range ( S' NC tranmit with number of time-lot δ from the offet O. Proof: To formally epre P S ( δ, O, we applied a tandard C tehnique of proof by ae to define P S ( δ, O. We epre C firt the bae ae of thi probability for CFP equal to 2 lot with N 2 (ee Equation.8 and for CFP equal to 3 lot N 3 (ee Equation.9. For CFP=2 lot; N 2 2 (1 P0 C ( δ, O = 2 2 1 0 2 A P 0 + A P0 (1 P0 + A (1 P0 N N N For CFP=3 lot; N 3 3 (1 P0 C ( δ, O = 3 3 2 2 1 2 0 3 A P0 + A P0 (1 P0 + A P0 (1 P0 + A (1 P0 N N N N Then imilarly, For CFP=K lot and following equation: And for N K K (1 P C O = 0 (δ, K K A P P0 N 0 (1 = 0 N K, K (1 P C = O 0 (δ, N N K N A P P P 0 0 N 0 (1 (1 = 0 IV. SIMULATION RESULTS (9 (10 we epre the (11 (12 In thi etion, we ondut a imulation tudy uing n-2 to evaluate and ompare the performane of our propoed heme, i.e., VA, with the eiting heme baed on 802.11p CF. We evaluate everal metri: 1 the end-to-end delay; 2 the outage probability; 3 the paet reeption rate and 4 the average delay. The end-to-end delay i involved when afety 598
related meage need to be relayed to other vehile in a multi-hop manner (e. pot-rah meage. The outage probability i defined a the ratio of the number of vehile eperiening paet loe higher than the given threhold to the total number of vehile in the VANET. The paet reeption rate i rate of meage reeived within a one-hop range. The average delay i the average delay within a one-hop range. A. Simulation Configuration We ue a topology ompoed of 80 vehile with 10 vehile in eah lane in an 8 lane highway (4 lane/ diretion The radio tranmiion range r tae one of the following value: 150m, 200m and 250m and the tranmiion interferene R of eah vehile i 550m. Alo, we fi the paet ize to 1000Byte. The parameter are preented in Table I. Table I. Sytem Parameter PHY radio model SINR Carrier Sene Range 550m Tranmiion range 150m, 200m, 250m TIM 32m Threhold paet lo 5% α 0,68 well time-ratio 50% CCH well Time lot 20 µ MAC type 802.11 (ued with SRC Channel bandwidth [Mbp] 6, 9,12,24 Traffi type CBR (UP Period of meage 100 diemination [m] Meage payload ize [byte] 1000 Number of vehile 80 Speed [m/h] 100 Traffi denity [veh/m/lane] 10 Number of lane 8 Simulation time [e] 60 The end-to-end delay with VA in imulation doe not eeed 390m; it i bounded by the TIM interval, whih i equal to 32m, multiplied by the imum number of hop in a path that i equal to 10 in our topology. Wherea the delay provided by CF inreae without any bound with the inreae of the offered load. For eample, the delay with CF reahe 725m at 8 Mbp and it reult in many more vehile ontending for the ommuniation hannel, auing many more olliion and reulting in both longer baoff and more frequent retranmiion. The delay improvement (for all load i about 40% for VA ompared to CF. Fig.4. The end-to-end delay of both VA and CF. 2 The outage Probability Study B. Reult Analyi 1 The elay Study We tudied the performane of ae method CF and VA when tranmitting data on the hared hannel. We ditinguih between low and high offered load ondition: In light offered load ondition (0.05 Mbp-0.5 Mbp, where olliion are very rare, CF ae method provide lower delay, a hown in Fig. 4, ine it tranmit almot intantaneouly in a random time lot no later than 0.68m. In low offered load ondition, VA wait longer period before tranmitting in a peifi reerved ontiguou time lot. Thi i beaue VAOP annot be heduled to tart until the end of a TIM period of 32m. Thi heduling i performed regardle of the abene of interferene and even if earlier time lot are available, ine it need ontiguou available time lot to tranmit paet. Therefore, the average ae delay i higher with VA ompared to that of CF when the offered load i low. However, one hould note that in low offered load ondition, delay are very low both in CF and VA, and the etra delay introdued by VA i very low. In high offered load ondition (0.75Mbp-12Mbp, VA outperform CF and dereae the delay by a fator of two (i.e., VA- 0.34 and CF-0.7 in average. Fig.5. The outage ratio of both VA and CF. Fig.5 how the outage for both VA and CF method when varying the offered load. VA outage i muh lower ine the VA heme allow vehile to ae the wirele medium at eleted time with a lower ontention than would otherwie be poible within two-hop neighborhood by the laial 802.11p CF heme. Therefore, VA outage whih i related diretly to meage loe will obviouly outperform CF method. It preent an improvement of 46% over all load a hown in Fig.5. In the following et of eperiment we fied the offered load to 2Mbp, and varied vehile denity in order to ae the paet reeption rate and the average delay in a one-hop range when many vehile are ontending for the medium in dene vehiular enario both for CF and VA method. 3 The Paet Reeption Rate Study Fig.6 learly illutrate the differene of reult between CF and VA method. CF method uffer from inevitable olliion. Therefore, it ha a ignifiant drop in reeption 599
probability. VA enhane heduling. The fat that VA tae into aount the vehile in the arrier ene range to guarantee that none of them ontend with the ender enure high paet reeption rate and low delay. VA outperform CF by 42% in term of reeption probability. We note that in average for all denitie, VA reeption probability equal 0.78 and for CF it equal 0.44. Fig.6. The paet reeption rate of both VA and CF. 4 The elay of Emergeny and Routine Meage Study Fig.7. The delay of both emergeny and routine meage in VA heme. From vehile afety point of view, it i ruial for vehile in the highway to reeive tatu update (routine meage from eah neighboring vehile in the tranmiion range frequently enough and in an evenly timed manner. For eventdriven meage (emergeny meage, the tranmiion delay requirement are even more trit. That i why it i very ueful to have an effiient heduling heme uh a VA that provide lower tranmiion delay epeially for emergeny afety meage. Fig.7 how the delay of both emergeny and routine afety meage when varying vehile denity. From the figure, we an ee that, on the one hand, VA enure a very low delay for all denitie that i le than 0.00112616. Thi i very deirable ine emergeny meage uually involve urgent life-ritial ituation. On the other hand, routine meage have an average delay equal to 0.33 whih i higher than the delay of emergeny meage but good enough for routine meage. Thi i the epeted behavior from VA, ine the heme prioritize emergeny meage over routine meage when heduling VAOP a hown in Fig. 2. V. CONCLUSIONS AN FUTURE WORK In thi paper, we how how we minimize ontention between high-priority afety-oriented routine or emergeny traffi and non-afety appliation traffi uing a determiniti ae method over 802.11p alled VA. VA provide bounded delay and low loe partiularly for emergeny meage. Uing imulation, we how that the propoed approah, integrating determiniti ae, outperform CF and ahieve good performane in term of delay and paet reeption rate. Currently, we plan to invetigate a mehanim that prevent interfering vehile not integrating VA with 802.11p, from aeing the heduled VA opportunitie (VAOP and ubequently the hared medium during reerved time-lot. VI. ACKNOWLEGEMENT Thi wor i part of CooPerCom, an international reearh projet (Canada-Frane. The author would lie to than the National Siene and Engineering Reearh Counil (NSERC of Canada for upporting thi wor. REFERENCES [1] H. Su and X. Zhang, "Clutering-Baed Multihannel MAC Protool for QoS Proviioning Over Vehiular Ad Ho Networ," IEEE Tranation on Vehiular Tehnology, vol. 56, pp. 3309-3323, 2007. [2] M. iaomin, C. Xianbo, and H. Refai Hazem, "Performane and Reliability of SRC Vehiular Safety Communiation: A Formal Analyi," EURASIP Journal on Wirele Communiation and Networing, vol. 2009, Artile I 969164, 13 page, 2009. doi:10.1155/2009/969164. [3] G. R. Hiertz, S. Ma, Y. Zang, T. Junge, and. enteneer, "IEEE 802.11 MAC fundamental, " in Pro. IEEE MehTeh, pp. 1 8, 2007. [4] G. R. Hiertz, S. Ma, T. Junge,. enteneert, and L. Berlemann, "IEEE 802.11 Meh determiniti ae, " in Pro. of 14th EW, pp. 1 8, 2008. [5] C. Cionetti, L. Lenzini, and E. Mingozzi, "Sheduling and dynami reloation for IEEE 802.11 meh determiniti ae, " in Pro. of IEEE SECON, pp. 19 27, 2008. [6] J. Rezgui, A. Hafid, and M. Gendreau, "itributed Admiion Control in Wirele Meh Networ: Model, Algorithm, and Evaluation," IEEE Tranation on Vehiular Tehnology, Volume: 59, Iue: 3, 2010. [7] J. Rezgui, A. Hafid, and M. Gendreau, "Admiion Control and QoS Proviioning in Multi-Servie MA in IEEE 802.11-baed Wirele Meh Networ," in Pro. of ISCT Conferene on Heterogeneou Networing for Quality, Reliability, Seurity and Robutne (Qhine, 2010. [8] B. Fan, et al., "Toward haraterizing and laifying ommuniationbaed automotive appliation from a wirele networing perpetive", in Pro. of IEEE Worhop on Automotive Networing and Appliation (AutoNet, 2006. [9]. Jiang, et al., "eign of 5.9 ghz dr-baed vehiular afety ommuniation", Wirele Communiation, IEEE, 13, pp. 36-43, 2006. [10] W. Zhe and H. Mahbub, "How muh of dr i available for non-afety ue?", in Pro. of the fifth ACM international worhop on VehiulAr Inter-NETworing, pp.23-29, 2008. [11] C. Qi,. Jiang, and L. elgroi, "IEEE 1609.4 SRC multi-hannel operation and it impliation on vehile afety ommuniation", in Pro. of IEEE Vehiular Networing Conferene (VNC, pp. 1-8, 2009. [12] H. Ching-Ling, et al., "Adaptive intervehile ommuniation ontrol for ooperative afety ytem", Networ, IEEE, Vol. 24, pp. 6-13, 2010. [13] R.Mangharam, et al., "Bounded-Lateny Alert in Vehiular Networ", in Pro. of Mobile Networing for Vehiular Environment (MOVE, pp. 55-60, 2007. 600