Research On Vehicle-active Methods For Traffic Status Detection In VANETs

Vehicular Ad-hoc Networks (VANETs) is a significant component of intelligenttransportation system. In recent years, with the rapid development of both vehicularwireless communication technology and sensor techniques, research of intelligenttransportation technology has become a hot issue in the Internet of Things. VANETsprovide not only promotion for vehicles in capability of proactive safety and drivingassistance, but also convenient guidance for drivers with which they may avoidcongest road section and thus achieve destinations more expediently by makingreasonable plans for the routes. Timely and accurate detection must go ahead inorder to establish a quick and easy traffic guidance system. VANETs facilitate thevehicle-to-vehicle and vehicle-to-RSU (Road Side Unit) communication, providingnew technology support for detection of urban traffic condition. This paper performsstudies focusing on the vehicle-active detection oriented to road condition inVANETs and gives further discussions about the data transmission and detectionscheme. The main achievements are summarized as follows.First, vehicle-active broadcast methods oriented to traffic condition detectionunder the schema of vehicle-to-RSU communication are discussed and a restricted2-hop broadcast method is proposed to complement the disadvantages of1-hopbroadcast and2-hop broadcast. On the basis of Nakagami wireless transmissionpattern, this novel method restricts the field in which the status messages arerebroadcast and relieves the redundant messages, thus effectively extends the RSUdetection range from300metres to450metres with no less than75%T-windowdependability. Moreover, a probability-based limited2-hop broadcast method isproposed due to the disadvantage of the limited2-hop broadcast which is unable todistribute the opportunity of rebroadcast reasonably. Compared with limited2-hopbroadcast, it successfully raises the average rebroadcast rate by8.5%on the300metres to450metres road section.Second, on consideration of the defects of GPSR routing protocol directly applied to traffic condition detection, a restricted multi-hop broadcast method isproposed. It is inspired by the greedy transmission of GPSR and a maximumtransmission range in introduced to make sure transmission reception rate is beyonda certain level. In comparison with restricted2-hop broadcast, road section extendsfrom350metres to400metres on which reception rate for RSU achieves no lessthan60%. In addition, two optimized strategies, instant piggyback (IP) and periodicpiggyback (PP), are proposed on basis of restricted multi-hop broadcast. Throughmathematical modeling analysis and simulation experiment comparison， it isconcluded that strategy of IP promotes a full use of wireless bandwidth resource forrestricted multi-hop broadcast，thus the RSU can receive more vehicle statusinformation and effectively extends the RSU detection range from450metres to550metres with no less than75%T-window dependability. However, strategy of2next-hop has advantage only in cases with plenty bandwidth.Third，to complement the defects of previous detection schemes that mostlybased on vehicle-to-vehicle communication, a vehicle-to-RSU-based scheme isproposed, in which the length of waiting queue at the crossroad (L) acts as asignificant indicator and a detection method with aid of tail vehicle’s activebroadcast is also presented. The precision of L can reach92.8%on condition thatthe average computational error is no larger than5metres. Capacity of crossroad isalso considered and a queue-position-based detection scheme is proposed，which canaccurately detect the queue length passing a crossroad in a green light duration (Δl).Based on the vehicles kinematic model, computing method for the maximumcapacity of crossroad (Max (Δl)) is further discussed. On the basis of L, Δ landMax (Δl), traffic index (α) is introduced，which can clearly reflect traffic conditionat the crossroad and lay good foundation for the next generation intelligent vehicularnavigation system.Fourth, considering the disadvantage that exiting detection schemes need allvehicles participation, a scheme based on the active broadcast of buses is proposed,in which buses act as probes. Under the scheme of V2R, computing method for timeof passing detected road section for buses (Δt) is discussed, and current roadcondition can be gained according to Δ tand the evaluation system for level of road service. To cope with the impact of individual factor and external environmenton detection results, all buses passing the detected road section in a certain durationshould be observed and the average passing time should be calculated. Twocalculating methods, T-window-average and N-window-average, are proposed. Thesimulation results show that T-window-average achieves satisfying timeliness. Whenthe density of traffic flow for buses reaches80per hour and window size is set to5minutes, the average delay is only2.5minutes. While N-window-average gainsbetter universality, for example, RSU can collect status information from7buses in5minutes if the density of traffic flow for buses maintains80per hour. The schemeproposed can objectively reflect the condition of detected road section and providestrong support of dynamic data for future intelligent bus dispatching system.