Enhanced stability of cluster-based location service mechanism for urban vehicular ad hoc networks

Abstract

Vehicular Ad Hoc Networks (VANETs) are gaining tremendous research interest in developing an Intelligent Transportation System (ITS) for smart cities. The position of vehicles plays a significant role in ITS applications and services such as public emergency, vehicles tracking, resource discovery, traffic monitoring and position-based routing. The location service is used to keep up-to-date records of current positions of vehicles. A review of previous literatures, found various locationbased service mechanisms have been proposed to manage the position of vehicles. The cluster-based location service mechanisms have achieved growing attention due to their advantages such as scalability, reliability and reduced communication overhead. However, the performance of the cluster-based location service mechanism depends on the stability of the cluster, and the stability of the cluster depends on the stability of the Cluster Head (CH), Cluster Member (CM) and cluster maintenance. In the existing cluster-based location service schemes, the issue of CH instability arises due to the non-optimal cluster formation range and unreliable communication link with Road Side Unit (RSU). The non-optimal cluster formation range causes CH instability due to lack of uniqueness of Centroid Vehicle (CV), uncertainty of participating vehicles in the CH election process and unreliability of the Cluster Head Election Value (CHEV). Also, the unreliable link with RSU does not guarantee that CH is stable with respect to its CMs and RSU simultaneously. The issue of CM instability in the existing cluster-based location service schemes occurs due to using instantaneous speed of the CH and fixed CM affiliation threshold values. The instantaneous speed causes the CM to switch the clusters frequently and fixed CM affiliation threshold values increase isolated vehicles. The frequent switching of isolated vehicles augment the CM instability. Moreover, the inefficient cluster maintenance due to non-optimal cluster merging and cluster splitting also contributes to cluster instability. The merging conditions such as fixed merging threshold time and uncertain movement of overlapping CHs within merging threshold time cause the cluster instability. Furthermore, the unnecessary clustering during cluster splitting around the intersection due to CH election parameters also increases cluster instability. Therefore, to address the aforementioned cluster instability issues, Enhanced Stability of Cluster-based Location Service (ESCLS) mechanism was proposed for urban VANETs. The proposed ESCLS mechanism consists of three complementary schemes which are Reliable Cluster Head Election (RCHE), Dynamic Cumulative Cluster Member Affiliation (DCCMA) and Optimized Cluster Maintenance (OCM). Firstly, the aim of the RCHE scheme was to enhance the stability of the CH through optimizing the cluster formation range and by considering communication link reliability with the RSU. Secondly, the DCCMA scheme focussed on improving the stability of the CMs by considering the Cumulative Moving Average Speed (CMAS) of the CH and dynamic CM affiliation threshold values, and finally, the OCM scheme enhanced the cluster stability by improving cluster merging conditions and reducing unnecessary clustering in cluster splitting. The results of the simulation verified the improved performance of the ESCLS in terms of increasing the location query success rate by 34%, and decreasing the query response delay and localization error by 24% and 35% respectively as compared to the existing cluster-based location service schemes such as HCBLS, CBLS and MoGLS. In conclusion, it is proven that ESCLS is a suitable location service mechanism for a wide range of position-based applications of VANETs that require timely and accurate vehicle locations.