A Hybrid (Active-Passive) Clustering Technique for VANETs

Clustering serves a vital role in the operation of Vehicular Ad hoc Networks (VANETs) by continually grouping highly mobile vehicles into logical hierarchical structures. These moving clusters help stabilize a global topology for message routing, and enable multi-channel operations: utilizing a long- range control channel for control data, and a short-range service channel for intra-cluster communications. Clustering techniques are partitioned in research into two categories: active and passive. Active techniques rely on periodic beacon messages from all vehicles containing location, velocity, and direction information. However, in areas of high vehicle density, congestion may occur on the long-range channel used for beacon messages limiting the scale of the VANET. Passive techniques use embedded information in the packet headers of existing traffic to handle clustering. In passive clustering, vehicles not transmitting traffic may cause cluster heads to contain stale and malformed clusters. This paper proposes a hybrid (active-passive) clustering technique, where the passive technique is used as a congestion control strategy when congestion is detected in the network. In this case, cluster members halt their periodic beacon messages on the control channel and embed position information in the packet header of traffic over the service channel to update the cluster head of their position. This multi-channel technique dynamically reduces the channel load of the control channel in urban VANET scenarios, increasing the scalability of the VANET. Simulation results show that the hybrid technique is effective at controlling congestion and reducing message delay in an urban VANET environment when compared to active clustering techniques.