A simulation study published in Peer-to-Peer Networking and Applications reports that a four-stage Candidate Cluster Head Nomination Algorithm significantly improves the structural stability of Vehicular Ad Hoc Networks while curbing the volume of control signaling required during network reorganization. Designed to balance scalability and low communication overhead, the proposed technique operates primarily under a single-cluster-head architecture but introduces an inactive candidate cluster head designated to take over immediately if the primary coordinator malfunctions or leaves the transmission vicinity. In numerical evaluations simulating multi-lane highway traffic, the protocol reduced the average number of cluster heads by one to three compared with established single-head schemes and cut reorganization message exchanges by up to six packets, demonstrating an effective approach to managing high-speed dynamic topologies without continuous dual-coordinator overhead.
Vehicular Ad Hoc Networks represent a specialized division of Mobile Ad Hoc Networks tailored for Intelligent Transportation Systems, driver assistance, passenger comfort, and road safety applications. Within these dynamic mobile environments, vehicles routinely establish two fundamental communication modes, comprising Vehicle-to-Infrastructure exchanges through roadside units and direct Vehicle-to-Vehicle transmissions between neighboring cars using dedicated short-range communication equipment. Unlike conventional ad hoc networks, vehicles traveling on highways move at considerable speeds, generate rapid shifts in topology, and are constrained to exchange mission-critical safety and navigation data within very narrow operational timeframes.
Traditional clustering models partition mobile nodes into clusters managed by designated cluster heads, which coordinate communication on behalf of surrounding cluster members. Most existing protocols implement a single-cluster-head architecture, which effectively suppresses cluster proliferation, shortens reorganization routines, and limits message overhead across the network.
To overcome the architectural shortcomings of both conventional paradigms, the research team structured the Candidate Cluster Head Nomination Algorithm into four distinct operational phases: a Cluster Head Nomination Stage, a Cluster Head Pruning Stage, a Candidate Cluster Head Selection Stage, and an adaptive Cluster Maintenance Stage. Rather than employing complex weighted coefficients that combine velocity, positioning, and acceleration into a rigid mathematical index, the nomination sequence evaluates specific vehicle parameters in a clear column-based priority order. Vehicles first initialize as unknown nodes and broadcast local coordinates and velocity metrics across their single-hop neighborhood. By examining localized data, the system prioritizes candidates exhibiting the lowest standard deviation of speed relative to nearby vehicles, followed by the greatest neighbor density, the shortest average spatial distance, and an unambiguous low vehicle identification tie-breaker.
The secondary pruning phase addresses overlapping transmission zones where adjacent groups of vehicles inadvertently elect duplicate cluster heads. During pruning, redundant coordinators in mutual radio range compare their structural parameters; the coordinator governing the larger neighbor count retains its leadership role, while inferior heads surrender leadership and revert to ordinary cluster member status. If vehicle counts match, the algorithm preserves the node displaying the lowest average speed standard deviation, applying identification numbers only as a final deterministic arbiter. By actively weeding out superfluous coordinators before finalizing cluster boundaries, the pruning routine ensures that vehicle groups consolidate around the most representative mobile nodes, establishing a lean single-head baseline that suppresses transmission clutter across shared wireless channels.
Once clusters stabilize under individual primary coordinators, the algorithm triggers the candidate selection phase to establish operational durability without incurring active synchronization overhead. The primary cluster head conducts a targeted evaluation within its active membership to appoint a single passive candidate cluster head, choosing the node with the lowest speed deviation, maximum neighbor connectivity, and minimal spatial separation. Unlike dual-head protocols such as the Fuzzy Logic-Based Algorithm or Double-Head Clustering, which maintain two active communicating heads with higher operational complexity, the nominated candidate remains entirely passive during normal routines. Operating essentially as a standard cluster member, the candidate consumes no continuous coordination bandwidth and activates only when primary leadership fails or departs the local neighborhood.
To sustain connectivity across evolving highway conditions, the final maintenance stage executes adaptive response protocols tailored to four specific field occurrences. First, newly arrived unknown vehicles broadcast join requests containing current speed indicators; they enter clusters whose heads exhibit matching speeds or create independent single-vehicle clusters if no coordinator replies. Fourth, neighboring clusters traveling at comparable speeds that overlap for a predefined threshold duration merge, provided the aggregate vehicle count remains below specified capacity limits.
The researchers verified the protocol through extensive MATLAB simulations modeling high-speed highway scenarios across a three-lane unidirectional corridor. Experiments deployed mobile populations of 30, 80, and 130 vehicles traveling at randomized velocities between 90 and 120 kilometers per hour, testing node transmission ranges configured to 100, 200, and 300 meters over 300-second simulation runs. Performance evaluations benchmarked the proposed algorithm against two established high-speed single-head algorithms, namely the Unified Framework of Clustering and Center-Based Stable Clustering, alongside internal ablation variants lacking either candidate nomination or maintenance mechanisms. To rigorously isolate the fundamental efficiency of cluster formation and structural retention, the investigation concentrated squarely on physical clustering metrics—specifically cluster head counts, cluster duration, and control signaling packets—intentionally setting broader network-layer metrics aside.
The resulting performance data confirmed that the nomination and pruning framework consistently formed fewer, better-consolidated clusters across both sparse and dense traffic conditions. At a 100-meter transmission radius, the algorithm decreased the average cluster head tally by one to three compared to both benchmark models, maintaining superior structural cohesion even when sparse distributions or restricted radio ranges challenged connectivity. In cluster lifetime assessments, the complete protocol achieved durability competitive with Center-Based Stable Clustering, differing by an average of approximately 12 seconds across varying conditions while closing the gap to just 1 second in specific sparse alignments. In sharp contrast, the ablation variant stripped of candidate heads suffered severe premature collapses, exhibiting lifetimes that were 68 seconds shorter on average and deteriorating by up to 83 seconds under dense traffic with short transmission ranges.
Analysis of control signaling demonstrated substantial reductions in operational packet overhead, establishing the algorithm’s communication advantage over competing frameworks. While both benchmark protocols required extensive continuous neighbor data collection and weighted mathematical processing that elevated network load, the candidate nomination method confined dense data gathering primarily to initial setup and targeted maintenance events. This targeted signaling curtailed control exchanges by up to six packets per reorganization episode. Although the authors note that fixed rule-based selections may repeatedly favor low-variance nodes under extreme mobility swings, the lightweight candidate design successfully achieved near-optimal cluster longevity, minimal coordinator numbers, and lower message burdens, establishing an effective foundation for future evaluations incorporating end-to-end packet delivery, latency, and edge-computing integration.
Subject of Research: Candidate cluster head nomination algorithm for vehicular ad hoc networks
Article Title: Increasing cluster stability in vanet by candidate cluster head nomination algorithm
Article References: Chiang, M.-L., Hsieh, H.-C., Tsai, W.-C., Lin, T.-L., & Lin, Y.-H. (2026). Increasing cluster stability in vanet by candidate cluster head nomination algorithm. Peer-to-Peer Networking and Applications, 19(4), Article 104. https://doi.org/10.1007/s12083-026-02263-1
Image Credits: AI Generated
DOI: 10.1007/s12083-026-02263-1
Keywords: vehicular ad hoc networks, clustering algorithms, cluster head selection, candidate cluster heads, intelligent transportation systems, network stability, message overhead, highway simulation, vehicle to vehicle communication, vehicle to infrastructure, dedicated short-range communication, speed standard deviation, cluster maintenance, MATLAB simulation, traffic density, connected vehicles
Cite Scienmag News
Reid Dalton. (November 21, 2026). Nomination Algorithm Stabilizes Vehicular Ad Hoc Networks by Reducing Cluster Heads. Scienmag. https://scienmag.com/nomination-algorithm-stabilizes-vehicular-ad-hoc-networks-by-reducing-cluster-heads/
Reid Dalton. "Nomination Algorithm Stabilizes Vehicular Ad Hoc Networks by Reducing Cluster Heads." Scienmag, 21 November 2026, https://scienmag.com/nomination-algorithm-stabilizes-vehicular-ad-hoc-networks-by-reducing-cluster-heads/. Accessed 21 November 2026.
Reid Dalton. "Nomination Algorithm Stabilizes Vehicular Ad Hoc Networks by Reducing Cluster Heads." Scienmag. November 21, 2026. https://scienmag.com/nomination-algorithm-stabilizes-vehicular-ad-hoc-networks-by-reducing-cluster-heads/








