The increasing capability of 3GPP NR V2X standard for next-generation vehicular systems, will support Vehicle-to-Vehicle (V2V) communication in the Millimeter Waves (mmWaves) spectrum to address the communication requirements of future intelligent automotive networks. This new connectivity will enable the evolution towards Cooperative and Intelligent Transportation Systems (C-ITSs) with the aim of delivering improved traffic safety and efficiency. This concept is crucial because vehicles, even with advanced sensor systems, may not perceive every detail of their surroundings. By establishing connectivity, vehicles can enable a collaborative approach for object perception, where they can collectively enhance their awareness of the environment. In V2V, data sharing puts a strain on traditional vehicular communication technologies due to high demands in data rate, reliability, and latency. Researchers are exploring new radio systems, like at mmWaves, to address these challenges. However, the propagation at these frequencies raises many concerns. So, simply increasing channel capacity may not meet the demanding Quality of Service (QoS) requirements of future automotive applications, especially in scenarios with challenging automation levels. Therefore, it is crucial to limit the amount of data broadcast over bandwidth-limited channels. In this thesis, clustering-based algorithms are studied and analyzed to demonstrate how the burden on the network can be reduced exploiting NR V2V connectivity: first the vehicles are grouped into clusters and exchange information with the master of their cluster using a sidelink connectivity, then the master will communicate with the Base Station (BS), thus reducing the number of vehicles that are exchanging information and enabling the masters to select the information to be shared in a smart way. Through simulations, we demonstrate how the processing delay required by the approaches that select information to transmit in a smart way may be too high to meet the strict requirements of critical environment such as vehicular networks; hence, a random approach where what to transmit is selected in a random way by the masters of the clusters, performs better. Furthermore, we evaluate the potential of the proposed cluster-based dissemination algorithms as a function of several parameters, including the channel condition and the number of vehicles.
La crescente capacità di 3GPP NR V2X supporterà le operazioni veicolo-veicolo (V2V) nello spettro delle onde millimetriche (mmWave) per affrontare le esigenze di comunicazione delle future reti automobilistiche intelligenti. Questa nuova connettività consentirà la creazione di Sistemi di Trasporto Intelligenti Connessi (C-ITS) con l’obiettivo di migliorare la sicurezza e l’efficienza del traffico. Questo concetto è cruciale perché i veicoli, anche con sistemi avanzati di sensori, potrebbero non percepire ogni dettaglio del loro ambiente circostante. Stabilendo la connettività, i veicoli possono adottare un approccio collaborativo per la percezione degli oggetti, dove possono migliorare collettivamente la loro consapevolezza dell’ambiente. Nel V2V, la condivisione dei dati mette a dura prova le tecnologie tradizionali di comunicazione veicolare a causa delle elevate richieste di velocità di trasmissione, affidabilità e latenza. I ricercatori stanno esplorando nuovi sistemi radio, come le mmWave, per affrontare queste sfide. Tuttavia, i problemi di propagazione alle frequenze superiori a 6 GHz pongono ostacoli. Pertanto, aumentare semplicemente la capacità del canale potrebbe non soddisfare le esigenze di Qualità del Servizio (QoS) delle future applicazioni automobilistiche, soprattutto in scenari con diversi livelli di automazione. È quindi cruciale limitare la quantità di dati trasmessi su canali a larghezza di banda limitata. In questa tesi algoritmi basati su clustering sono studiati e analizzati per capire come ridurre il carico sulla rete sfruttando la connettività NR V2X: innanzitutto, i veicoli vengono raggruppati in cluster e scambiano informazioni con il capo del loro cluster utilizzando un collegamento laterale, quindi il capo comunicherà con la stazione base, riducendo così il numero di veicoli che scambiano informazioni e consentendo ai capi di selezionare in modo intelligente le informazioni da condividere. Attraverso simulazioni, dimostriamo come il tempo di elaborazione richiesto dagli approcci che selezionano in modo intelligente le informazioni da trasmettere potrebbe essere troppo elevato per soddisfare i requisiti rigorosi di ambienti critici come le reti veicolari; pertanto, un approccio casuale, in cui le informazioni da trasmettere vengono selezionate in modo casuale dai capi cluster, risulta più efficace. Inoltre, valutiamo il potenziale degli algoritmi di disseminazione basati su cluster proposti in funzione di diversi parametri, inclusi le condizioni del canale e il numero di veicoli.
Study and Analysis of Clustering-Based Algorithms for Cooperative Perception in Vehicular Networks
MEROTTO, ALBERTO
2023/2024
Abstract
The increasing capability of 3GPP NR V2X standard for next-generation vehicular systems, will support Vehicle-to-Vehicle (V2V) communication in the Millimeter Waves (mmWaves) spectrum to address the communication requirements of future intelligent automotive networks. This new connectivity will enable the evolution towards Cooperative and Intelligent Transportation Systems (C-ITSs) with the aim of delivering improved traffic safety and efficiency. This concept is crucial because vehicles, even with advanced sensor systems, may not perceive every detail of their surroundings. By establishing connectivity, vehicles can enable a collaborative approach for object perception, where they can collectively enhance their awareness of the environment. In V2V, data sharing puts a strain on traditional vehicular communication technologies due to high demands in data rate, reliability, and latency. Researchers are exploring new radio systems, like at mmWaves, to address these challenges. However, the propagation at these frequencies raises many concerns. So, simply increasing channel capacity may not meet the demanding Quality of Service (QoS) requirements of future automotive applications, especially in scenarios with challenging automation levels. Therefore, it is crucial to limit the amount of data broadcast over bandwidth-limited channels. In this thesis, clustering-based algorithms are studied and analyzed to demonstrate how the burden on the network can be reduced exploiting NR V2V connectivity: first the vehicles are grouped into clusters and exchange information with the master of their cluster using a sidelink connectivity, then the master will communicate with the Base Station (BS), thus reducing the number of vehicles that are exchanging information and enabling the masters to select the information to be shared in a smart way. Through simulations, we demonstrate how the processing delay required by the approaches that select information to transmit in a smart way may be too high to meet the strict requirements of critical environment such as vehicular networks; hence, a random approach where what to transmit is selected in a random way by the masters of the clusters, performs better. Furthermore, we evaluate the potential of the proposed cluster-based dissemination algorithms as a function of several parameters, including the channel condition and the number of vehicles.File | Dimensione | Formato | |
---|---|---|---|
Merotto_Alberto.pdf
accesso aperto
Dimensione
1.56 MB
Formato
Adobe PDF
|
1.56 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/72866