混合式車載隨意感測網路之多維配置問題

Abstract

為了使車載隨意網路(VANETs)具有感測道路狀況的能力，混合式車載隨意感測器網路(Hybrid VANET -Sensor network; HVSNs)加入了大量配置的感測器，以提供安全的道路環境；在過往的網路配置研究當中，考量的問題大都過於簡化，且少有研究專注在HVSNs的配置問題之上，因此，為了考量實際情況上的需要，本研究分析在HVSNs中該如何配置路邊設施(RSUs)以服務三種不同維度道路上的感測器且滿足兩個目標(i.e., 最小化總傳送距離及最小化感測器至其最近RSU的總跳耀數)，並加入兩個QoS的限制，此兩個目標可有效地減少傳輸過程的干擾的機率，使傳輸時間更短，進而快速告知駕駛突發事件的發生。由於格網道路和稠密HVSN的最佳配置問題無法在有效時間內解決，因此本研究提出該問題的數學規劃模型，並進一步提出改進和聲搜尋演算法來解決該問題，最後並透過模擬在不同大小的實例和統計分析，證實所提出的方法具有前瞻性。

To enhance vehicular ad hoc networks (VANETs) with the ability to detect road conditions, hybrid VANET-sensor networks (HVSNs) additionally deploy a large number of wireless sensors on roads, so as to provide a safer driving environment. Previous works on network deployment problems often considered too simplified problem settings, and few of them focused on deploying HVSNs. To meet more practical requirements, this paper investigates the problem of deploying roadside units (RSUs) to serve sensors on three different-dimensional road settings in HVSNs with two objectives (i.e., to minimize the total distance and the total number of hops from sensors to their respective closest RSUs) and two QoS constraints. Achieving the two objectives can effectively decrease the probability of transmission interference and lead to shorter transmission time so as to inform drivers of unexpected events. Since the optimal deployment problems for a grid road and a dense HVSN cannot be solved efficiently, this paper creates mathematical programming models for the problems, and further proposes an improved harmony search algorithm. By simulation in different-size instances and statistical analysis, the proposed approaches show promising performance.