近年,協調運転を実現するためにクラウドが広範囲に存在する車両からセンサー情報を収集・解析を行い,車両に制御情報を提供するクラウド型車両制御とMobile-Edge Computing (MEC) によるエッジサーバを用いて低遅延な車両制御を実現するMEC型車両制御が提案されている.我々はこれらの長所を生かすために,基地局に配置され低遅延な処理が可能なLower Edge Server (LoES)とP-GWやクラウドに配置され広範囲から車両のセンサー情報を収集可能なUpper Edge Server (UpES) が連携する車両制御システムを提案しており,プロトタイプを用いた評価から,遅延の大きさが不安定に変化する状況でも安定して制御できることを示した.
しかし, 提案システムは実通信環境を反映させた十分な評価を行えていない.現在,日本国内には様々なMNO, MVNO事業者が存在しており,これらキャリアの通信経路の特性も考慮し車両制御の評価を行う必要がある.

Recently, there have been proposed cloud-based vehicle control systems collect sensor information from vehicles and control them from cloud data center and MEC-based vehicle control systems that control vehicles utilizing ultra low latency. We have proposed a layered vehicle control system[1],[2] taking advantage of both cloud and MEC based systems where we enable coordinated control between Upper Edge Server (UpES) that aggregates sensor information from a large number of vehicles, and Lower Edge Server (LoES) that processes the sensor information in low latency feedback loop. The proposed system, however, has not been sufficiently evaluated using actual delay information in real world. In this report, we evaluate our proposed system using real latency data taken from various MNO and MVNO operators in Japan. In more detail, we measure the actual latency between various cloud services and UE and show that communication service quality is highly dependent on carriers and differs significantly. In addition, we simulate the proposed system using the measured latency and evaluate the effectiveness of our coordinated control. While there is inherent tradeoff between the number of vehicles to be optimized and the feedback latency in control, we show that by adjusting the location of UpES, our system can provide the optimized coordination between LoES and UpES mediating the tradeoff between the scope and the latency of vehicle control according to the requirements for the vehicle control.

佐々木健吾, 牧戸知史, 中尾彰宏. “協調運転実現のための多層エッジサーバによる車両制御システム ~ インターネット遅延計測と計測結果を用いた車両制御評価 ~”. 信学技報, vol. 117, no. 459, NS2017-234, pp. 375-380, 2018年3月. copyright©2018 IEICE