ACM Transactions on Autonomous and Adaptive Systems
https://dl.acm.org/journal/taas/editorial-board?fbclid=IwAR020SIc0limwvVpuufkIbYFncXwhkm9I40FrlpWKhY8woCeEBQTb8cnHKw
上記に「道路安全 自己適応システムの安全分析」を投稿すべくgithubで資料整理をはじめました。投稿に連名いただける方を募集します。6ヶ月以内を目標にしています。2ヶ月後にTOPPERSのコンテストに応募し、１年後にはIEEEに投稿予定です。その２つも連名者を募集しています。それぞれ、オープンソースで模擬試験をすることと、分析の結果に基づいた提言を想定しています。
arxivでRoad Safetyで検索したら544。参考文献に入れるのは100件未満に絞ろうとしました。読んで行くと絞れずに、参考文献の参考文献の一覧を作って篩にかけようと思いました。
https://github.com/kaizen-nagoya/autonomous_and_adaptive_system/tree/main
https://www.acm.org/publications/authors/submissions
LaTeX
Best Practices for Submitting a LaTeX Paper/Article
https://www.acm.org/publications/taps/latex-best-practices
交通安全
自動車の交通安全は、
道路安全
自動車安全
運転者安全
の３つに分類できる。
自動車安全
運転者安全
は、広く検討が進んでいる。
道路安全は、費用対効果がおおきくないと推測し、
進んでいないかもしれない。
以下のいくつかの資料では、道路安全に焦点を絞っていない。
自動車安全
自動車安全資料を仮訳してみた。分析はこれから。
https://qiita.com/kaizen_nagoya/items/c28c707ef226e65b780e
自動運転の安全分析
https://qiita.com/kaizen_nagoya/items/b7f4a714f6016d3c1bf3
仮説(２１）交通事故死を減らすのにプログラマが協力できる仮説13選
https://qiita.com/kaizen_nagoya/items/4d46bbf0dde44d7bb99a
子供向け自動車安全（ソフトウェアに何ができるか）
https://qiita.com/kaizen_nagoya/items/a6e922b6630cd05cd244
猫が安全な世界は、子供も安全。
仮説（25) 猫中心設計
https://qiita.com/kaizen_nagoya/items/d409c1185df30f660adc
仮説（24) 高齢者・障害者設計指針(1) 点字ブロック・手すり・階段
https://qiita.com/kaizen_nagoya/items/47e0b3a83c2e250fa80c
<この項は書きかけです。順次追記します。>
道路安全
道路安全は、
道路安全
歩行者安全
自転車安全
二輪車安全
乗用車安全
貨物車安全
乗合自動車安全
などに分類できる。
道路
交差点
信号交差点
信号のない交差点
中央分離帯
ガードレール
歩行者
歩道
横断歩道
自転車
自転車専用道路
二輪車
乗用車
貨物車
積載重量超過
積載方法不備
乗合自動車
乗員超過
考察
北海道と愛知県が交通死亡事故で１、２位を争っていた。
当時、北海道立工業試験場と一緒に自動車安全に関する事業を行なっていた。
残念ながら、事業の間には、１位、２位を脱却することはできなかった。
力不足だった。
情報産業における自治体の都道府県連携に関する１与件
https://qiita.com/kaizen_nagoya/items/01699284626c48b22e20
仮設類
道路を安全設計すると、非安全部分との境界の危険性が高まるかもしれない。
安全設計の不整合により、これまでにない危険性が確認できるかもしれない。
天敵による捕食行動が昆虫の繁殖力を増加させる
https://www.okayama-u.ac.jp/up_load_files/press_r3/press20210608.pdf
ある行為が、線形に有効になるのではなく、反作用をもたらし、逆に有効性がなくなる可能性を示している。
参考資料
安全（0）安全工学シンポジウムに向けて
https://qiita.com/kaizen_nagoya/items/c5d78f3def8195cb2409
自動運転資料集(1)
https://qiita.com/kaizen_nagoya/items/42eb2129e281f25eaab8
自動運転資料集(2)
https://qiita.com/kaizen_nagoya/items/6c5231e6030020c7307c
「ソフトウェアエンジニアが「トヨタ」の伝統を革新して、ソフトとハードの融合した「モノづくり」を推進する時代へ」Qiita Zine記事を読み解く
https://qiita.com/kaizen_nagoya/items/20a3144e0ccd4d3655b5
プログラマによるプログラマのプログラマの子供のための自動車安全絵本企画
https://qiita.com/kaizen_nagoya/items/0ab47d8fca2933f8877a
仮説(66)プログラマが安全な系のためにできること
https://qiita.com/kaizen_nagoya/items/a9667ab0d1e48438edba
安全分析においてHAZOPで想定外を洗い出すために
https://qiita.com/kaizen_nagoya/items/11f1ace6f4c150248903
「ワークショップ「ソフトウェア開発におけるHAZOP入門」の結果」の分類
https://qiita.com/kaizen_nagoya/items/e62e91cb019c6275d6c1
仮説(152）水素自動車用カーナビの基本仕様
https://qiita.com/kaizen_nagoya/items/fb3b0ee76dda9780b6da
自動車の電源・電池と計算機・通信
https://qiita.com/kaizen_nagoya/items/f749754c2c9a15d2b70e
プログラマが電池設計に寄与できるプログラムを書くために
https://qiita.com/kaizen_nagoya/items/73e44e4f1ebf161f58cc
模型駆動開発(Model Driven Design)への道
https://qiita.com/kaizen_nagoya/items/bb4d73bfb3cbba88727f
充電電池関連ソフトウェアおよび安全分析
https://qiita.com/kaizen_nagoya/items/26d898911c6e43c9e8e6
Motor Bench and Hardware In the Loop Simulator
https://qiita.com/kaizen_nagoya/items/3b5cf163663ad1d3b83e
電動機制御算譜（プログラム）設計における３つの罠６つの教訓（実機）
https://qiita.com/kaizen_nagoya/items/b39b6b7ba0d90dff471d
プログラマが電動機(electric motor)制御する際に陥る穴
https://qiita.com/kaizen_nagoya/items/1f26d75bb1a8fb0f9e1b
電動機故障診断(ACサーボモータを中心に)
https://qiita.com/kaizen_nagoya/items/756d19527d5f862e8085
bike informatics
https://qiita.com/kaizen_nagoya/items/13c16288aa6c5bcabbee
プログラマが安全工学シンポジウムで発表する動機、題材、技法
https://qiita.com/kaizen_nagoya/items/b7adf3001eb325166e52
安全工学シンポジウムへの投稿を１週間で作成する方法
https://qiita.com/kaizen_nagoya/items/181f2ced6d418de9c7a6
名古屋で自動車関係のソフトウェア設計する際にあるといい知識
https://qiita.com/kaizen_nagoya/items/9f01d55e4bd0bd931c96
@kazuo_reveさんの「自動車の故障診断に関連するプログラマーになりたての方が参照するとよさそうな情報」の読み方
https://qiita.com/kaizen_nagoya/items/0c6b8373f93ce52def33
「国土交通省自動車局 自動運転車の安全技術ガイドライン」を超えて
https://qiita.com/kaizen_nagoya/items/8d219f57dfd71b526c96
Bus Master
https://qiita.com/kaizen_nagoya/items/746144ef912ac3b07854
"SAFETY FIRST FOR AUTOMATED DRIVING" に追加するとよいかもしれないこと
https://qiita.com/kaizen_nagoya/items/0bab4a2c184c8fbfb0ef
自動車技術会 2020年春季大会 Summarized Paper 単語帳
https://qiita.com/kaizen_nagoya/items/758922c754be557571a4
自動車技術会学術講演会原稿執筆要領を電子ファイルで作成する際にすることと困ったこと
https://qiita.com/kaizen_nagoya/items/ef23e836132b5f289b51
自動車技術会投稿要項にもとづいたLaTeXファイルを作るには
https://qiita.com/kaizen_nagoya/items/c9e99ca1181e2a20d987
自動車のソフトウェアで解決できるリコールをIT業界の視点で推測してみる
https://qiita.com/kaizen_nagoya/items/7a45b2911649982fa4f5
QC検定に落ち「たか」らかける記事。20,000人の方に読んでいただけ「たか」ら書ける記事。「たかたか」分析の勧め。
https://qiita.com/kaizen_nagoya/items/2a371ee8c8f1b78cd5bb
arXiv
Adaptive Semi-Persistent Scheduling for Enhanced On-road Safety in Decentralized V2X Networks
https://arxiv.org/pdf/2104.01804.pdf
Avik Dayal1, Vijay K. Shah1, Biplav Choudhury1, Vuk Marojevic2, Carl Dietrich1, and Jeffrey H. Reed1 1Wireless@VT, Bradley Department of ECE, Virginia Tech, VA, USA
2Electrical and Computer Engineering, Mississippi State University, MS, USA
{ad6db, vijays, biplavc, cdietrich, reedjh}@vt.edu and vuk.marojevic@ece.msu.edu
Abstract—Decentralized vehicle-to-everything (V2X) networks (i.e., Mode-4 C-V2X and Mode 2a NR-V2X), rely on periodic Basic Safety Messages (BSMs) to disseminate time-sensitive information (e.g., vehicle position) and has the potential to improve on-road safety. For BSM scheduling, decentralized V2X networks utilize sensing-based semi-persistent scheduling (SPS), where vehicles sense radio resources and select suitable resources for BSM transmissions at prespecified periodic intervals termed as Resource Reservation Interval (RRI). In this paper, we show that such a BSM scheduling (with a fixed RRI) suffers from severe under- and over- utilization of radio resources under varying vehicle traffic scenarios; which severely compromises timely dissemination of BSMs, which in turn leads to increased collision risks. To address this, we extend SPS to accommodate an adaptive RRI, termed as SPS++. Specifically, SPS++ allows each vehicle – (i) to dynamically adjust RRI based on the channel resource availability (by accounting for various vehicle traffic scenarios), and then, (ii) select suitable transmission opportunities for timely BSM transmissions at the chosen RRI. Our experiments based on Mode-4 C-V2X standard implemented using the ns-3 simulator show that SPS++ outperforms SPS by at least 50% in terms of improved on-road safety performance, in all considered simulation scenarios.
Index Terms—Decentralized V2X, C-V2X, NR-V2X, Semi- Persistent Scheduling, Basic Safety Message, On-road Safety
I. INTRODUCTION
Vehicle-to-everything (V2X) 1 communications is a promis- ing technology for next generation of intelligent transportation systems (ITS), mainly due to its potential of improving on-road safety, leading to the prevention/reduction of road accidents and more efficient traffic management [1]. There are two com- peting technologies that enable V2X communications, Ded- icated Short Range Communications (DSRC) and Cellular- V2X (C-V2X) [2]. Dedicated Short Range Communications (DSRC) is a decentralized wireless technology based on the 802.11p standard [3]. On the other hand, 3GPP introduced C-V2X communications in Release 14, and standardized two modes of operation termed Mode 3 and Mode 4 based on the scheduling preferences. Mode 3 C-V2X employs a centralized scheduling approach under the coverage of eNodeB, where two vehicles can communicate directly. The selection of radio
This research is supported by the Office of Naval Research (ONR) under MURI Grant N00014-19-1-2621. This paper has been accepted for publication in IFIP Networking 2021. This is a preprint version of the accepted paper.
1V2X refers to vehicle-to-infrastructure (V2I), vehicle-to-vehicle (V2V), and vehicle-to-pedestrian (V2P) communications.
resources are managed by the control signaling from the cellu- lar infrastructure over the Uu interface (uplink and downlink) [4]. Mode 4 C-V2X adopts new PC5 interface for direct communication among vehicles without the need for coverage from the eNodeB [5]. Since the cellular connectivity can not be assumed ubiquitous, Mode 4 C-V2X is considered as the baseline mode for C-V2X. Recently, New Radio V2X (NR- V2X) has been envisioned in Release 16, and it is expected to have decentralized Mode 2a NR-V2X [6].
There are two major V2X use cases: (i) Cooperative safety applications that rely on basic safety messages (BSMs), which are periodic messages that contain critical safety information, e.g., sender vehicle’s position and speed, and (ii) Cooperative traffic efficiency messages, which are event triggered messages that are intended to help vehicular flow. Since BSMs carry time-sensitive information, BSMs enable cooperative safety applications, such as, forward collision warnings [7] and blind spot/lane change warnings [8]. This is mainly because BSMs facilitate accurate positioning or localization of neighboring vehicles. Outdated BSMs (due to large BSM scheduling inter- vals) and/or lost BSMs (due to channel congestion) negatively impacts the performance of safety applications as it leads to increased collision risky situations, mainly due to wrong localization of neighboring vehicles, or in other words, high tracking error. Tracking error can be defined as the difference between a vehicle’s actual and perceived location (via most recent BSM) by its neighboring vehicles. From the above discussion, it is evident that BSM scheduling (or time intervals at which BSMs are broadcasted) becomes a critical parameter for ensuring BSM timeliness and minimal channel congestion, and thus, improved on-road safety of vehicles.
In Mode 4 C-V2X, the BSM scheduling is achieved by uti- lizing sensing-based semi-persistent scheduling (SPS), where the vehicle sense the radio resources (or channel medium) and select suitable (unutilized or underutilized) radio resources for the transmission of BSMs at prespecified fixed time intervals, termed, Resource Reservation Interval (RRI). More formally, RRI can be defined as the inter-transmission time interval between two consecutive BSM transmissions by a vehicle. The value of RRI is usually set to 100 ms (equivalently, BSM rate to 10 Hz) in the C-V2X standard. The 3GPP standard allows other values of RRI, such as, 20 ms and 50 ms [2].
In this paper, we show that the SPS algorithm with fixed RRIs (e.g., 100 ms) has limitations in terms of improving on-
arXiv:2104.01804v1 [cs.NI] 5 Apr 2021
road safety performance of Mode-4 C-V2X networks. Con- sider the following instances: (i) high vehicle density scenarios – C-V2X networks would likely suffer from overly congested radio channels (and thus, large number of lost or dropped BSM packets), which will lead to increased tracking error and thus, degraded on-road safety, and (ii) low vehicle density – The radio resource are under-utilized in such low density scenarios. The on-road safety can be greatly improved by choosing lower value of RRIs (e.g., 20 ms), as lower value of RRIs will improve the timeliness of BSMs without compromising the channel congestion (as the channels can support lower RRIs for fewer vehicles). Furthermore, notice that due to vehicle mobility and other contextual factors, the vehicle densities and available channel resources may change over time even for a given C-V2X scenario. Motivated by this, we propose significant enhancements to conventional (fixed RRI) SPS, and develop an improved, adaptive BSM scheduling algorithm, termed SPS++. At any time slot, SPS++ allows each vehicle to adapt (or adjust) the RRI for a given network scenario (depending upon vehicle density around it), and thus, ensure timely BSM transmissions, which in turn, promises to improve on-road safety performance. Decentralized Mode 2a NR-V2X is expected to utilize SPS as the standard BSM scheduling protocol, thus, we anticipate that our SPS++ algorithm will also be applicable to the new NR-V2X standard [6].
In summary, the paper makes following key contributions:
• We show that the conventional SPS algorithm for Mode-4 C-V2X suffers severely from under- and over- provisioning of radio channel resources (depending upon the vehicle densities). This in turn negatively impacts the timely successful delivery of BSMs and thus, compro- mises on-road safety performance of Mode-4 C-V2X.
• To address the limitations of SPS, we propose an im- proved, adaptive SPS, termed SPS++, that ensures BSM timeliness and thus, improves on-road safety performance of Mode-4 C-V2X. SPS++ allows each vehicle to dynam- ically adjust RRIs at each time slot, and judiciously utilize radio resources for BSM transmissions while accounting for varying C-V2X vehicle traffic scenarios.
• To measure the on-road safety performance of C-V2X, we present collision risk model, which is based on the concepts of tracking error and time-to-collision (TTC).
• Our extensive experiments based on ns-3 simulator show that the proposed SPS++ greatly outperforms the conven- tional SPS in terms of road safety performance, measured as collision risk, across all considered C-V2X vehicle sce- narios. When compared to SPS with RRI as 20 ms, 50 ms, and 100 ms, the results show a significant improvement in the reduction of collision risk respectively by 51.27%, 51.20%, and 75.41% for 80 vehicles/km.
The rest of the paper is organized as follows: Section II discusses the related works. Section III discusses Mode-4 C- V2X and SPS algorithm. Section IV discusses on-road safety performance model. In Section V, we highlight the limitations of SPS in the context of on-road safety performance, and pro-
pose SPS++ algorithm. Section VI describes the performance evaluation, followed by concluding remarks in Section VII.
II. RELATED WORK
Recent works have simulated and proposed resource allo- cation improvements to Mode-4 C-V2X. Molina-Masegos et al. [2] and Chen et al. [9] develop and use simulations to assess the performance of the SPS algorithm in highway and city scenarios, and showed the improved performance of C- V2X over DSRC. Research in [10], [11], and [5] investigate the performance of parameters used in SPS such as RRI, probability of reselection, and selection window on the packet delivery ratio (PDR). Gonzalez-Martin et al. [12] build the first known analytical model of SPS based C-V2X. Recently, [13] proposed short term sensing before resource selection to help reduce packet collisions and improve SPS performance. Halder et al. [4] and Lee et. al [14] suggest adjusting the transmission power and RRI respectively, to improve the overall performance of SPS, but both only focus on improving the PDR performance of C-V2X.
As opposed to prior works which focus on communication- centric metrics such as, throughput and PDR, we focus on on- road safety performance of decentralized V2X networks, such as, tracking error and collision risks. It is worth mentioning that designing optimal beacon rate in case of DSRC for improved safety performance of DSRC based V2X networks has been explored fairly well [3], [15]–[18]. However, note that designing rate control algorithms for DSRC are fundamentally different than designing scheduling protocols for decentralized Mode-4 C-V2X networks, and existing rate control protocols can not be directly applied to improving on-road safety performance in our context of Mode-4 C-V2X networks. To the best of our understanding, this is the first work which investigates designing BSM scheduling protocol for improved on-road safety performance of Mode-4 C-V2X.
III. C-V2X MODE-4
In this section we discuss the Mode-4 C-V2X standard based on 3GPP Release 14. We start with the C-V2X Mode 4 physical layer and then detail the inner workings of sensing- based SPS that allows vehicles to find and reserve suitable transmission opportunities.
A. Physical Layer
At the physical layer, Mode-4 C-V2X is similar to the LTE uplink and uses single-carrier frequency division mul- tiple access (SC-FDMA). C-V2X supports 10 and 20 MHz channels at 5.9 GHz. In Mode-4, the time-frequency resources are divided into resource blocks (RBs), subframes (sf), and subchannels. Frames are 10 ms in length, and are comprised of 10 subframes. Subframes are typically comprised of 2 resource blocks in time [11]. A resource block is the smallest schedulable unit of time and frequency that can be allocated to users. Each RB is made up of 7 SC-FDMA symbols and is 180 kHz wide in frequency with 12 equally spaced subcarriers. Each RB occupies 0.5 ms in time, with a minimum allocation
time interval of 1 ms, also referred to as the transmission time interval (TTI). Subchannels are composed of resource blocks contiguously located in frequency. In C-V2X, the data packet to be transmitted, called the Transport Block (TB), typically takes up one or more subchannels. Each TB is transmitted in the same subframe along a sidelink control information (SCI), that contains the modulation and coding scheme (MCS) [5].
B. Semi-Persistent Scheduling (SPS)
candidate subframes (See Step 2). T1 ≤ 4 and T2 ≤ RRI are the start and end subframes for the selection window. The RRI refers to the time interval between two consecutive BSM transmissions (See Fig. 2(a) for illustration). Each vehicle utilizes Nsensing subframes (obtained in Step 1) for identifying and subsequently, selecting the available subframe within the selection window for BSM transmission as follows.
- The vehicle sets – (i) the RSRP threshold, Pth, to a minimum RSRP value, Pmin (Step 3) and (ii) ini- tializes set SA as all subframes in the selection win- dow, i.e., SA = [sfn+T1 , sfn+T1+1, · · · , sfn+T2 ], and SB as an empty set (See Step 5).
- As shown in Step 6, the vehicle excludes all candi- date subframes from set SA if one of the follow- ing conditions are met – (i) the vehicle has not monitored the corresponding candidate subframe in the sensing window (i.e., Nsensing) and (ii) the lin- ear average RSRP measurement for corresponding candidate subframe is higher than Pth. The RSRP exclusion criteria for the ith subframe (for jth sub- channel) in the selection window can be written as
10 to find the average RSRP. - If the remaining subframes in SA is less than 20%
of total available subframes (Step 7), then Pth is increased by 3 dB (Step 8), and Steps 4 to 7 are repeated.
• Resource Selection (Steps 9-10): If more than 20% of available channel resources are identified, then, as shown in Step 9, the vehicle populates SB with the first 20% of candidate subframes which has the lowest average S- RSSI in set RA. The vehicle then randomly selects a candidate subframe from set SB as selected resource for first BSM transmission (See Step 10).
At the MAC layer, Mode-4 C-V2X utilizes SPS that uses sensing to determine suitable semi-persistent transmission opportunities, i.e., set of subframes, for BSM transmission. Fig. 1 depicts the SPS algorithm 2, and is explained below. We use sfij to refer to a single-subframe resource where i is the subframe index and j is the subchannel index of J total subchannels. See Fig. 2(a) for the illustration of subframe and other key terminologies (e.g., sensing window) related to SPS.
• Sensing (Step 1): Each vehicle continuously monitors the subframes by measuring the reference signal received power (RSRP) and the sidelink received signal strength indicator (S-RSSI) across all J subchannels; and stores sensing measurements for a prespecified last Nsensing subframes, known as the sensing window. Nsensing is set to 1000 subframes. Let subframe sfn denote the first subframe after the sensing window. Then we can write the sensing window at sfn as the following set of single-subframe resources for the jth subchannel:
• Identifying available resources (Steps 2-8): Each vehi-
• Resource Reselection (Steps 11-13): Each vehicle can reserve the same subframe (selected in Step 10) for next Resource Counter (RC) 3 number of subsequent transmissions with the same transmission interval, i.e., RRI. The RC varies with the RRI to ensure that the selected subframe/resource is in use for at least 0.5 s and at most 1.5 s. This means that for a 20 ms RRI,
3Resource Counter (RC) is the maximum number of transmissions a certain vehicle is allowed (by utilizing the selected subframe/resource in the current selection window) before having to reselect a new set of resources.
cle initializes a selection window with a set of consecutive 2Please refer to [2] and [11] for a detailed discussion on SPS.
RSRP sfn+T1+i−Nsensing+k·RRI ≥ Pth (1) where K = Nsensing . If RRI = 100 ms and
Fig. 2. (a) Illustration of the sensing window, selection window, RRI, and scheduling for SPS and (b) Illustration of Collision Risky scenario caused by a large tracking error.
25≤RC≤75,for50msRRI,10≤RC≤30,andfor a 100 ms RRI, 5≤RC≤15
After RC reaches 0, the vehicle can either continue utilizing the preselected resources with a probability pr or reselect new resources for BSM transmissions with a probability (1 − pr ) (See Steps 11-13).
IV. ON-ROAD SAFETY PERFORMANCE
In this section, we propose a collision risk model, which measures the on-road safety performance of Mode-4 C-V2X networks. Our proposed collision risk model is inspired from risk model presented in literature [15], and is based on tracking error and time-to-collision (TTC) as discussed below.
A. Tracking Error (TE)
Tracking error (TE), etrack, is defined as the difference uv
between the ground truth location of the sender vehicle u and u’s location as estimated by the neighboring receiver vehicle v. Let the most recent BSM received at v from u was generated at time t′ . At time t > t′, the tracking error (TE) that v has in tracking u can be calculated as follows:
u’s location information contained in the most recent BSM received from u. We consider x-coordinate to calculate the tracking error as the vehicle moves in x-direction only 4. The TE is usually significant because - Each BSM takes non-zero channel delay (e.g., propaga- tion delay) to be successfully delivered to the receiver after being generated, and the sender would have moved a non-zero distance during the time period.
- The RRI (or in other words, inter-BSM transmission interval) is significant. For example, if RRI is 100 ms, it means there is at least 100 ms inter-reception delay between two consecutive BSM from sender vehicle u at
the receiver v even if the channel delay is zero. Thus, TE at vehicle v would be significant as vehicle u would have moved to a different location during this time interval. - Lost or delayed BSM, mainly due to channel congestion, will further deteriorate TE.
Note that lower TE value at receiver vehicle v means that v is able to track u well (i.e., accurately position vehicle u). Thus, TE across all vehicles is used to measure TTC and Collision risk as discussed in next subsection.
B. Time-to-Collision (TTC) and Collision Risk
TTC for a pair of vehicles is defined as the time needed for the distance between the two vehicles to become zero, which denotes a potential collision between them. We relate the collision risk (or on-road safety performance) to the TTC, which in turn, utilizes tracking error.
At any time t, the receiver v can estimate TTC with respect to its neighboring vehicle u based on the BSM sent by u as:
t′ t
= |xu −xv| (4)
at time t. su,v is the relative velocity between them. However, since sender u is currently at location xtu at time
t(andnotxt′),thetrueTTCattimetis: u
TrueTTC, TTCt =|xtu−xtv| (5) uv su,v
Using Equations 3, 4, and 5, we get
We designate
is clear that the tracking error affects the TTC calculation. This
estimated TTC, T T C can lead to collision risky situations, uv
and its significance in collision warning was studied in [19].
track t t′ euv =|xu−xu|
su,v
where xu is u’s location as per the last received BSM at v
(3) xt is the actual location of u at time t and xt′ is the
from u with generation time t′ and xtv is the v’s actual location
Estimated TTC, TTCt
model with no lateral movements across lane.
For the ease of presentation, we consider a simple tracking error (TE)
For example, consider that v overestimated/underestimated 5
the TTC to u and is about to take a route/maneuver based t
At any time t, using Eq. 9, we count the number of instances
between each pair of vehicles in which TTC error exceeds the
CRt= uv suv
ms. For instance, the average C for SPS with low RRI
on this errorneous value of T T C uv . Such a situation can be safely avoided by manual intervention if the time taken by the driver to react and apply the brakes to make the vehicle stop, defined as, TTC threshold (TTCth), is less than the true TTC (i.e., TTCutv) at any given time t. Mathematically, TTC threshold, TTCth is given by:
T T Cth = tbrake + treact (7)
where treact is 1s and represents the time taken by the driver to respond to the situation and apply the brakes [20]. tbrake is the time taken by the vehicle to complete to a stop after the brakes have been applied. Assuming vehicle u has velocity su and every vehicle has a maximum deceleration a as 4.6 m/s2 (from [20]) makes tbrake = su/a. If the true TTC, i.e., TTCutv, between a pair of vehicles exceeds the TTCth (i.e., driver’s controllability), it results in a collision risky scenario.
It means, given TTCth, and TTCutv, the collision risk, CRutv, for a given vehicle pair u and v at any given time t, can be computed as follows:
ratio suv and true TTC, T T Cuv , is within TTC threshold, TTCth, as the measure of collision risks.
V. ADAPTIVE SPS++ SCHEDULING PROTOCOL
In this section, we present the limitations of conventional SPS protocol in terms of improving on-road safety perfor- mance of Mode-4 C-V2X, followed by detailed discussion on the proposed adaptive SPS++ scheduling protocol.
A. Limitations of SPS on On-Road Safety Performance
1 TTCt ≤TTC
CRt= uv th (8)
We present the limitations of SPS through a simple C- V2X example (See Fig. 3). The example C-V2X network consists of three clusters 6, of vehicles, where cluster 1, 2, and 3 respectively have 20, 50, and 100 vehicles. We make the following assumptions for all example C-V2X scenarios.
uv 0 otherwise Using Eq. 6, Eq. 8 can be rewritten as:
We assume T1 = 0 and T2 = RRI, which means, the size of selection window is equal to the RRI.
C-V2X physical layer consists of 2 subchannels only. Each BSM transmission uses both the subchannels and takes 1 ms to transmit. This means if the selection window is 100 ms (i.e., RRI = 100 ms), then at most 100 distinct vehicles have unique BSM transmission op- portunities (assuming no collision in resource selection) Each cluster of vehicles is sufficiently spaced apart from each other so that there is no inter-cluster interference. This means, for example, no transmissions from cluster 2 interfere with any transmissions from cluster 1, and vice-versa.
1 (TTCt −eTTC)≤TTC
CRt= uv uv th (9)
uv 0 otherwise
From Eq. 9, it is evident that improving on-road safety
performance of C-V2X, i.e., reducing collision risky scenarios,
is directly proportion to minimizing TE, etrack, between each uv
pair of vehicles in the C-V2X networks.
Fig. 2(b) illustrates how the TTC error causes a collision
risky scenario. In Fig. 2(b), vehicle u estimates vehicle v to
have a TTC, TTCuv beyond TTCth, i.e., in a ”safe” position.
However in reality, vehicle u is close to vehicle v and the true TTC T T Cuv is within the T T C t , and causes collision
risky scenario. Thus, ideally TTC error should be non-existent
(or zero) in order to prevent such a collision risky scenario.
However, as discussed earlier, TTC error exists as the TE
is significant (mainly, due to scheduling, transmission and
communication delays). As per the recommendations by SAE-
Under the above assumptions, let us look at the (i) Channel
Occupancy Percentage Coccup, defined as, the percentage
of the number of vehicles transmitting to the total number
of available subframes transmission opportunities., and (ii)
Probability of Successful Reception (Psuc). For simplicity, let
Psuc is given by 1 where N is the number of vehicles using N
Fig. 3.
C-V2X example network with three clusters of vehicles.
J2945 [21], there exists a TE threshold etrack = 0.5m which
does not lead to a collision risky situation. We use this TE threshold to rewrite Collision risk in Eq 9 as follows.
otherwise
su,v is set to be the average relative velocity between a pair
of vehicles u and v.
5Both overestimating and underestimating TTC are hazardous as overesti- mating means the vehicle gets too close, and decision based on understimated values will impact the other vehicles.
the same subframe for BSM transmissions. (In reality, P gets worse as the number of vehicles N increases.)
Table I depicts the average Coccup and Psuc observed in
the example C-V2X network under conventional SPS with
three different values of RRIs, i.e., 20 ms, 50 ms, and 100
6Each vehicle is at 1-hop (i.e., within the transmission range) of every other vehicles belonging to a certain cluster of vehicle
|etrack| 1(eTTC>th
)and(TTC <TTC) uv th
RRI ×(Coccup in each cluster i) . Coccup for
= 20 ms is given by
each cluster can be computed as follows: Since RRI is 20 ms, there are 20 transmission opportunities (or subframes), (i) in cluster 1, 20 vehicles attempt to transmit, it means
CONVENTIONAL SPS WITH FIXED RRIS VS SPS++ SCHEDULING WITH ADAPTIVE RRI
C1 = 20 × 100 = 100%, (ii) in cluster 2, 50 vehicles occup 20
attempt to transmit, which results in C2 = 50 × 100 = occup 20
250%, and (iii) in cluster 3, 100 vehicles attempt to transmit, resulting in Co3ccup = 500%. Thus, the average Coccup is
20×C 1 +50×C 2 +C 3
occup occup occup × 100 = 379.4%. The average
Ps can be computed in the similar fashion, and it turns out to be 0.35 in case of SPS with RRI as 20 ms. On contrary for SPS with high RRI = 100 ms the average Coccup and Ps are 75.88% and 1 respectively.
Note that SPS with low RRI such as, 20 ms leads to overly congested radio channels (379.4%), and thus, large number of dropped BSM packets (0.35), particularly, in clusters 2 and 3 with > 20 vehicles). The lost packets result in high tracking error, which compromise the on-road safety performance of considered C-V2X network. Whereas, in case of SPS with high RRI as 100 ms, the radio resources are under-utilized (75%), particularly in cluster 1 and 2 with < 100 vehicles. On-road safety performance can be significantly improved by choosing lower value of RRI as lower value of RRIs will improve timely delivery of BSMs. From the above discussion, it is evident that SPS with fixed RRI (irrespective of the chosen value of RRI) is limited in the context of improving overall on-road safety performance of C-V2X networks.
To address the limitations of conventional SPS (and as detailed in next subsection), we propose an improved schedul- ing strategy, termed, SPS++, which allows each vehicle to adapt its RRI, based on its neighboring vehicle density at any given time. In case of example C-V2X network, under such SPS++ strategy, each vehicle in cluster 1 (with 20 vehicles) will choose RRI = 20 ms. Similarly, SPS++ will choose RRI = 50 ms for cluster 2 (with 50 vehicles) and RRI = 100 ms for cluster 3 (with 100 vehicles) – which will result in Coccup = 100% and Ps = 1 (See Table I). It means that the proposed SPS++ protocol strategy with adaptive RRI enables judicious utilization of the radio resources This in turn reduces the tracking error and enhances the on-road safety of C-V2X networks.
B. SPS++ Algorithm Description
This subsection discusses in detail the proposed SPS++ algorithm. As shown in Fig. 4, SPS++ makes significant enhancements to the conventional SPS algorithm. The dotted boxes represent the new or modified steps in SPS++ (not present in conventional SPS), whereas the solid boxes rep- resent the steps borrowed from conventional SPS.
SPS, SPS++ continuously monitors the previous Nsensing subframes by measuring RSRP and S-RSSI and stores the sensing measurements for the sensing window, i.e., Nsensing subframes (Step 1). In Step 2, SPS++ initializes
the estimated RRI to RRImin. SPS++ also initializes
the RSRP threshold (Pth) to a minimum value Pmin.
Unlike SPS, the selection window is not initialized before
starting the resource selection (see Step 2 in Fig. 1),
and is varied as available resources are identified and the
estimated RRI is updated.
• Identifying available resources under the chosen RRI (Steps 3-10): Each vehicle utilizes Nsensing subframes (obtained in Step 2) for identifying the available sub- frames and subsequently, selecting the minimum RRI possible in between transmission while ensuring that there remain resources (or subframes) for other vehicles. - Like SPS, SPS++ sets and updates the Pth(See Step 3). However, in Step 4, SPS++ updates the
estimated RRI and initializes the selection window
withT1 =1andT2 =RRI.T1 isfixedto1to
maximize subframe resources. - Similar to SPS, each vehicle populates set SA with
all subframes in the selection window and SB as a empty set (See Step 5). The candidate subframe exclusion criteria is also borrowed from SPS, except
that, RRI is an adjustable parameter in SPS++. - If the remaining subframes in SA is less than 20%
of total available subframes (Step 7), then, SPS++,
unlike SPS, first checks whether the RRI < RRImax
(Step 8). If yes, RRI is increased by ∆ as shown in
Step 9, and Steps 4 - 7 are repeated. Once RRI has reached RRImax, then Pth is increased by 3 dB in Step 10, and Steps 3 - 7 are repeated.
selection (Step 13) are similar to SPS. However, in
case of SPS++, we choose a reselection counter (RC)
value such that the resource reservation is restricted to
0.5 s, irrespective of chosen RRI. Once RC is zero, unlike SPS (see Steps 12 - 13 in SPS flowchart), SPS++ does not allow reservation of subframe resources. Both these modifications are to ensure that SPS++ allows each vehicle to adjust its RRI at every 0.5 s and account for changing vehicle traffic conditions.
VI. SIMULATION DESCRIPTION AND RESULTS
This section describes the simulation setting, followed by experimental results for the proposed SPS++ compared against conventional SPS protocol.
A. Simulation Setting
Both SPS and SPS++ protocols are implemented on top of a modified system-level network simulator (ns-3) which supports C-V2X mode-4 communications. NIST originally implemented Mode 1 and 2 of the LTE sidelink in Network simulator (ns-3) [22], which was extended to support Mode-3 and Mode-4 C-V2X Communication by Nabil et al. [10].
The highway mobility model assumes vehicles to be moving in a six lane highway road, with three lanes in each direction. While existing work [2], [10] and [23] largely assume a constant velocity model, we consider a more realistic vehicle velocity model where the velocity of a certain vehicle follows a Gaussian distribution centered around vavg (in forward direction, i.e., lane 1, 2 and 3) and −vavg (in reverse direction, i.e., lane 4, 5, and 6). In our experiments, the mean vavg is set to 19.44 m/s (70 km/hr) and the variance is set to 3.0 m/s. [24] supports the assumption of Gaussian random variables as reliable for modelling highway traffic speeds. When each vehicle reaches the end of the highway segment, a warp is applied that moves the vehicle immediately to the other end of road segment, and the vehicle is kept in the same lane. Vehicles with velocities normally distributed around vavg are assigned to lanes 1, 2, and 3 (the 3 bottom most lanes in Fig.1), while vehicles normally distributed around −vavg are assigned to lanes 4, 5, and 6. Each vehicle in the same lane travelled in the same direction and had a maximum deceleration of 4.6 m/s2. A Poisson distribution was used for the initial placement of vehicles along the highway.
For extensive analysis, we compare the performance of SPS++ against conventional SPS with three different fixed RRIs: 20 ms RRI, 50 ms RRI, and 100 ms RRI. The simula- tions were run with varying vehicle densities, ranging from 40 to 160 vehicles. For the purposes of our simulations we assume a 10 MHz channel, though this can be easily adjusted. The initial positions and velocities were kept the same for the same considered vehicle density scenario across the SPS++ and conventional SPS simulations. Each scenario had a simulation time of 8 seconds, and results were averaged across 10 trials. Table II summarizes the key simulation parameters for both SPS++ and conventional SPS with fixed RRIs.
We use the following two performance metrics for the evaluation of SPS++ against conventional SPS.
• Tracking error (etrack)- etrack is the difference between transmitting vehicle u’s actual location and u’s location obtained from the most recent BSM received from u at receiver vehicle v. (see Section IV-A).
• Collision Risk Ratio - It measures the overall on-road safety performance of Mode-4 C-V2X networks, and is defined as the ratio of number of collision risky instances (computed using Eq. 9) to that of both non-risky and risky instances between each pair of vehicles.
• Packet Delivery Ratio (PDR) - The probability that
all vehicles within the range of the transmitting vehicle
receives the transmitted packet. The PDR is calculated
as PDRu = PRu , where PDu is the number of BSMs PDu
transmitted by vehicle u and P Ru is the number of BSMs sent by vehicle u received by neighboring vehicles.
C. Experimental Results
Before we discuss the comparative analysis of SPS++ and SPS in terms of aforestated performance metrics, we first briefly discuss how SPS++ is able to adjust the RRI at each vehicle given varying vehicle densities and over time.
Fig. 5(a) shows how SPS++ chooses, on average, the RRI across all vehicles for each considered vehicle density scenario (i.e., 40, 80, 120 and 160) under varying simulation time. We observe over 10 trials that the average chosen RRI increases over simulation time as vehicles enter the simulation for each considered vehicle density. After 2 seconds, the RRI chosen for each vehicle setting converges to a certain unique RRI value. The chosen RRI in case of 40 and 80 vehicle density (i.e., sparse) is almost four to five times than that of 120 or 160 (i.e., dense) vehicle setting, thanks to the ability of SPS++ to adjust RRI in real-time and adapt to the varying C- V2X environment. Vehicles also take a longer time in dense vehicle settings (5 and 6 seconds for 120 and 160 vehicles respectively) to converge to a chosen RRI value, likely because it takes longer for all vehicles to enter the simulation.
Figs. 6(a)-6(d) show average SPS++ RRI distributions over
Fig. 5. (a) RRI chosen by SPS++ over time and varying vehicle densities: (b) Average high risk tracking error, and (c) Collision Risk(b)
the last second of the simulation 7 and show a large RRI variance in the 80, 120 and 160 vehicle density simulations. Not that the chosen RRI is unique This is due to the varying vehicle traffic densities, and vehicles choosing different RRIs depending on the number of vehicles in their vicinity. The average RRI of 40 vehicles, on the other hand, converges to the RRImin of 20 ms. This is because there are sufficient channel resources to support 20 ms RRI for each vehicle in case of 40 vehicles, irrespective of varying neighborhood sizes. Notice from Figs. 6(c) and 6(d) that while the RRI distribution is skewed to RRImax, there are a number of vehicles still transmitting at lower RRIs.
(c) losses at higher densities. Notice that the tracking error for SPS++ is significantly lower than the tracking error associated with SPS (either with RRI 20 ms, 50 ms or 100 ms) across all vehicle densities. On average, SPS++ outperforms 50 ms RRI SPS (the best performing fixed RRI) by almost 17% and 23% repectively at 60 and 80 veh/km. This indicates that SPS++ protocol is able to choose a suitable RRI at each vehicle that minimizes packet losses and enable up-to-date BSM sharing between a certain vehicle and its neighboring nodes. The tracking error improvements found with SPS++ expectedly to extend to the collision risk. Fig. 5(c) shows a significant decrease in high risk scenarios at 60 and 80 veh/km (55.6 % and 51.20 % respectively) as compared to 50 ms RRI SPS, which is the best performing fixed RRI scheme.
These results shows that the lower RRI chosen is able to help vehicles in both low and high density situations. Observe that even at highly dense situations (80 vehicles/km), SPS++ outperforms the 100 ms SPS protocol, although the average RRI chosen for SPS++ was 100 ms. It is suspected that a small number of vehicles could be using smaller RRIs during the course of the simulation which help reduces the number of high risk situations.
Fig. 6. Average RRI distribution across different vehicle densities
Tracking Error and Collision Risk Analysis. The tracking error (Fig. 5(b)) and the collision risk (Fig. 5(c)) are both used to compare the effectiveness of SPS++ to 20 ms, 50 ms, and 100 ms fixed RRI SPS, in terms of improved road safety performance of C-V2X networks. Fig. 5(b) shows that the tracking error increases as the vehicle density increases, and 20 ms RRI SPS and 50 ms RRI SPS tracking error alternate as the best performing fixed RRI SPS. At higher vehicle density scenarios, 50 ms RRI performs better in terms of tracking error, likely because the 20 ms RRI leads to increased packet
7The reason we use the last second is because by this time, all vehicles have entered the simulation, and chosen an suitable RRI. See Fig. 5(a)
Packet Delivery Ratio Analysis. Fig. 7 shows that SPS++ performs significantly better in terms of PDR when compared
(a) PDR for 40 vehicles (b) PDR for 80 vehicles (c) PDR for 120
Packet Delivery Ratio
Packet Delivery Ratio
Packet Delivery Ratio
Number of Transmissions
Number of Transmissions
Number of Transmissions
Packet Delivery Ratio
Number of Transmissions
Resource Reservation Interval
Tracking Error (m)
Collisions Ratio
to that of SPS with RRIs 20 ms and 50 ms, under all considered vehicle density scenarios. Among the SPS with different RRIs, 100 ms SPS PDR performs the best of all three fixed RRIs. Larger RRIs led to a better PDR, but also yield a large tracking error and collision risk (See Figs 5(b) and 5(c)). Also notice that other than 100 ms RRI SPS, none of the 20 ms RRI, 50 ms RRI, or SPS++ protocols have PDRs near 100%, although at low densities there should be enough subframes for all vehicles. This reduced PDR is likely an effect of the hidden terminal problem, where two transmitters might be out of the sensing range of each other, even if they are both transmitting to the same destination. The likelihood of a hidden terminal collision increases with smaller RRIs, as there are fewer subframes that vehicles can use. However, in SPS++ the PDR remains the same across vehicular densities, and outperforms the 50 ms fixed RRI PDR at higher densities. This is likely because of the larger RRI chosen by SPS++ in the dense scenarios, which likely leads to less congestion and fewer packet collisions.
Discussion. Unlike conventional SPS protocol, the proposed adaptive SPS++ is successfully able to adapt to the considered time-varying C-V2X scenarios and allows each vehicle to dynamically choose the best RRI for the efficient and reliable BSM sharing with neighboring vehicles This greatly reduces the tracking error between each pair of vehicles, which in turn significantly enhances the road safety of C-V2X networks. Thus, SPS++ is a significant step forward for enabling the next-generation of C-V2X mode-4 based connected vehicles and intelligent transportation systems.
VII. CONCLUSION
In this work, we proposed an adaptive, sensing-based semi- persistent scheduling protocol, named, SPS++ for improved on-road safety performance of decentralized V2X networks. Specifically, SPS++, unlike conventional SPS, allows each vehicle to dynamically adjust RRIs based on the availability of channel resources and select suitable transmission opportu- nities for timely BSM transmissions at adjusted RRIs, while accounting for various vehicle traffic scenarios. Our extensive experiments based on Mode-4 C-V2X standard implemented using ns-3 simulator demonstrated that SPS++ protocol signif- icantly outperformed conventional SPS in terms of improved on-road safety performance in all considered C-V2X scenarios. In the future, we will explore designing reinforcement learning (RL) and deep RL, based BSM scheduling protocol that learns vehicle traffic patterns and other contextual factors over time, and selects suitable RRIs.
REFERENCES
[1] J.Thota,N.F.Abdullah,A.Doufexi,andS.Armour,“V2vforvehicular safety applications,” IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 6, pp. 2571–2585, 2020.
[2] R. Molina-Masegosa and J. Gozalvez, “Lte-v for sidelink 5g v2x vehicular communications: A new 5g technology for short-range vehicle- to-everything communications,” IEEE Vehicular Technology Magazine, vol. 12, no. 4, pp. 30–39, 2017.
[3] J. Aznar-Poveda, E. Egea-Lopez, A. Garcia-Sanchez, and P. Pavon- Maria ́o, “Time-to-collision-based awareness and congestion control for vehicular communications,” IEEE Access, vol. 7, pp. 154 192–154 208, 2019.
[4] A. Haider and S.-H. Hwang, “Adaptive transmit power control algorithm for sensing-based semi-persistent scheduling in c-v2x mode 4 communication,” Electronics, vol. 8, no. 8, 2019. [Online]. Available: https://www.mdpi.com/2079- 9292/8/8/846
[5] B. Toghi, M. Saifuddin, H. N. Mahjoub, M. O. Mughal, Y. P. Fallah, J. Rao, and S. Das, “Multiple access in cellular v2x: Performance anal- ysis in highly congested vehicular networks,” in 2018 IEEE Vehicular Networking Conference (VNC), 2018, pp. 1–8.
[6] G. Naik, B. Choudhury, and J. Park, “Ieee 802.11bd and 5g nr v2x: Evolution of radio access technologies for v2x communications,” IEEE Access, vol. 7, pp. 70 169–70 184, 2019.
[7] R.Sengupta,S.Rezaei,S.E.Shladover,D.Cody,S.Dickey,andH.Kr- ishnan, “Cooperative collision warning systems: Concept definition and experimental implementation,” Journal of Intelligent Transp. Systems, vol. 11, no. 3, pp. 143–155, 2007.
[8] C. Huang, Y. P. Fallah, R. Sengupta, and H. Krishnan, “Adaptive intervehicle communication control for cooperative safety systems,” IEEE Network, vol. 24, no. 1, pp. 6–13, 2010.
[9] S. Chen, J. Hu, Y. Shi, and L. Zhao, “Lte-v: A td-lte-based v2x solution for future vehicular network,” IEEE Internet of Things Journal, vol. 3, no. 6, pp. 997–1005, 2016.
[10] A.Nabil,K.Kaur,C.Dietrich,andV.Marojevic,“Performanceanalysis of sensing-based semi-persistent scheduling in c-v2x networks,” in 2018 IEEE 88th Vehicular Technology Conference (VTC-Fall), 2018, pp. 1–5.
[11] A. Bazzi, G. Cecchini, A. Zanella, and B. M. Masini, “Study of the impact of phy and mac parameters in 3gpp c-v2v mode 4,” IEEE Access, vol. 6, pp. 71 685–71 698, 2018.
[12] M. Gonzalez-Mart ́ın, M. Sepulcre, R. Molina-Masegosa, and J. Goza- lvez, “Analytical models of the performance of c-v2x mode 4 vehicular communications,” IEEE Transactions on Vehicular Technology, vol. 68, no. 2, pp. 1155–1166, 2019.
[13] X. He, J. Lv, J. Zhao, X. Hou, and T. Luo, “Design and analysis of a short term sensing based resource selection scheme for c-v2x networks,” IEEE Internet of Things Journal, pp. 1–1, 2020.
[14] T. H. Lee and C. F. Lin, “Reducing collision probability in sensing- based sps algorithm for v2x sidelink communications,” in 2020 IEEE REGION 10 CONFERENCE (TENCON), 2020, pp. 303–308.
[15] B. Choudhury, V. K. Shah, A. Dayal, and J. H. Reed, “Experimental analysis of safety application reliability in v2v networks,” in IEEE Vehicular Technology Conference (VTC2020-Spring), 2020, pp. 1–5.
[16] S. Kaul, M. Gruteser, V. Rai, and J. Kenney, “Minimizing age of infor- mation in vehicular networks,” in 2011 8th Annual IEEE Communica- tions Society Conference on Sensor, Mesh and Ad Hoc Communications and Networks. IEEE, 2011, pp. 350–358.
[17] A.Dayal,E.Colbert,V.Marojevic,andJ.Reed,“Riskcontrolledbeacon transmission in v2v communications,” in 2019 IEEE 89th Vehicular Technology Conference (VTC2019-Spring), 2019, pp. 1–6.
[18] N. Lyamin, A. Vinel, and M. Jonsson, “Does etsi beaconing frequency control provide cooperative awareness?” in 2015 IEEE International Conference on Communication Workshop (ICCW), 2015, pp. 2393–2398.
[19] S.E.ShladoverandS.-K.Tan,“Analysisofvehiclepositioningaccuracy requirements for communication-based cooperative collision warning,” Journal of Intelligent Transp. Systems, vol. 10, no. 3, pp. 131–140, 2006.
[20] M. Green, “”how long does it take to stop?” methodological analysis of driver perception-brake times,” Transportation Human Factors, vol. 2, no. 3, pp. 195–216, 2000. [Online]. Available: https://doi.org/10.1207/STHF0203 1
[21] “On-board system requirements for v2v safety communications,” in Soc. Autom. Eng. SAE International, 2016.
[22] R. Rouil, F. J. Cintro ́n, A. Ben Mosbah, and S. Gamboa, “Implementa- tion and validation of an lte d2d model for ns-3,” in Proceedings of the Workshop on ns-3, 2017, pp. 55–62.
[23] 3GPP, “Study on lte-based v2x services,” 3rd Generation Partnership Project (3GPP), Technical Report (TR) 36.885, 06 2011, version 14.0.0. [Online]. Available: https://portal.3gpp.org/desktopmodules/ Specifications/SpecificationDetails.aspx?specificationId=2934
[24] T. Camp, J. Boleng, and V. Davies, “A survey of mobility models for ad hoc network research,” Wireless Communications and Mobile Computing, vol. 2, no. 5, pp. 483–502, 2002. [Online]. Available: https://onlinelibrary.wiley.com/doi/abs/10.1002/wcm.72
仮訳
「
分散型V2X網における道路安全を強化する適応型半永続的時間割
Avik Dayal1)、Vijay K. Shah1)、Biplav Choudhury1)、Vuk Marojevic2)、Carl Dietrich1)、Jeffrey H. Reed1)
- 1Wireless @ VT、ECE、バージニア工科大学、米国
- 米国ミシシッピ州立大学電気およびコンピュータ工学
{ad6db、vijays、biplavc、cdietrich、reedjh}@vt.edu, vuk.marojevic@ece.msu.edu
要約-分散型の車両対全て（Vehicle-to-Everything:V2X）網（つまり、Mode-4C-V2XおよびMode2a NR-V2X）は、定期的な基本安全伝言（BSM）に依存して、時間に敏感な情報（車両の位置など）を配布する。路上での安全を向上させる可能性がある。 BSM時間割の場合、分散型V2X網は感知に基づく半永続的時間割（SPS）を利用する。この場合、車両は無線資源を検知し、資源予約間隔（RRI）と呼ぶ事前に指定した周期間隔でBSM送信に適した資源を選択する。この白書では、このようなBSM時間割（固定RRIを使用）が、さまざまな車両交通筋書きでの無線資源の深刻な過少および過大使用に悩まされていることを示す。これにより、BSMの適時な普及を著しく損ない、衝突の危険が高まる。これに対処するために、SPS ++と呼ぶ適応RRIに対応するようにSPSを拡張する。具体的には、SPS ++により、各車両は–（i）経路資源の可用性に基づいて（さまざまな車両交通筋書きを考慮して）RRIを動的に調整し、（ii）選択したRRIで適時なBSM送信に適した送信機会を選択できます。 ns-3模擬試験を使用して実装されたMode-4C-V2X標準に基づく実験では、検討されたすべての模擬試験筋書きで、SPS ++が路上安全性能の向上に関してSPSを少なくとも50％上回っていることを示す。
索引用語-分散型V2X、C-V2X、NR-V2X、半永続的筋書き、基本的な安全伝言、交通安全
I.はじめに
Vehicle-to-Everything（V2X）1通信は、主に交通安全を改善し、交通事故の防止/削減などにつながる可能性があるため、次世代の高度道路交通系（ITS）にとって有望な技術です。効率的な交通管理[1]。 V2X通信を可能にする2つの競合技術、専用狭域通信（DSRC）とセルラーV2X（C-V2X）があります[2]。専用狭域通信（DSRC）は、802.11p規格に基づく分散型無線技術です[3]。一方、3GPPは資源14でC-V2X通信を導入し、時間割設定に基づいて状態3と状態4と呼ぶ2つの動作状態を標準化しました。状態3C-V2Xは、eNodeBの網羅の下で集中型時間割方法を採用しており、2台の車両が直接通信できる。無線の選択。
この研究は、MURI Grant N00014-19-1-2621の下で海軍研究局（ONR）によって支援している。この論文は、IFIP Networking 2021での公開を承認する。これは、承認論文の事前印刷版です。
1V2Xは、車両間（V2I）、車両間（V2V）、および車両間（V2P）通信を指す。
資源は、Uu界面（上りおよび下り）を介した携帯基盤からの制御信号で管理する[4]。状態4C-V2Xは、eNodeBからの網羅を必要とせずに車両間の直接通信のために新しいPC5界面を採用しています[5]。携帯接続はユビキタスであると想定できないため、状態4C-V2XをC-V2Xの基準線状態と見なす。最近、新しい無線V2X（NR- V2X）を配布16で想定しており、分散型状態2aNR-V2Xを搭載する予定です[6]。
V2Xの主な使用例は2つある。（i）基本安全伝言（BSM）に依存する協調安全応用。これは、送信者の車両の位置や速度などの重要な安全情報を含む定期的な伝言である。（ii）協調交通効率伝言、車両の流れを助けることを目的とした行事引き金伝言です。 BSMは時間に敏感な情報を運ぶため、BSMは、前方衝突警告[7]や死角/車線変更警告[8]などの協調的な安全応用を可能にします。これは主に、BSMが隣接する車両の正確な測位または位置特定を容易にするためです。古いBSM（大きなBSM時間割間隔による）および/または失われたBSM（経路の混雑による）は、主に隣接する車両の誤った局所化が原因で、衝突の危険な状況が増加するため、安全応用のパフォーマンスに悪影響を及ぼします。言い換えれば、高いトラッキングエラー。トラッキングエラーは、車両の実際の位置と、隣接する車両が（最新のBSMを介して）認識した位置との差として定義できます。上記の議論から、BSM時間割（またはBSMを放送する時間間隔）がBSMの適時性と最小限の経路輻輳を保証するための重要な引数になり、したがって交通安全が向上することは明らかです。
」
一覧
物理記事 上位100
https://qiita.com/kaizen_nagoya/items/66e90fe31fbe3facc6ff
量子(0) 計算機, 量子力学
https://qiita.com/kaizen_nagoya/items/1cd954cb0eed92879fd4
数学関連記事１００
https://qiita.com/kaizen_nagoya/items/d8dadb49a6397e854c6d
統計(0)一覧
https://qiita.com/kaizen_nagoya/items/80d3b221807e53e88aba
言語・文学記事 １００
https://qiita.com/kaizen_nagoya/items/42d58d5ef7fb53c407d6
医工連携関連記事一覧
https://qiita.com/kaizen_nagoya/items/6ab51c12ba51bc260a82
自動車 記事 １００
https://qiita.com/kaizen_nagoya/items/f7f0b9ab36569ad409c5
通信記事１００
https://qiita.com/kaizen_nagoya/items/1d67de5e1cd207b05ef7
日本語（０）一欄
https://qiita.com/kaizen_nagoya/items/7498dcfa3a9ba7fd1e68
英語(0) 一覧
https://qiita.com/kaizen_nagoya/items/680e3f5cbf9430486c7d
転職(0)一覧
https://qiita.com/kaizen_nagoya/items/f77520d378d33451d6fe
仮説（0）一覧（目標100現在40）
https://qiita.com/kaizen_nagoya/items/f000506fe1837b3590df
Qiita(0)Qiita関連記事一覧（自分）
https://qiita.com/kaizen_nagoya/items/58db5fbf036b28e9dfa6
鉄道（０）鉄道のシステム考察はてっちゃんがてつだってくれる
https://qiita.com/kaizen_nagoya/items/26bda595f341a27901a0
安全（0）安全工学シンポジウムに向けて: 21
https://qiita.com/kaizen_nagoya/items/c5d78f3def8195cb2409
一覧の一覧( The directory of directories of mine.) Qiita(100)
https://qiita.com/kaizen_nagoya/items/7eb0e006543886138f39
Ethernet 記事一覧 Ethernet(0)
https://qiita.com/kaizen_nagoya/items/88d35e99f74aefc98794
Wireshark 一覧 wireshark(0)、Ethernet(48)
https://qiita.com/kaizen_nagoya/items/fbed841f61875c4731d0
線網（Wi-Fi）空中線(antenna)(0) 記事一覧(118/300目標)
https://qiita.com/kaizen_nagoya/items/5e5464ac2b24bd4cd001
OSEK OS設計の基礎 OSEK(100)
https://qiita.com/kaizen_nagoya/items/7528a22a14242d2d58a3
Error一覧 error(0)
https://qiita.com/kaizen_nagoya/items/48b6cbc8d68eae2c42b8
++ Support(0)
https://qiita.com/kaizen_nagoya/items/8720d26f762369a80514
Coding(0) Rules, C, Secure, MISRA and so on
https://qiita.com/kaizen_nagoya/items/400725644a8a0e90fbb0
プログラマによる、プログラマのための、統計(0)と確率のプログラミングとその後
https://qiita.com/kaizen_nagoya/items/6e9897eb641268766909
なぜdockerで機械学習するか 書籍・ソース一覧作成中 (目標100)
https://qiita.com/kaizen_nagoya/items/ddd12477544bf5ba85e2
言語処理100本ノックをdockerで。python覚えるのに最適。:10+12
https://qiita.com/kaizen_nagoya/items/7e7eb7c543e0c18438c4
プログラムちょい替え（0）一覧:4件
https://qiita.com/kaizen_nagoya/items/296d87ef4bfd516bc394
Python(0)記事をまとめたい。
https://qiita.com/kaizen_nagoya/items/088c57d70ab6904ebb53
官公庁・学校・公的団体（NPOを含む）システムの課題、官（０）
https://qiita.com/kaizen_nagoya/items/04ee6eaf7ec13d3af4c3
「はじめての」シリーズ ベクタージャパン
https://qiita.com/kaizen_nagoya/items/2e41634f6e21a3cf74eb
AUTOSAR(0)Qiita記事一覧, OSEK(75)
https://qiita.com/kaizen_nagoya/items/89c07961b59a8754c869
プログラマが知っていると良い「公序良俗」
https://qiita.com/kaizen_nagoya/items/9fe7c0dfac2fbd77a945
LaTeX(0) 一覧
https://qiita.com/kaizen_nagoya/items/e3f7dafacab58c499792
自動制御、制御工学一覧（０）
https://qiita.com/kaizen_nagoya/items/7767a4e19a6ae1479e6b
Rust(0) 一覧
https://qiita.com/kaizen_nagoya/items/5e8bb080ba6ca0281927
小川清最終講義、最終講義（再）計画, Ethernet(100) 英語(100) 安全(100)
https://qiita.com/kaizen_nagoya/items/e2df642e3951e35e6a53
＜この記事は個人の過去の経験に基づく個人の感想です。現在所属する組織、業務とは関係がありません。＞
This article is an individual impression based on my individual experience. It has nothing to do with the organization or business to which I currently belong.
文書履歴(document history)
ver. 0.01 初稿 20240616
最後までおよみいただきありがとう4ざいました。
いいね 💚、フォローをお願いします。
Thank you very much for reading to the last sentence.
Please press the like icon 💚 and follow me for your happy life.
<この項は書きかけです。順次追記します。>
This article is not completed. I will add some words in order.
文書履歴(document history)
ver. 0.01 初稿 20210621
ver. 0.02 ありがとう追記 20230527
最後までおよみいただきありがとうございました。
いいね 💚、フォローをお願いします。
Thank you very much for reading to the last sentence.
Please press the like icon 💚 and follow me for your happy life.