Commercial Viability Assessment and Planning of Safety-Critical Embedded SW of Electrified Road Vehicles


Shah Alias Sangani A., Harman C., Kinav E., Göl M., Hartavi A. E.

SAE 2021 WCX Digital Summit, Virtual, Online, Amerika Birleşik Devletleri, 13 - 15 Nisan 2021 identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.4271/2021-01-0132
  • Basıldığı Şehir: Virtual, Online
  • Basıldığı Ülke: Amerika Birleşik Devletleri
  • Anahtar Kelimeler: Battery Management System, Cost Estimation, detailed COCOMO, Electric Vehicle, Safety-Critical Embedded Software, Software Planning
  • İstanbul Gelişim Üniversitesi Adresli: Evet

Özet

© 2021 SAE International. All rights reserved.Recent extraordinary technological progress in the field of high-voltage-batteries has led an evolution in the automotive industry, resulting in vehicle manufacturers to shift from conventional powertrains towards electric ones. However, electrified road vehicles are amazingly complex. Today, the number of operations for a modern electric vehicle grew from millions to billions per second. This is mainly fueled by the electrification and safety requirements in addition to critical core functionalities. This emphasizes the importance of software development cost, effort, and production planning in the automotive industry. In this paper, in the framework of EU funded H2020 OBELICS project, a detailed-COCOMO approach is proposed for a manually coded safety-critical embedded SW for an electric vehicle not only to plan the project well in advance but also to assess its commercial viability using quantifiable cost metrics to make the process more objective and repeatable. In this context, a case study is given for a battery-management software using an algorithmic model. The case-study has demonstrated not only the planning requirements but also the impact of project type on effort and time estimation in each phase of the software lifecycle. The results have shown that inaccurate software project type estimation can lead an error of up to 16% in effort and 19% in development time according to the phase. It also shows that reliability, database size and complexity are the major contributors of the total time, effort and hence of the cost. The results have also demonstrated the impact of model-based design and automated code generation tools on planning for different phases of the V-cycle briefly.