101 lines
9.3 KiB
Markdown
101 lines
9.3 KiB
Markdown
tags:: [[#zotero]]
|
|
title:: @Model-based systems engineering toolchains for software development of robotic autonomous systems
|
|
item-type:: [[journalArticle]]
|
|
original-title:: Model-based systems engineering toolchains for software development of robotic autonomous systems
|
|
language:: en
|
|
authors:: [[Calvin Cheung]], [[Andrew Pfiel]], [[Mark Petrotta]], [[Catherine Haggerty]]
|
|
library-catalog:: Zotero
|
|
links:: [Local library](zotero://select/library/items/XCZ58PV4), [Web library](https://www.zotero.org/users/1039502/items/XCZ58PV4)
|
|
|
|
- [[Abstract]]
|
|
- Robotic autonomous system development requires a system-of-systems approach to manage the complexity inherent in integrating disparate fields of engineering into a single system. While model-based systems engineering (MBSE) provides the tools and processes needed in the management of complex robotics system development, there has traditionally been challenges in leveraging MBSE in rapidly changing software-intensive domains such as autonomy. To bridge the gap between MBSE and software engineering, we developed model import toolchains that automate the generation of SysML models. These toolchains are tailored for the Robot Operating System and leverage model import files that are maintained by the software developer, ensuring alignment between the models and codebase. By leveraging the model import toolchain, we have been able to successfully integrate an MBSE approach with our software engineering processes, improving efficiency in the development process and the quality of robotic systems.
|
|
- ### Attachments
|
|
- [PDF](zotero://select/library/items/4NSUQ82S) {{zotero-imported-file 4NSUQ82S, "Cheung et al. - Model-based systems engineering toolchains for software development of robotic autonomous systems.pdf"}}
|
|
- ### Notes
|
|
collapsed:: true
|
|
- # Annotazioni
|
|
(3/12/2024, 22:47:54)
|
|
|
|
- “there has traditionally been challenges in leveraging MBSE in rapidly changing software-intensive domains such as autonomy.” (Cheung et al., p. 1) #a28ae5
|
|
|
|
- “model import toolchains that automate the generation of SysML models” (Cheung et al., p. 1) #a28ae5
|
|
|
|
- “By leveraging the model import toolchain, we have been able to successfully integrate an MBSE approach with our software engineering processes, improving efficiency in the development process and the quality of robotic systems.” (Cheung et al., p. 1) #ffd400
|
|
*The goal of the paper is not clear by looking at the abstract.*
|
|
|
|
- “combined to create systems that integrate together towards the common goal of a mobile autonomous system” (Cheung et al., p. 1) #a28ae5
|
|
|
|
- “reliable systems engineering approach is vital to the overall success of RAS development.” (Cheung et al., p. 1) #5fb236
|
|
|
|
- “improving the processes used in integrating robotic software engineering and model-based systems engineering (MBSE) via the development of a systems modeling language (SysML) import toolchain.” (Cheung et al., p. 1) #a28ae5
|
|
*This is supposed to be the novelty of the paper even though it is not clear what are the "competitors" and their limitations that authors want to address with this paper.*
|
|
|
|
- “reducing the friction experienced by software engineers when working within a systems engineering framework and ensuring synchronicity between the software and the overall system models.” (Cheung et al., p. 1) #ffd400
|
|
*it's not clear what are the considered processes their limitations and how authors are addressing them!!*
|
|
|
|
- “the formalized application of modeling to support system requirements, design, analysis, and verification and validation activities beginning in the conceptual design phase and continuing throughout development and later lifecycle phases” (Cheung et al., p. 1) #5fb236
|
|
|
|
- “The overall purpose of models is to aid in understanding, decision making, and predict events.” (Cheung et al., p. 1) #5fb236
|
|
|
|
- “specification of the structure, behavior, requirements, and parametrics of a system.” (Cheung et al., p. 1) #5fb236
|
|
|
|
- “Fig. 1 shows the nine diagram types defined by SysML, each of which conveys information about different aspects of a system. Model elements across each diagram type can be linked together to describe the interrelations amongst the different parts of a system.” (Cheung et al., p. 1) #ffd400
|
|
*What is the need for this figure, or in general show the different types of diagrams in SysML*
|
|
|
|
- “challenges in integrating MBSE with software engineering methodologies exists. M” (Cheung et al., p. 1) #a28ae5
|
|
|
|
- “toolchains that reduced the burden placed upon software engineers in the creation of models” (Cheung et al., p. 2) #5fb236
|
|
|
|
- “we wanted to be able to allow for the automated generation of SysML models based on structured, text-based, and machine-readable files managed by the developers” (Cheung et al., p. 2) #ffd400
|
|
*AUtomated generationof SysML models starting from textual descriptions? How do they look like?*
|
|
|
|
- “ROS is a framework for the development of robotic software, leveraging a collection of libraries, tools, and conventions to enable collaborative software development and reuse” (Cheung et al., p. 2) #5fb236
|
|
|
|
- “The instantiation import files on the other hand define the structure of a ROS system as collection of individual components.” (Cheung et al., p. 2) #ffd400
|
|
*In the end we are condiguring a ROS based system by means of XML files. Isn't it?*
|
|
|
|
- “Definition Phase: Import ROS constructs and components into SysML model elements” (Cheung et al., p. 2) #ffd400
|
|
*Here it seems that we are generating SysML models starting from XML specifications.*
|
|
|
|
- “Applying these toolchains to our ROS development efforts enabled us to apply MBSE to our software engineering efforts in a number of different ways, increasing efficiency of processes and consistency of the system” (Cheung et al., p. 3) #ffd400
|
|
*The process is not clear at all!*
|
|
|
|
- “he model elements and diagrams generated by our model import toolchains have enabled integration between MBSE and software engineering processes in a synchronized fashion, ensuring consistency between the two domains” (Cheung et al., p. 3) #ffd400
|
|
*What do you mean?*
|
|
|
|
- “In order to perform these reviews, accurate diagrams of both the overall RAS architecture and the architecture of the components are required to compare interfaces towards integration.” (Cheung et al., p. 3) #5fb236
|
|
|
|
- “This linkage between requirements and capabilities allows for greater system consistency and reduction in errors” (Cheung et al., p. 3) #5fb236
|
|
|
|
- “Changes to capabilities can immediately flag the affected requirements and vice versa, allowing for rapid response in ensuring that overall system needs are met” (Cheung et al., p. 3) #ffd400
|
|
|
|
- “MBSE utilization in RAS development, there may still be situations where more traditional forms of documentation are preferred.” (Cheung et al., p. 3) #5fb236
|
|
|
|
- “XML was decided upon due to the prevalence of the XML file format in both ROS and Doxygen, a code documentation generator” (Cheung et al., p. 3) #5fb236
|
|
|
|
- “TABLE I. ROS CONSTRUCTS” (Cheung et al., p. 3) #5fb236
|
|
|
|
- “s on the benefits of different technologies cannot be in a bubble. The entire development environment and ecosystem must be taken into consideration for design choices related to selection of competing technologies.” (Cheung et al., p. 4) #5fb236
|
|
|
|
- “inclusion of MBSE” (Cheung et al., p. 4) #ffd400
|
|
*what kind of integration has been done? How was before and how is it after the adoption of the presented approach?*
|
|
|
|
- “From the perspective of a software developer, they may see their current practices as sufficient, and any additional effort adds no value to their workflow” (Cheung et al., p. 4) #5fb236
|
|
|
|
- “Incremental adaptation of new MBSE workflows that meet developers within their comfort zone yielded better results than trying to add new processes.” (Cheung et al., p. 4) #a28ae5
|
|
|
|
- “needed a ROS profile to use in our toolchain to generate SysML diagrams.” (Cheung et al., p. 4) #5fb236
|
|
|
|
- “we developed a ROS profile for our toolchain that defines common ROS constructs in SysML.” (Cheung et al., p. 4) #5fb236
|
|
|
|
- “increasing the variety of ROS constructs that the model import toolchain can represen” (Cheung et al., p. 4) #5fb236
|
|
|
|
- “While the constructs described in Table I can be used to provide a basic overview of a ROS system, there are additional constructs that can be represented that would give a more detailed understanding of the system.” (Cheung et al., p. 4) #ffd400
|
|
*If you keep adding ROS constructs, you risk to get too complex. It is important to keep and leverage the concept of abstractsions when possible.*
|
|
|
|
- “we would be able to increase the fidelity at which a ROS system can be modeled and represented in MBSE” (Cheung et al., p. 4) #ffd400
|
|
*Yes, at the price of model complexity. What would be the related benefit then?*
|
|
|
|
- “we developed SysML model import toolchains that automated the generation of SysML model elements based on ROS constructs and model import files.” (Cheung et al., p. 4) #5fb236
|
|
|
|
- “utility of MBSE in RAS development to provide a starting point for the greater robotics community to follow suite in their own efforts.” (Cheung et al., p. 4) #5fb236 |