The PolarSys Rover project will receive contributions based on features developed by the Dortmund University of Applied Sciences and Arts in the A4MCAR project. The A4MCAR is a multi-core, radio-controlled car used to demonstrate parallel applications in embedded systems. Part of this development was funded through the Google Summer of Code (GSOC). The stipendium winner, Mustafa Ozcelikors, describes his work and what we can expect in the Rover project.
What was your objective for the GSOC project?
We wanted to investigate and evaluate the performance of the Eclipse APP4MC project in a distributed multi-core system by developing a demonstrator that involves typical automotive application features. To improve APP4MC's usability, we investigated various tracing methodologies, options to reduce power consumption, and software deployment techniques across multiple cores for GNU/Linux platforms.
What basic functionality have you and your colleagues developed for the A4MCAR?
The A4MCAR features not only low level functionalities such as sensor and motor driving but also high level features such as image processing, camera streaming, server-based wireless driving via a Web-UI, bluetooth connectivity via an Android application, system core monitoring and analysis features and a touchscreen UI.
We provide detailed documentation on Github on the project page. https://mozcelikors.github.io/a4mcar/, https://a4mcar.readthedocs.io The software can be downloaded via the git repository:
Your team is also working on the PolarSys Rover. Can you describe the technology transfer that may happen from the A4MCAR features to the Rover?
My team and me are working on developing applications on the PolarSys Rover to demonstrate multi-agent car-to-cloud communication to support the development in the open source APPSTACLE (Eclipse KUKSA) project. We are working on extensions to the Rover software to provide not only mobile robot functionality (such as environment perception, actuation, and human-machine interfaces) but also IoT and Cloud related technologies (using Eclipse Hono Cloud, WiFi, Bluetooth etc.).
I am currently working on transferring the image processing application that I've created for the RCCar, and the ability to drive over a web UI. Also another student is working with me on bluetooth-based driving implementations for the Rover.
The trace functionality is also an important implementation that I transferred from the A4MCAR to the Rover. In fact, the Rover has higher potential than the A4MCAR, because it makes use of lower-level libraries and has a uniform toolchain. This makes it easier to develop the tool support to do performance analyses of threads and processes.
Do you plan to be involved further in the Rover project?
Of course. I am now developing a fully functional web back-end using new technologies such as Node.js, socket.io, and REST APIs for the Rover project, which I think project partners would find very interesting! Furthermore, I created a sensor layer board and a software development kit (SDK) extension to make it an open project not only in the software sense but also in hardware and modular cross-compiling support. I am also currently creating comprehensive documentation of my development work. We provide documentation on our current implementations which is accessible with the link https://app4mc-rover.github.io/rover-docs.
Would you recommend GSOC for other students?
I would definitely recommend it to anyone who is interested in learning about open-source development and get a Google referral. They will not only learn a lot about open-source workflows, licensing, proper and standardized coding techniques but also will be part of a bigger community that consists of the smart people participating in GSOC. During my participation, I have learned about various tools for version control, documentation, standard toolchains and their configurations, which is a great know-how for those who would like to develop for the open-source community in the future. Another great incentive is that they will receive a scholarship after each evaluation and gifts from Google such as a t-shirt, pen, notepad, certificates, and yearly open-source community subscription (for lwn.net).