1. |
- Berger, Christian, 1980, et al.
(författare)
-
Containerized development and microservices for self-driving vehicles: Experiences & best practices
- 2017
-
Ingår i: Proceedings - 2017 IEEE International Conference on Software Architecture Workshops, ICSAW 2017: Side Track Proceedings. - 9781509047932 ; , s. 7-12
-
Konferensbidrag (refereegranskat)abstract
- In this paper, experiences and best practices from using containerized software microservices for self-driving vehicles are shared. We applied the containerized software paradigm successfully to both the software development and deployment to turn our software architecture in the vehicles following the idea of microservices. Key enabling elements include onboarding of new developers, both researchers and students, traceable development and packaging, convenient and bare-bone deployment, and traceably archiving binary distributions of our quickly evolving software environment. In this paper, we share our experience from working one year with containerized development and deployment for our self-driving vehicles highlighting our reflections and application-specific shortcomings, our approach uses several components from the widely used Docker ecosystem, but the discussion in this paper generalizes these concepts. We conclude that the growingly complex automotive software systems in combination with their computational platforms should be rather understood as data centers on wheels to design both, (a) the software development and deployment processes, and (b) the software architecture in such a way to enable continuous integration, continuous deployment, and continuous experimentation.
|
|
2. |
- Nguyen, Björnborg, 1992, et al.
(författare)
-
Application and evaluation of direct sparse visual odometry in marine vessels
- 2022
-
Ingår i: IFAC-PapersOnLine. - : Elsevier BV. - 2405-8963. ; 55:31, s. 235-242
-
Konferensbidrag (refereegranskat)abstract
- With the international community pushing for a computer vision based option to the laws requiring a look-out for marine vehicles, there is now a significant motivation to provide digital solutions for navigation using these envisioned mandatory visual sensors. This paper explores the monocular direct sparse odometry algorithm when applied to a typical marine environment. The method uses a single camera to estimate a vessel's motion and position over time and is then compared to ground truth to establish feasibility as both a local and global navigation system. Whilst it was inconsistent in accurately estimating vessel position, it was found that it could consistently estimate the vessel's orientation in the majority of the situations the vessel was tasked with. It is therefore shown that monocular direct sparse odometry is partially suitable as a standalone navigation system and is a strong base for a multi-sensor solution.
|
|
3. |
- Nguyen, Björnborg, 1992
(författare)
-
Bio-inspired retinal optic flow perception in robotic navigation
- 2021
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis concerns the bio-inspired visual perception of motion with emphasis on locomotion targeting robotic systems. By continuously registering moving visual features in the human retina, a sensation of a visual flow cue is created. An interpretation of visual flow cues forms a low-level motion perception more known as retinal optic flow. Retinal optic flow is often mentioned and credited in human locomotor research but only in theory and simulated environments so far. Reconstructing the retinal optic flow fields using existing methods of estimating optic flow and experimental data from naive test subjects provides further insight into how it interacts with intermittent control behavior and dynamic gazing. The retinal optic flow is successfully demonstrated during a vehicular steering task scenario and further supports the idea that humans may use such perception to aid their ability to correct their steering during navigation. To achieve the reconstruction and estimation of the retinal optic flow, a set of optic flow estimators were fairly and systematically evaluated on the criteria on run-time predictability and reliability, and performance accuracy. A formalized methodology using containerization technology for performing the benchmarking was developed to generate the results. Furthermore, the readiness in road vehicles for the adoption of modern robotic software and related software processes were investigated. This was done with special emphasis on real-time computing and introducing containerization and microservice design paradigm. By doing so, continuous integration, continuous deployment, and continuous experimentation were enabled in order to aid further development and research. With the method of estimating retinal optic flow and its interaction with intermittent control, a more complete vision-based bionic steering control model is to be proposed and tested in a live robotic system.
|
|
4. |
- Nguyen, Björnborg, 1992, et al.
(författare)
-
Systematic benchmarking for reproducibility of computer vision algorithms for real-time systems: The example of optic flow estimation
- 2019
-
Ingår i: IEEE International Conference on Intelligent Robots and Systems. - : IEEE. - 2153-0858 .- 2153-0866. ; , s. 5264-5269
-
Konferensbidrag (refereegranskat)abstract
- Until now there have been few formalized methods for conducting systematic benchmarking aiming at reproducible results when it comes to computer vision algorithms. This is evident from lists of algorithms submitted to prominent datasets, authors of a novel method in many cases primarily state the performance of their algorithms in relation to a shallow description of the hardware system where it was evaluated. There are significant problems linked to this non-systematic approach of reporting performance, especially when comparing different approaches and when it comes to the reproducibility of claimed results. Furthermore how to conduct retrospective performance analysis such as an algorithm's suitability for embedded real-time systems over time with underlying hardware and software changes in place. This paper proposes and demonstrates a systematic way of addressing such challenges by adopting containerization of software aiming at formalization and reproducibility of benchmarks. Our results show maintainers of broadly accepted datasets in the computer vision community to strive for systematic comparison and reproducibility of submissions to increase the value and adoption of computer vision algorithms in the future.
|
|