Emulation of Camera Failure PDF Print E-mail
Tools

A python library that aims to simulate failures that may occur in a camera during the acquisition/processing phase.

To support the definition of safe and robust vehicle architectures and intelligent systems, we define the failures model of a vehicle camera, together with an analysis of effects and known mitigations. As a natural consequence, here we present a software library for the generation of the corresponding failed images. These images are then fed the trained agent of an autonomous driving simulator: the misbehavior of the trained agent allows a better understanding of failures effects and especially of the resulting safety risk. 

Some examples of the failures injected:

GoldenRun

Github is available at: https://github.com/francescosecci/Python_Image_Failures

More information:

  • Thesis of Francesco Secci, "On failures of RGB cameras and their effects in autonomous driving applications", Master Thesis at the University of Florence, Italy (in Italian only), July 2020. Supervisor: Andrea Ceccarelli (link to thesis page)
  • Paper: Francesco Secci, Andrea Ceccarelli, "On failures of RGB cameras and their effects in autonomous driving applications", in press -- to appear at the 31st International Symposium on Software Reliability Engineering (ISSRE 2020), available on arXiv

GoldenRun

NOISE

BLUR

Broken Lens

Brightness

Sharpness

Chromatic Aberration

ICE

DIRT

 
RCL visits INPE at Sao Josè dos Campos - SP PDF Print E-mail
Projects_1

Today is the day in which Maria Hafiza Maqsood, PhD student at the RCL group, ends her 1-month secondment at INPE (Instituto Nacional de Pesquisas Espacial) in the Sao Paulo State!

It was a very inspiring secondment within the ADVANCE project, as it allowed to establish connections between UNIFI and INPE researchers. Thanks Maria and thanks Fatima's group at INPE for hosting!

WhatsApp Image_2019-10-09_at_15.48.34

 
Farewell to "Brazilian" Researchers PDF Print E-mail
People

To celebrate the two visiting "Brazilians", RCL organized a social lunch with some delicious middle-east food thanks to the chef of the group!

WhatsApp Image_2019-02-11_at_18.13.55

This was the perfect time to celebrate the farewell of Leandro Dias da Silva (Universidade Federal do Alagoas - UFAL - Brazil) and Leonardo Montecchi, a former RCL member which now has a position at UNICAMP, the Universidade Estadual do Campinas, in Brazil.

We hope to see you again in Florence, it was a pleasure working with you!

 
RCL receives brazilian researchers PDF Print E-mail
DEVASSES

Starting from the middle of November, the Resilient Computing Lab is hosting three brazilian students as part of the DEVASSES project. All of them are coming from Universidade Federal de Alagoas (UFAL), in Maceió, Alagoas, and are supervised by professor Baldoino Neto.

Anderson Santos is a graduate student working on Anomaly Detection Benchmarks, and he will stay in Firenze for six months. Caio Barbosa and Felipe Falcão are both undergraduate students visiting Firenze for three months: Caio is studying an Anomaly Detector Framework and Felipe is addressing the Classification of Anomaly Detection Algorithms.

2016.11.24

 
RCL hosted the 5th IRENE F2F meeting PDF Print E-mail
IRENE

RCL hosted the 5th face-to-face meeting of the project IRENE (now at month 19/30), which involved attendees from all the partners of the consortium. In particular, the meeting was directed to discuss the novel architectures and the modeling tool that are being developed within WP3 and WP4 of the project.

During the meeting, RCL promoted the discussion about the integration of the tools that are being developed in the context of the project (modeling, threat analysis, load prediction), aiming to find the best way to connect them. In addition, a presentation about model-based evaluation of the IRENE methods and framework was shown, to support and complement the analysis including evaluations related to dependability and security assessment. These evaluations will be conducted using techniques such as extensions of SAN (Stochastic Activity Networks).

For some of the participants this was the first time in Firenze. Consequently the session of the 26th April ended in a dinner in one of the most famous restaurants of the city centre. Moreover, a brief Firenze-by-night tour was performed before coming back to the hotels.

Meeting Attendees

 
DEVASSES/CECRIS Joint Seminars and Prof. Henrique Madeira visit PDF Print E-mail
CECRIS

Very enjoyable international collaboration is happening in these days at RCL!

Today we had a seminar joining together people with both the DEVASSES and the CECRIS projects. Beáta Hergovics and Máté Tóth are testing engineers from Prolan in Budapest, Hungary, and are currently visiting RCL for a short period to work on the CECRIS project.

Lucas Leal is a Master's student at the Universidade Estadual de Campinas, in Brazil, and he will stay in Florence for a three-months secondment to work in the DEVASSES project. During his secondment he will work on "Model-Based Online Testing for Dynamic SOA"; this seminar was the opportunity to present the initial idea and discuss the topic with the other researchers.

Prof. Henrique Madeira, from the University of Coimbra in Portugal was also attending the workshop. In this period Prof. Madeira is visiting University of Florence, where he is also holding a set of seminars on "Design of experiments for resilience assessment and benchmarking". The goal is to use real data from concrete resilience/dependability assessment experiments, particularly raw data from fault injection experiments, to provide a comprehensive set of examples on how to design and perform correct experiments. The course is covering basic concepts on fault injection for dependability and resilience assessment, steps of designing an experiment, how to perform correct measurements, and hypothesis testing.

 
<< Start < Prev 1 2 3 4 5 6 7 8 9 10 Next > End >>

Page 8 of 16

Resilient Computing Lab, 2011

Joomla - Realizzazione siti web