European Safety & Reliability Conference (ESREL)
3 September 2017
The European Safety & Reliability Conference took place from Monday 19th June 2017 to 22nd June 2017 in Portorož, Slovenia. This conference focusses on a wide variety of topics related to safety and reliability. In its 27th iteration this conference attracted many academic and industry delegates from around the world, including numerous delegates from outside the European Union.
The conference was attended by Darius Roman & Ross Dickie from Cohort 3 of the Centre for Doctoral Training in Embedded Intelligence who presented their paper entitled “A Review of the Role of Prognostics in Predicting the Remaining Useful Life of Assets". This is a paper which highlights case studies from their field of Prognostics and Health Management and illustrates the distinctions between data-driven and model-based prognostics and discusses the blending of these approaches through so-called fusion prognostic methods. Their work uses the case studies to highlight features of these distinct prognostic methods.
The work was presented in the afternoon of the first day of the conference as the first presentation in the afternoon parallel session “Prognostics & Health Management 2", as a review paper this provided a good framing and context for the work subsequent presentations within the session.
The experience at ESREL gave Darius and Ross an opportunity not only to showcase their work in an international setting, first-hand but also the opportunity to gain exposure across both the industrial and academic worlds due to the mixing provided in this conference. The conference also provided the researchers an opportunity to learn about cutting edge work in their field and discuss their work with a diverse group of delegates within the conference.
Team ROSteK consisted of myself, Dr Thomas Steffen (my supervisor), Mauricio Silva (University of Northampton) and Ashton Bowen (De Montfort University). We decided to tackle the challenge of building a CAV simulation. TSC provided us with data, equipment (including a driving simulator and an Omnideck virtual reality simulator), and expert mentors in the field of autonomous vehicles and simulation. We recognised that a key challenge to testing autonomous vehicles is latency of feedback of the controller during simulation. For this reason, we wanted to find a way of putting the actual controller of the autonomous vehicle within the simulation, rather than modelling it. We achieved this by creating an MQTT broker to bridge the virtual world in Unity (windows-based gaming engine) and the controller in Linux-based Robotic Operating System (ROS). This allowed us to test the quality of a controller, similar to how you would test an engine on a test bed before putting it in a vehicle in the automotive industry.