Job opportunity: Researcher in Trustworthy Autonomous Systems Verification

Do you work in verification? Our TAS functionality node is looking for a researcher to join our team in sunny Bristol. 

This role involves working with us to identify and formalise requirements for the verification and validation of autonomous systems with evolving functionality, as well as identifying design principles that facilitate this. You will also undertake fundamental and applied research into verification and validation techniques and apply these techniques to case studies in swarm robotics, soft robotics and machine learning uncrewed air vehicles.  

We would love to hear from you if you have a strong background in verification, including knowledge and hands-on skills in, e.g., formal methods, simulation, test-based techniques and runtime verification and have excellent programming, modelling and analysis skills- but we don’t mind where you learnt those skills. 

This position is part of the multidisciplinary UKRI Trustworthy Autonomous Systems Node in Functionality which is composed of a team of researchers from ethics, sociology, computer science and engineering working together to develop design-for-trustworthiness techniques for autonomous systems. 

The position will be based both at the Bristol Robotics Laboratory, the largest centre for multidisciplinary research in the UK, and at the University of Bristol campus within the Trustworthy Systems Laboratory, where we research, develop and evaluate techniques that enable designers and engineers to gain confidence in the trustworthiness of the systems they design and build. This position will also involve close interaction with the Bioinspired Flight Lab, Soft Lab, and Hauert Lab. 

Visit the University jobs site for full details of the position including how to apply online.  

Vacancy summary 

Research Associate or Senior Research Associate or Research Fellow in Trustworthy Autonomous Systems Verification (Job ref ACAD106198) 

Department of Computer Science, University of Bristol  

 Salary: £34,304 to £50,296  

 Closing date: 9 August 2022 

  

For informal enquiries contact: Prof Kerstin Eder (Kerstin.Eder@bristol.ac.uk) or Dr Shane Windsor (shane.windsor@bristol.ac.uk) 

Robot Hackathon

In June this year the universities of York, Manchester and Sheffield hosted a hackathon – a competition designed to collaboratively solve a problem in a particular field. The competing teams were a mix of undergraduates, postgraduates and post-doctoral researchers. The goal of the competition was to produce some code that would allow a single operator to control a group of 5 robots. The TAS functionality node was represented at the event by Dr Sabine Hauert and Dr James Wilson.

Over the three days attendees heard from an expert panel discussing key challenges around swarm robotics, which included Dr Hauert, and helped develop solutions for remotely operating robot swarms, using Pi-puck and MONA robots.  On the final day teams made presentations, with prizes for the best solutions.

The competition involved the robots moving around a small arena and gathering in locations (generated virtually) to score points. The locations themselves were labelled from 1-5, which was the amount of robots you needed gathered there in order to gain points. Most of the devised solutions involved control over a single agent with others following along autonomously, though some groups put together fully autonomous solutions. Some groups were even able to run code on swarms remotely at other universities which was fun!

Thanks to the universities, organisers, hosts and speakers for a brilliant time.

 

 

Autonomous vehicle testing using game engines

We’re excited to announce the recent publication by some members of our team, Professor Kerstin Eder and Dr Greg Chance of the Trustworthy Systems Laboratory (https://www.bristol.ac.uk/tsl), of research into the determinism of game engines used for simulation-based autonomous vehicle verification. These simulations are useful in developing control systems and test environments for the vehicles; but only if they behave deterministically, and therefore yield reliable and repeatable results. Unfortunately, this isn’t a characteristic of game engines and they often produce different outcomes for the same initial conditions. The paper, published in IEEE Transactions on Intelligent Transportation Systems, presents the causes and effects of this non-deterministic behaviour, and sets forth a case study showing the shortcomings of a particular simulator, and a methodology to assess and minimize simulation precision. To read the paper please go here: On Determinism of Game Engines Used for Simulation-Based Autonomous Vehicle Verification