Robohub.org
 

Oxbotica’s autonomous vehicle software learns its environment and how it changes over time


by
10 October 2016



share this:
Oxbotica testing Selenium. Credit: Oxbotica/YouTube

Oxbotica testing Selenium. Credit: Oxbotica/YouTube

Oxbotica, a UK technology company with a focus on mobile robotics and driverless vehicles, has created Selenium, an autonomous software system acting similar to a ‘brain’ for a vehicle. Selenium can work in pedestrianised environments as well as roads and motorways, and is not reliant on GPS to operate – meaning it can transition between indoor and outdoor settings, overground or underground. The system has been developed to be “vehicle agnostic” and can be applied to cars, self-driving pods (e.g. for campuses and airports), and warehouse truck fleets.

The software developed by Oxbotica’s team of scientists, mathematicians and engineers gives any vehicle an awareness of where it is, what surrounds it, and how it should move to complete a task.

Selenium’s system uses patented algorithms giving vehicles a next generation level of intelligence to autonomously perform a range of mobility tasks, including motion control, braking, calibration, navigation, static and dynamic obstacle detection. One of the main focusses at the moment is low-speed autonomous transport, using low-cost vision sensors, explains Dirk Gorissen, Senior Engineer at Oxbotica.

“While you are travelling at a fairly low speed, it is a much more unstructured and unpredictable environment than, say, a highway. There is a lot of talk of level 4 or even 5 autonomy but, to our knowledge, nobody has reached this reliably yet. Even Google with their millions of autonomously driven miles still needs a safety driver. There is a lot of work to get there, and we have our own roadmaps in place that will enable our customers to be class leading,” he said.

https://www.youtube.com/watch?v=cfTst2AV6M8

“Robotics outside of simulation is difficult and getting fully reliable solutions that work anytime and any place are where many others have failed,” said Gorissen. “Many components need to work together seamlessly in order to create a safe, reliable system. Particularly in the context of autonomous vehicles where you have to deal with human behaviour from other drivers and road users. In our launch projects in Greenwich and Milton Keynes, we are working in pedestrianised areas and working amongst pedestrians, cyclists, cats and dogs, as well as, regular road users.”

https://www.youtube.com/watch?v=0-Io1JMalSo

Selenium is set to be deployed at a series of autonomy trials where Oxbotica is the sole supplier of autonomy software, including the £8 million GATEway project in Greenwich and the LUTZ Pathfinder self-driving pod project in Milton Keynes. The company is also working with manufacturers in a broad spectrum of mobile autonomy domains, driverless cars.

“A key strength of the Oxford heritage is our focus on systems that learn from experience. The more often a vehicle traverses a given route the more it learns about its environment and how it changes over time (rain, fog, snow, road works, etc.),” said Gorissen. “We don’t believe in building a single all-encompassing model of the world or that autonomy is a binary on/off switch that is either active all the time or never. Rather we believe in a system that will offer autonomy as and when it is confident enough to do so. Scaling can then be achieved by sharing experiences of the world between vehicles.”

“We are very carefully focused – intentionally – on the things that we do well. For the applications and domains we have set our sights on we are fully on track and you will be able to see our Selenium autonomous control system running public demos of pods around Greenwich in East London from early 2017,” he said.

Oxbotica is a spin-off from Oxford University’s Mobile Robotics Group.


If you liked this article, you may also want to read:

See all the latest robotics news on Robohub, or sign up for our weekly newsletter



tags: , , , ,


Kassie Perlongo Kassie is the Managing Editor at Robohub.
Kassie Perlongo Kassie is the Managing Editor at Robohub.





Related posts :



Robot Talk Episode 119 – Robotics for small manufacturers, with Will Kinghorn

  02 May 2025
In the latest episode of the Robot Talk podcast, Claire chatted to Will Kinghorn from Made Smarter about how to increase adoption of new tech by small manufacturers.

Multi-agent path finding in continuous environments

  01 May 2025
How can a group of agents minimise their journey length whilst avoiding collisions?

Interview with Yuki Mitsufuji: Improving AI image generation

  29 Apr 2025
Find out about two pieces of research tackling different aspects of image generation.

Robot Talk Episode 118 – Soft robotics and electronic skin, with Miranda Lowther

  25 Apr 2025
In the latest episode of the Robot Talk podcast, Claire chatted to Miranda Lowther from the University of Bristol about soft, sensitive electronic skin for prosthetic limbs.

Interview with Amina Mević: Machine learning applied to semiconductor manufacturing

  17 Apr 2025
Find out how Amina is using machine learning to develop an explainable multi-output virtual metrology system.

Robot Talk Episode 117 – Robots in orbit, with Jeremy Hadall

  11 Apr 2025
In the latest episode of the Robot Talk podcast, Claire chatted to Jeremy Hadall from the Satellite Applications Catapult about robotic systems for in-orbit servicing, assembly, and manufacturing.

Robot Talk Episode 116 – Evolved behaviour for robot teams, with Tanja Kaiser

  04 Apr 2025
In the latest episode of the Robot Talk podcast, Claire chatted to Tanja Katharina Kaiser from the University of Technology Nuremberg about how applying evolutionary principles can help robot teams make better decisions.

AI can be a powerful tool for scientists. But it can also fuel research misconduct

  31 Mar 2025
While AI is allowing scientists to make technological breakthroughs, there’s also a darker side to the use of AI in science: scientific misconduct is on the rise.



 

Robohub is supported by:




Would you like to learn how to tell impactful stories about your robot or AI system?


scicomm
training the next generation of science communicators in robotics & AI


©2025.05 - Association for the Understanding of Artificial Intelligence


 












©2025.05 - Association for the Understanding of Artificial Intelligence