Categories
Uncategorized

dRISK’s Tesla spoofing rig is shown at CENEX Connected Autonomous Mobility event

After almost 18 months without any UK based physical events to attend, Chess, Nils and Rav proudly demonstrated dRISK’s ground breaking work at CENEX LCV/CAM. This is the UKs biggest Connected and Autonomous Mobility event held from the 20th to the 23rd September, at the UTAC Millbrook facility in Bedford. The UK government invited us to show case our work at this prestigious event attended by Trudy Harrison MP, Parliamentary Under Secretary of State at the Department for Transport.

Curious attendees were drawn to our stand which featured a Tesla Model 3 (also confused as Ferrari), a SAE level 2 Autonomous vehicle fitted with our bespoke spoofing rig. The team talked attendees through the process of spoofing the Tesla to believe that it was using its vision based detection system to ‘see’ projected images such as lane lines, cars and pedestrians in front of it. A projected pedestrian wearing standard everyday clothing was easily recognised by the car, whilst a pedestrian wearing white noise clothing against a white noise background was not registered by the car. Using camouflaged clothing on the simulated pedestrian against a woodland background again tricked the car into not seeing the pedestrian, this is clearly a possible failure mode for the Tesla, vision based perception system and could point to potential accidents if to occur on the roads.

The event bought us together with other government sponsored projects managed by Innovate UK (Endeavour, ServCity and ALEAD). The dRISK stand raised lots of enthusiasm and we are looking forward to engaging with those who expressed interest in our technology.

Chess introduced dRISK lead engineer Nils Goldbeck and AI engineer Lorenzo Niccolini who presented ‘Improved Hazard Detection through training and testing on dRISK Edge Case Scenarios’ on the virtual days of the event. The talk can be viewed here.  Nils and Lorenzo summarised that training on dRISK edge case scenarios improves the detection of high risk events by six times with twice the confidence.

We would like to thank Innovate UK, The UK Centre for Connected and Autonomous Vehicles (CCAV) and CENEX for this fantastic opportunity.