Critical data ‘will speed arrival of driverless cars in the UK’
A three-year Government-funded project to accelerate the adoption of autonomous vehicles in the UK has concluded that the key is to use only critical data from test vehicles rather than all the data.
The MOVE_UK programme, which concludes next month, has demonstrated that 8,500 hours of driving can be distilled into just 450 short driving sequences where on-board systems detected a potential hazard. The other driving data doesn’t teach the system anything new.
The so-called ‘Connected Validation’ methodology was developed during real-world trials in Greenwich, south London. Critical data recording is triggered by events such as hard braking.
Only 25 sequences were classed as ‘critical’ braking situations making them highly relevant for the validation of the next generation of driver assistance systems.
By limiting the amount of information collected, Connected Validation significantly reduces the time it takes to process and analyse the data and bring a roadworthy system to market.
In total, the fleet of vehicles completed 100,000 hours since the project’s start in 2016. Over 100,000 speed limit signs were successfully detected and 1000 event-triggered video sequences were recorded.
The trials – led by Tier 1 supplier Bosch and including partners auto maker Jaguar Land Rover, insurance firm Direct Line, auto research centre TRL, telematics start-up The Floow and the Royal Borough of Greenwich – freatured five sensor-equipped Land Rover Discovery Sport vehicles driven by Greenwich council and TRL employees.
Dr Sam Chapman, chief innovation officer of The Floow.
The advanced sensors – including video cameras, webcams and radar, but not Lidar – give a full 360-degree view, ensuring these tests provide a much better understanding of the behaviour of surrounding vehicles which will be fundamental for the development of safe and ‘human-like’ level 3 and 4 automated driving features.
‘Ask anyone about their first few months behind the wheel since passing their test and they’ll tell you how much they've learned,’ said Dr Sam Chapman, chief innovation officer of The Floow.
‘Real world experience counts for a lot and we’ve discovered that, not unexpectedly, humans do not drive to the letter of the law - people speed, cut each other up and make all manner of mistakes.
‘Their interpretation of hazards and safety is markedly different than the Highway Code would have you expect. Autonomous vehicles will need to be programmed to accommodate this real-world behaviour.
‘This will be a critical factor to consider when planning for autonomous mobility in the future and determining how risk is managed.’
The team behind the £5.5 million trials – part funded by a £3.4 million Government grant – say that the Connected Validation approach could also be applied after launch of a new system to market, allowing system engineers, and potentially approval authorities and insurers, to monitor safety performance of production systems fitted to end users’ vehicles.
The MOVE_UK research could also allow consortium partners Direct Line Group and The Floow to build more accurate insurance models which will help towards providing future insurance products and pricing that is more closely linked to risk.
‘Understanding the impact of autonomous vehicles on insurance is vital for us as a business and for the customers we serve. The Connected Validation process will allow us to get a much quicker understanding on how autonomous cars will interact with other cars, pedestrians and infrastructure and will play an important role when it comes to identifying the risks that will give consumers the confidence to embrace the technology,’ said Dan Freedman, director of motor development at Direct Line Group.
The final MOVE_UK report will be released next month.
By Ray Molony, Technology Editor
Monday 24 June 2019