The results of Move_UK’s three year project have determined that automated driving systems will need to deal with the unexpected.
The MOVE_UK consortium, led by Bosch, is now completing the final phase of its research programme, designed to accelerate the development and deployment of automated driving systems using a new method of testing which they call ‘Connected Validation’.
The project is a joint effort with Jaguar Land Rover, Transport Research Laboratory, telematics experts The Floow, insurance provider Direct Line and the Royal Borough of Greenwich, with part funding provided by the UK government.
Real-world trials of the consortium’s ‘Connected Validation’ methodology, conducted in and around the Royal Borough of Greenwich, have found that 8,500 hours of driving can be distilled into just 450 short driving sequences where on-board systems detected a potential hazard. Only 25 of these sequences were classed as ‘critical’ braking situations which are highly relevant for the validation of the next generation of driver assistance systems.
A fleet of five sensor-equipped Land Rover Discovery Sport vehicles driven by Greenwich council and TRL employees were used to collect MOVE_UK’s data. This fleet will have completed over 100,000 miles on public roads by the time the project concludes at the end of next month.
The advanced sensors fitted to the vehicles give a full 360-degree view. MOVE_UK says that this ensures that the 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.
The technique used by MOVE_UK aims to limit the amount of information collected, by only recording data from what it deems the most relevant events. MOVE_UK states that this significantly reduces the time it takes to process and analyse the data and bring a roadworthy system to market.
When a ‘significant’ driving event such as sharp braking or cutting in takes places, vehicle data and radar data are recorded and immediately uploaded into the cloud. The data is then managed and immediate analysis of data is carried out by engineers and or developers.
Academy Director at TRL Richard Cuerden commented, ‘MOVE_UK has developed analysis techniques, using big data collected by vehicle sensors and cameras, to assess the real world performance of these complex vehicle systems operating in silent mode.’
‘Understanding how advanced driver assistance and automated systems perform in the real world context is critical. Efficient validation and verification methodologies are key components to help get safer and approved vehicles on our roads as quickly as possible, ultimately preventing deaths and injuries,’ added Cuerden.
The research programme hopes to also allows consortium partners Direct Line Group and The Floow to build more accurate insurance models that hopes to help towards providing future insurance products and pricing that will be more closely linked to risk.
Director of Motor Development at Direct Line Dan Freedman said, ‘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.’
In terms of the projects success, the Connected Validation process has already improved roadside infrastructure, and has been involved in a traffic sign recognition case. A 30mph traffic sign that was recorded during a data collecting drive was ‘covered’ and nearly unnoticeable, and as a result has since been maintained so that it can be fully visible. Other roadside traffic signs have also been moved or rotated after being recorded and deemed ‘misleading’ to drivers.
MOVE_UK has said that a significant outcome of the project so far, has been that it has determined that automated driving systems will need to be able to deal with the unexpected. Sudden sharp braking, cars cutting across lanes, pedestrian choices on roads, are just a few examples of the unexpected events that an automated vehicle will have to interact with whilst on the road. MOVE_UK’s data hopes to be able to inform those in the automotive and autonomous industry, that these unexpected events will need to be dealt with in order to successfully and safely get autonomous vehicles driving on our roads.