MOVE_UK - Phase 2: Automated Driving Data Analysis Report
Earlier this month the latest report from MOVE_UK was released which covered the findings of phase 2 of this three year project.
MOVE_UK is contributing to the progression towards automated driving through connected systems validation and analysis of ‘big data’. The project is part funded by the UK Government’s CCAV department and is a collaboration between a consortium of six industry-leading organisations - Bosch, Jaguar Land Rover, TRL, Royal Borough of Greenwich, Direct Line Group and The Floow.
The project began in August 2016 and is set to conclude in July 2019 with phase 2 having run from December 2017 until September 2018. The project sees a series of Jaguar Land Rover trail vehicles driving to, from and around the test area of the Royal Borough of Greenwich. Each vehicle has a multitude of monitoring equipment in it which records the driving miles so journey data from within the vehicle systems, new sensors and controls can be analysed to inform insights and analysis.
For phase 2 of the project, advanced forward facing radar sensors were added to all six of the trial vehicles, extending the vehicles’ sensor perception alongside existing sensors. This addition allowed investigation of new safety systems helping to trial and test driver safety and assistive systems by combining new use of the sensor data.
As a first phase analysis, we reviewed the accuracy of raw sensor data in differing conditions and how data is used in existing safety systems. The project’s data recording and analysis tools were also updated and extended to allow easy access to statistics and visualisations of the radar data collected. This data was used to evaluate new vehicle technologies that assist (or save) the vehicle occupants in certain operating conditions e.g. forward facing automated radar emergency braking systems.
The primary focus of the report covers various project partners analysis covering wide topics. Fundamentally, phase 2 of the project work explores maximising the value from newly added radar data. This is explored in four key new use cases making use of the Phase 2 radar capabilities, these are:
- Lead Vehicle Statistics (LVS) - This adds detailed understanding of how drivers ‘follow’ other vehicles. This specifically looks at the gaps maintained between vehicles in various conditions and situations. This analysis helps not only how to make autonomous driving more comfortable to end users but it also allows The Floow to better understand risk in creating valuable statistics on driving behaviours in differing conditions, this alone offers strong value to risk scoring configurations supporting future telematic score refinements.
- Radar-based Autonomous Emergency Braking (ARB) - The radar sensors collate data specifically around extreme deceleration and braking events when rapidly approaching obstacles. This helps to understand near crash scenarios and fine grained driver behaviour using dense in vehicle data. In particular this helps not only develop next generation emergency braking technology but also providing The Floow with unique insights into evolving vehicle risks around new assistive technologies.
- Cut-in Scenarios (CIN) - MOVE_UK’s cut-in scenario use case shows the benefits of real-world event data to help understand driver behaviour, of both the driver and surrounding drivers. For instance, when and how do drivers react to the lane changes and relative movements of another vehicle. This has great value for designing future autonomous systems but it also gives clear insight into key risk scenarios allowing detailed study into the risk and frequency of specific maneuvers.
- Further telematics (Telematics 2) - This specific use case supports the investigation of data value from connected vehicles to understand consistency, reliability and how potential future telematics products can be enhanced with connected car data.
For The Floow, MOVE_UK allows us to understand not just future technologies and new potentials but also how this can test and refine our current cutting edge scoring. Ultimately for us we see the telematics aspect of this project as the most important for future development as these insights help us better understand driver risk helping strengthen our understanding of risk for those we work with, such as insurers, fleet managers and auto manufacturers.
The driver behaviour insights we have gathered in this project and beyond it are beneficial for many including insurers helping them to price policies more accurately, to car makers to help develop safer cars and governments and local authorities to design road systems and make changes which will improve road safety whilst also helping them deal with the challenges of autonomy.
During the third and final phase of the project, information from the additional corner Radar systems will be combined with the existing analysis to help work towards a refined model of risk. By enabling a 360-degree understanding of proximity, this enriched data will grant us access to an enhanced view of vehicles during autonomous operation.
One of the objectives that we have with the MOVE_UK project is to reach an enhanced understanding of behaviour from richer and more comprehensive mobility data. The aim of risk estimation research is to identify new behaviour clusters and factors that can aid the understanding of the likelihood and outcomes of incidents or real risk scenarios. Or in other words, existing telematics is used to seek new ways to distinguish journeys that appear very similar, but which actually present divergent levels of risk when viewed from the perspective of additional data.