Written evidence submitted by Vesos Solutions Limited (SDV0024)
This response is been submitted by VESOS Solutions Limited to highlight a key issue about driverless vehicle adoption, namely that the UK is not yet able to fully exploit connected vehicle data that has been already available for some time.
We focus on our response on the lack of exploitation of the eCall system, deployed in all newly approved cars and vans in the UK since 2018, as there are key lessons learned that can help smooth the rollout of driverless vehicles and a quick win benefit for the safety of UK roads.
In simple terms, every day nearly 200 vehicles send connected vehicle eCall data about an emergency situation that can reduce deaths, injuries, accidents, and congestion, including on smart motorways. But it is not yet fully exploited and so raises serious questions about effective roll out of driverless vehicles unless lessons are learned and a new approach to connecting roads and vehicles adopted.
Connected and Automated – what’s the difference?
To many people, they are the same thing. But there are key differences.
Any human operated vehicle or moving object (from HGV to pedestrian or cycle or escooter) can be connected, giving data on its location and status, or receiving information about conditions around it. The technology is already available, through devices built into the vehicle, links to smartphones and fleet management devices. Any age of vehicle can be connected (demonstrated by the RAC Foundation’s Connected London to Brighton Run ), unlike self-driving vehicles which must be newly manufactured. Benefits can apply to all road users not simply new vehicles.
Driverless vehicles are therefore often connected, as they work far better with information about their neighbouring roads and need to broadcast where they are. But not all connected vehicles are driverless and so can bring benefits far more quickly and to more road users than driverless cars. They are the Cinderella of the technology world because they are not as “sexy” as driverless cars and require co-operation with roads authorities to deploy at scale. This is a key challenge – moving from small pilots and demonstrations to at scale national services needs commitment and change to the status quo.
Making the most of existing connected vehicles of all forms is therefore a key path to deploying automated ones, but the UK is slow to exploit at scale world-leading work in pilots done in the UK and is falling behind our neighbours.
Connected vehicle opportunities
Because of VESOS’s work with road operators, vehicle makers, data providers and emergency services, we have seen many untapped opportunities for connected data. These are well examined in a report to the RAC Foundation  but in summary are:
- sending information to vehicles to better manage roads, improve traffic signal waiting times, increase safety and advise and pay for parking. The Talking Traffic system deployed in the Netherlands is an example of what could be done for all road users. All of these apply to automated vehicles that need this information to work well; and
- using data from connected vehicles to better manage road networks so saving emissions, prioritise roadworks to save money and detect collisions and stopped vehicles to save lives.
It is in this second group we now focus with a specific example – the lack of use of data from the eCall system by roads authorities and emergency services. We show this example as it includes all the key areas you cover in your questions.
What is ecall?
eCall has been built in to all newly type-approved cars and light vans in the UK since 2018. When activated, a voice call is set up to the 999-emergency call handling service, where an operator asks if assistance is required (police, fire or ambulance).
eCall can be automatically activated, typically by the air bag being deployed; or manually, by pressing an SOS button. eCall has a microphone, speaker and mobile SIM built-in so your vehicle does not need to be carrying a mobile phone. The eCall SIM uses 3G cellular communications working across any mobile operator. It is inactive until used so cannot track drivers.
More information is available at https://nationalhighways.co.uk/road-safety/ecall/
eCall volumes are currently at over 5,500 calls per month connected to the 999 services, with average growth rates of 50% per year.
The eCall activation also importantly sends data, similar to an SMS, called a Minimum Set of Data (MSD). This contains data critical to the incident response, including: vehicle ID, last three locations, confidence in location, direction of travel, number of occupants, fuel type, and whether the alert was manual or automatic.
Whilst the voice element of eCall is already in wide use, the MSD is not used by emergency services or road operators such as National Highways. This means that data needs to be transcribed / dictated by phone (difficult for a precise GPS location and 17-digit Vehicle Identification Number (VIN)) instead of being sent from “machine to machine” as data.
To make the most of eCall, the MSD could be transmitted independently of the voice through to the emergency services and road operators to alert them to collisions and incidents including stopped vehicles. The MSD does not replace the voice channels. Instead, it provides an “early warning” and added intelligence, for example:
- vehicle make and model from the VIN
- the fuel type – very useful for Fire Services but also recovering electric vehicles that cannot be towed away from breakdowns
- the location, direction and previous location points – so that for example a vehicle crossing over a central reserve can be detected, or a vehicle that has left the road can be found
- the number of occupants
Using eCall data to detect collisions could have challenges from false alarms and calls away from the road network. We have however demonstrated these can be overcome through checking the MSD before it is sent on.
The MSD is also of value for manually reported events raised by pressing the button – for example to gain an exact location of a stopped vehicle without eCall on a smart motorway but reported by an eCall user.
Key statistics about current use of eCall are:
eCalls passed to the emergency services in July 2022:
178 ecalls per day need 999 assistance
Of which automatic activations (Jul 22):
Usually involving air bag deployment which also require recovery of the vehicle
One automatic eCall every 65 minutes
Proportion of licenced vehicles in the UK with eCall fitted
Based on registered sales to date
New vehicles per year with eCall
Depends on car sales
eCall is one of the first instances of data in traffic operations sourced directly from vehicles, so if road emergency responders are not using this data, it is not a strong foundation for driverless vehicles. It is also the first example of a mass scale nationwide connected vehicle data source, but has not yet been fully exploited. The lessons behind this lack of adoption will be invaluable moving forward.
What benefits might it bring?
VESOS undertook a study in 2019 for Highways England that showed eCall could be used alongside roadside radar to detect stopped vehicles on smart motorways. The benefits of this were at that time:
Average time to report an incident per emergency responder in the chain by voice alone
For example, National Highways are notified by the Police, who are notified by 999
VESOS estimate of time to notify emergency responder using MSD
Based on a demonstration product
Our work showed this would give significant reductions in collisions above and beyond radar, but would also be of use on the many stretches of road where there is no CCTV or other collision detection, e.g., rural dual carriageways such as the A303. These roads have a much higher collision rate than smart motorways.
VESOS is also a major contributor to the Conference of European Directors of Roads’ SHADAR project, which reviews stopped vehicle detection methods, and makes recommendations for best practice. Our SHADAR work shows that eCall (some vehicles, all locations) complements roadside detection such as radar (all vehicles, some locations).
For the emergency services, eCall data could help in judging time critical factors like:
- the type of response needed – e.g., an electric vehicle needs specialist techniques
- the number of emergency vehicles needed – e.g., ambulances for a 9-seat minibus collision
- where to approach the scene from – e.g., a vehicle crossing a dual carriageway may need both directions
Why has the data not yet been used?
There are many lessons about the uptake of eCall data that apply to driverless vehicles:
- Out of date research – existing business cases may be stale and require updating in light of new technology and changing conditions. For eCall, we still find people referring to a 2006 business case that included the cost of eCall units as an option, even though eCall is mandated and the owners pay for it in the price of the vehicle. eCall data is “free” and compared to the installation and maintenance costs of roadside technology, cheap to implement.
- lack of stakeholder understanding / awareness – People do not know about eCall. It is not widely promoted by vehicle manufacturers as it is mandated for all makes so offers no competitive advantage, and drivers have not (until recently) been educated in its use. A recent National Highways campaign revealed this lack of understanding in drivers but this also occurs in road operators and government. There are also myths about eCall- that it is “Waze”-like crowd-sourced data, has no business case as above, and only comes from premium vehicles that have not been overcome but persist.
- There is a lack of skills and understanding of data in potential users notably the roads sector, where data is seen as a competitor to existing infrastructure such as radar and loops rather than an augmenting feed. Integrating eCall data into highway systems is perceived as a challenge but for National Highways at least, these systems are ready to accept eCall derived data.
- There is a “not invented here syndrome” in the roads sector, where data from vehicles is poorly understood and needs to be validated before use. Questions are always asked for example about false airbag activations, when there are incredibly high levels of reliability for these developed over many years.
- Because data is involved, people become immediately concerned re GDPR and data privacy, and understandably want full details else eCall is put into the “too difficult” pile.
- There is a lack of central co-ordination about eCall data and its use, with every fire brigade, ambulance service and police force, and every highway operator, taking different approaches.
- Internal politics in user organizations means that whilst they may recognise the value of eCall data, actually using it is subject to much internal approval and discussion, so nothing actually happens.
- There has been a feeling that “we can put off using it until there are enough vehicles equipped.” The UK now has around 7.5m vehicles equipped and receives 20 automatic ecalls a day. There are enough vehicles.
- Users don’t want raw data from eCall – and our multiple 999 services (over fifty fire services alone) and 150 odd road operators just want data for their geographic area. Hence VESOS has developed technology to filter and send eCall messages to the right place
- Finally, using data from connected vehicles is not part of the day job for stretched emergency services and road operations teams, they rely on others who may not be aware of the progress in eCall.
All the above will apply to both connected and automated vehicle roll outs. There are many lessons for automated vehicles in the above.
What is the solution?
VESOS has been working with road operators and emergency services to develop tools to access eCall data and process it. But we can only use these if:
- There is wider awareness of eCall data by potential users especially in roads operators, perhaps through a pilot (which would require funds equivalent to the catering budget of some driverless vehicle projects)
- Government produces guidance perhaps via DfT’s Road Safety Investigation Branch to co-ordinate awareness of eCall
- There is a seed change in attitude to data from vehicles being less “proven” than that from roadside sensors
Wider lessons from the eCall experience
Having been involved with eCall since 2006, and being frustrated that a lifesaving data source is not fully used, we suggest the following lessons need to be applied:
Likely uses: eCall is fitted to all newly approved cars and vans since 2018, but is also available as an app for motorbike and horse riders, and via a dashcam for any vehicle. HGVs and buses, along with line fit motorbikes, are being developed.
Key lesson: because of the average age of vehicles in the UK being over 8 years (and due to Covid sales slump) it has taken until 2022 for 20% of vehicles to be equipped with eCall, although this will now accelerate as new car sales recover. Using Smart phones and similar to add connected services will be much quicker and more flexible than fitting them in the vehicle.
Progress of research: The early UK business case for eCall was not strong, as it included the cost of extra in vehicle equipment. But as its mandated anyway, we may as well exploit the data available as costs are low. In addition, the early business case for eCall did not consider stopped vehicle detection on all lane running, a use case that does give considerable benefits not initially considered.
Currently there is some proposed research on eCall data for trauma research, and VESOS is self-funding our own work developing filtering tools, but there is no mining or harvesting of the eCall data set. It was identified ecall data could be of use to the forthcoming Roads Safety Investigation Branch
Key Lesson: Research needs to be revisited as circumstances and technology changes. There was a national opportunity to monitor the early feeds of eCall data from 2018 but due to the “not invented here” syndrome, nothing was done. When research moves to business-as-usual services as for eCall, monitoring is still needed.
Implications for infrastructure: Like many in vehicle and other systems, eCall uses 3G. Although it connects to any provider and is designed for poor comms, nationwide mobile coverage would help remote areas (and needed for connected and autonomous).
Key Lesson: Mobile comms are good enough in most areas for eCall, as the unit works for any service provider with coverage (as do 999 calls). But full national coverage of more remote areas might add to the benefits of finding lost vehicles and improve emergency responses.
Regulatory Framework: eCall was adopted by a Directive across Europe but the UK was not equipped to make the most of this regulatory change.
Key Lesson: Regulation is not enough – work needs to be done to monitor deployment and learn from data collected.
Data protection and perceptions of privacy issues
We often encounter false perceptions of data privacy issues; that the MSD contains personal information, it tracks people, or the driver needs to give consent. We have researched the Data Protection and GDPR issues thoroughly and rebut these views when they come up, but they are often used as an excuse for inaction.
Key Lesson: Data Protection and GDPR is vital but must not be wielded as a “too difficult answer switch”. The starting position should always be “how can we”
Safety and perceptions of safety
eCall could be used to report automatically a Collison in a smart motorway, manually report a vehicle fire on a dual carriageway or report a cyclist injury in an urban road – it works anywhere. It will need to be fitted to automated vehicles so they can report similar events as today's vehicles, and the data then needed will be far more detailed.
eCall has been perceived by some commentators as “too late” – that once the collision has occurred there is little value, and efforts should be placed instead on areas such as Automatic Emergency Braking. This is also of value (and is now mandated) but ignoring eCall would remove a layer from the often quoted “Swiss cheese” accident causation model . A similar argument could be falsely applied to airbags or even seatbelts.
Key Lesson: Fact, guidance and information rather than perceptions of safety are required about connected and automated vehicles
The role of Government and other bodies, such as National Highways and local authorities;
eCall has come from the vehicle and emergency service community, and hence knowledge by roads operators and authorities and the transport side of government is poor. It is seen as a communications service, not a road safety and operations opportunity. Above all the key lesson is for government to remove siloes and focus on obtaining low hanging benefits from connected vehicle data especially eCall.
Guidance on data use, support for new skills and a “why not” rather than “too difficult” attitude are needed. Without this, full autonomy has little practical hope.
Appendix: About VESOS
VESOS Solutions Ltd, established by four industry experts Andy Rooke, Danny Woolard, Andy Graham and Alan Gentle, offers project consultancy, design services and solutions designed to improve road safety and response times based around eCall technology and services
eCall is now mandated on most vehicle types in Europe and is being adopted by many other countries. It has the potential to save lives by providing connectivity, both voice and data, to emergency services and hence road operators in the event of a vehicle accident or a vehicle occupant pressing the SOS button. It can dramatically reduce response times for stopped vehicle detection too. However, there are challenges to maximising the benefits of eCall.
Our team has been working together in eCall for over three years, proving its value in incident generation, traffic management, and historical data insights. Individually, we go back over 20 years in eCall standards and implementation, connected vehicles, telematics and traffic operations.
Following our early projects, we developed our TeCall© platform. This can provide high quality incident data to support traffic operations by harvesting eCall/SOS alerts, qualifying their validity, enhancing the data, prioritising, and augmenting vehicle information to make it meaningful for road operations and response teams.
 https://vesos.co.uk/sites/default/files/downloads/HeCall Phase 1 report.pdf
 https://en.wikipedia.org/wiki/Swiss_cheese_model is an established model for safety – multiple layers of safety interventions aligned like Swiss cheese where holes rarely go all the way through