Thatcham Research is urging regulators to update how Intelligent Speed Assist (ISA) systems are assessed after real-world testing revealed significant differences between regulatory approval results and on-road performance.
ISA technology, which uses camera and GPS data to identify speed limits and communicate them to drivers, became mandatory on all new vehicles sold in the UK and European Union from July 2024. While the technology has the potential to improve road safety and reduce collisions caused by speeding, Thatcham Research believes current testing methods do not accurately reflect how systems perform in everyday driving conditions.
The organisation’s latest research highlights what it describes as a regulatory gap in the assessment of Advanced Driver Assistance Systems (ADAS), raising concerns that inconsistent ISA performance could undermine driver confidence and limit the technology’s safety benefits.
Current EU regulations assess ISA systems based on their accuracy over the total distance travelled. Although this provides a benchmark for approval, it can mask performance issues when speed limits change. Thatcham Research argues that these transition points are critical to the driver experience and should play a greater role in how systems are evaluated.
Using an event-based testing methodology, which measures performance every time a speed limit changes, Thatcham Research assessed three vehicles in real-world driving conditions. The results revealed a substantial difference between distance-based and event-based accuracy scores.
The poorest-performing vehicle achieved 91.3% accuracy under the current distance-based assessment. However, when measured using the event-based approach, accuracy fell to 74.3%, meaning the system displayed the wrong speed limit in approximately one in four instances.
Even the strongest-performing vehicle, which achieved a distance-based accuracy score of 98.39%, recorded an event-based accuracy score of 90.3%. This indicates that around one in ten speed limit changes were incorrectly identified, highlighting a level of inaccuracy that could affect driver trust and acceptance of the technology.
According to Thatcham Research, widespread adoption of ISA depends on drivers having confidence that the system will consistently provide accurate information.
The testing also uncovered examples of ISA systems displaying speed limits that do not legally exist on UK roads, including 5mph, 10mph, 15mph and 100mph. Such errors could create confusion for drivers and increase the risk of inappropriate vehicle responses when ISA is linked to Adaptive Cruise Control (ACC), potentially resulting in unexpected braking or acceleration.
Yousif Al-Ani, Principal ADAS Engineer at Thatcham Research, commented: “Systems should be aligned with the defined speed limit parameters for each market. Where readings fall outside those recognised limits, they should be filtered to avoid unintended responses.
“Presenting a driver with readings that fall outside recognised limits can erode confidence in the technology, and that is the trend we are seeing across the vehicles we assess.”
Thatcham Research warns that repeated inaccuracies may encourage drivers to disable ISA altogether, reducing the effectiveness of legislation designed to improve road safety.
Jonathan Hewett, CEO of Thatcham Research, said: “ADAS technologies represent one of the most significant opportunities we have to improve road safety, but that opportunity is only realised if the systems work correctly and earn the confidence of the people using them. ISA is a case in point, the intent behind the legislation is sound, helping drivers stay within speed limits saves lives. But a system that misreads limits, intervenes unexpectedly or presents drivers with speed data that bears no relation to the road they are on does not assist them. It frustrates and distracts them, and they turn it off.
“The automotive industry has the capability to deliver ISA that is accurate, consistent and genuinely useful. What is needed now is a regulatory standard that demands exactly that – one that measures performance at the moments that matter, rather than allowing systems to pass approval while failing drivers in real-world conditions. Getting this right is not optional. The safety case for ADAS depends on drivers trusting these technologies enough to keep them switched on. We will continue to assess these systems and feed it into the safety pillar of our new Vehicle Risk Rating system for insurers.”
Thatcham Research is now calling on regulators and industry stakeholders to modernise ISA approval standards by incorporating event-based testing alongside existing assessment methods. The organisation believes this would provide a more accurate reflection of real-world performance and help improve consumer confidence in the technology.
While achieving higher event-based accuracy will require improvements in camera recognition, GPS mapping data and sensor fusion technologies, Thatcham Research says these developments are both achievable and essential if Intelligent Speed Assist is to deliver its full road safety potential.
