Modern cars require lane-keeping – or “lane tracing” – technology to achieve a five-star safety rating.

Tiny cameras ‘read’ the lane markings on the road ahead – like an extra pair of eyes – and automatically adjust the vehicle’s steering and/or brakes to keep the car within the lane.

But some car brands cut corners to pass the test in a staged environment – while the technology ‘zig-zags’ on real roads and in some cases puts drivers at risk instead of trying to prevent a crash.

Now the peak car safety body, the Australasian New Car Assessment Program (ANCAP), has announced the first steps in cracking down on car brands that deliver dodgy technology.

New testing will measure the accuracy and sensitivity of lane-keeping systems – and apply the results to a vehicle’s overall safety score.

If the lane-keeping tech is found to be flawed, a five-star rating will not be applied.

Above: An Isuzu D-Max ute with a robotic steering test rig fitted.

It means future cars should have more accurate and less jerky lane-keeping systems.

But first, ANCAP is going to retest cars it has already assessed, to validate its new procedures.

While no cars will be stripped of their existing five-star scores, the safety authority hopes the new round of assessments – which do not require vehicles to be crashed – will provide a clearer picture of the current crop of good and bad performers.

Above: A robotic car used by ANCAP to test autonomous emergency braking systems on real vehicles. The Toyota Hiace van is the communications hub.

In a visit to Australia last year, Euro NCAP Secretary-General, Dr Michiel van Ratingen conceded some car companies had cut corners with lane-keeping technology, just so they could get a five-star score.

The safety expert told a media briefing at Sydney’s Crash Lab in November 2023:

When we started to develop these tests (for lane-departure warning), our understanding was that vehicle manufacturers would not bring systems in that would be upsetting their customers. 

We focussed on evaluating critical scenarios where the systems should intervene, and whether they will do a good job. 

Let’s imagine a crash is about to happen, how can the system effectively intervene? 

So we (created) a test on a test track and … that would form part of the basis on which we give five stars. 

As it turns out, many vehicle manufacturers basically used that (test) as a blueprint for (lane-keeping systems), which was never the intention for us. They basically said, if I meet the test, that’s fine. That’s enough. 

And because of (vehicle engineering) development times getting shorter … what we actually see happening is that they (lane-keeping systems) were not properly calibrated in the real world by some car manufacturers.

Above: The test track in central-western NSW where lane-keeping systems are assessed.

The top European safety executive said some car companies blamed Euro NCAP and Australasian NCAP for the implementation of lane-keeping assistance systems, however Michiel van Ratingen said:

We are upset that the vehicle manufacturers (told the media) that Euro NCAP was forcing them to develop (lane-keeping) technology. 

We never forced them to do that. We assumed they would have the ethical courage to implement a good system. 

And if they could not deliver that, then they wouldn’t go for it. But they went for it just to get the points (towards a five-star safety rating).

So now it’s back to (Euro NCAP and Australasian NCAP) to actually adapt a new test procedure and to into include real-world testing to measure the accuracy and the aggressiveness of these systems.

We’re talking about technology that wasn’t even there five years ago. So we’re talking about first generation systems that are coming to the market that may have some flaws.

We need to have a new way of evaluating these systems, not just looking at a critical scenario (on a test track) and how they intervene or whether they do a good job, but we also have to take the car out on the road.

There’s a few rotten apples in the fleet that screw up everything for everyone.

It’s an auto manufacturer’s job to make sure (lane-keeping) systems are not annoying, and don’t have false (alarms)..

Above: A robotic pedestrian used by ANCAP to test autonomous braking systems on new cars.

In a statement issued by ANCAP today, Australia’s top car safety body said it is about to start retesting cars with lane-keeping systems to help develop new test protocols.

The aim is to make sure future lane-keeping technology makes no sudden moves and does not jerk the steering wheel so aggressively it sends the car zig-zagging down the road.

ANCAP says it will assess:

  • Vehicle path and position – is the autonomous intervention smooth and intuitive?
  • Steering angle and velocity – is the intervention gradual, rapid, sharp – or unnecessary?
  • Steering torque – is the vehicle’s response difficult for the driver to override, leading to a feeling of loss of control?
  • Lateral vehicle acceleration – is the sideways force experienced by the driver severe?

ANCAP says the research follows feedback from consumers and media who raised concerns about the jerkiness of certain lane-keeping systems.

In a media statement, ANCAP Chief Executive Officer, Carla Hoorweg said properly calibrated lane-keeping technology is “critical to consumer acceptance”.

Unfortunately the behaviour of (certain) vehicles is having consumers question the benefits of these systems, and in some cases, turn them off.

What we don’t want to see is these systems being badged as ‘annoying’ and switched off.

What we want to demonstrate are the differences in vehicle behaviour, and by sharing these results, encourage manufacturers to improve their systems. This will in turn improve the acceptance of these systems by their customers.

We saw a great example of manufacturer response to this type of feedback last week with Mitsubishi rolling out an update to its Driver Monitoring System to improve the driving experience of its Triton.

ANCAP says the results of this preliminary project will be shared with car companies and used to inform refinements to ANCAP’s upcoming 2026-2028 test protocols and criteria.