Photo: Euro NCAP/Markus Spiske/edited by autoevolution
Soon after Tesla Model Y excelled in Euro NCAP tests, achieving the highest score of any vehicle ever tested, doubts were cast about the results. White hacker @greentheonly found evidence that the cars used for tests had modified software. Euro NCAP investigation found no evidence of cheating, though.
One month ago, @greentheonly discovered a piece of code in the Tesla software that identified safety organizations that conduct safety assessment tests. Specifically, Tesla added “ANCAP,” the Australian version of Euro NCAP, to the software. Speculations ran wild about the reasons for this addition, and Tesla was accused of cheating on those tests. Following this, Euro NCAP started its own investigation into the matter, and the first results are now out.
According to an email Euro NCAP sent to CNN last week, Euro NCAP program director Aled Williams confirmed that the investigation had “so far” not revealed any evidence of Tesla’s attempt to cheat the tests. As for why Tesla introduced the additional ANCAP code into the software, Williams said it is only used to identify how a vehicle is configured for a particular region.
“Different regions (such as Europe, Australasia, etc.) differ in terms of legislation as well as road conditions/markings, etc. The recent addition of ANCAP to the code of Model Y coincides with the start of sales of that vehicle in Australasia,” Williams said.
Paul Maric, the CarExpert journalist who broke the initial news about the potential Tesla cheating code, has revised his original story to include the new findings. Speaking with both @greentheonly and Euro NCAP’s Williams, Maric expanded into the investigation and found that the reason for the software alteration might actually be beneficial to Tesla customers.
According to the new story, Tesla vehicles tested by agencies like Euro NCAP feature a development software that was not deployed to customers’ fleets at the time of the tests. Green and Euro NCAP confirmed that they could not compare the development software with the production version to see what exactly is changed safety-related that could affect the test results. Initial reports speculated that code changes could affect how Tesla vehicles perform during the ADAS tests.
Nevertheless, although the cars were tested with the development code, the software would afterward be deployed to customers’ vehicles, improving safety for all. Euro NCAP investigation revealed that Tesla was indeed deploying the development software on the production vehicles following the tests.
“Euro NCAP allows manufacturers to update their vehicles up to the point they are tested. It is to the benefit of car buyers that manufacturers should seek to offer them the most highly-developed, best-performing systems available,” Williams said to CarExpert. “However, Euro NCAP requires that the cars tested should be the same as those in production at the time the safety rating is published.”
So far, Euro NCAP has been able to match the chronology of software updates for the Tesla Model 3, Model Y, and Model X. In all cases, Euro NCAP has found that the software updates have aligned with the expected timeline. Specifically, after Euro NCAP randomly selects the cars for testing, Tesla continues to improve the software. The software is frozen and installed on the test vehicles close to the start of Euro NCAP tests.
Results can take up to three months from the start of the tests until publication. It’s plenty of time for Tesla to introduce the development software into production. The improved software is pushed to vehicles via OTA before the Euro NCAP results are published. This means that Tesla used the Euro NCAP test results to improve its cars’ software further before the results were published. It also ensures that the software version used for tests is also on customers’ vehicles when the results are published.
Euro NCAP admits that it couldn’t verify the changes in the development code of the tested vehicles. Nevertheless, Tesla assured them that the safety-related elements in the development code were deployed to customers via software over-the-air updates. Williams confirmed to CarExpert that Euro NCAP would publish software versions for future tests to make the process more transparent.
According to an email Euro NCAP sent to CNN last week, Euro NCAP program director Aled Williams confirmed that the investigation had “so far” not revealed any evidence of Tesla’s attempt to cheat the tests. As for why Tesla introduced the additional ANCAP code into the software, Williams said it is only used to identify how a vehicle is configured for a particular region.
“Different regions (such as Europe, Australasia, etc.) differ in terms of legislation as well as road conditions/markings, etc. The recent addition of ANCAP to the code of Model Y coincides with the start of sales of that vehicle in Australasia,” Williams said.
Paul Maric, the CarExpert journalist who broke the initial news about the potential Tesla cheating code, has revised his original story to include the new findings. Speaking with both @greentheonly and Euro NCAP’s Williams, Maric expanded into the investigation and found that the reason for the software alteration might actually be beneficial to Tesla customers.
According to the new story, Tesla vehicles tested by agencies like Euro NCAP feature a development software that was not deployed to customers’ fleets at the time of the tests. Green and Euro NCAP confirmed that they could not compare the development software with the production version to see what exactly is changed safety-related that could affect the test results. Initial reports speculated that code changes could affect how Tesla vehicles perform during the ADAS tests.
Nevertheless, although the cars were tested with the development code, the software would afterward be deployed to customers’ vehicles, improving safety for all. Euro NCAP investigation revealed that Tesla was indeed deploying the development software on the production vehicles following the tests.
“Euro NCAP allows manufacturers to update their vehicles up to the point they are tested. It is to the benefit of car buyers that manufacturers should seek to offer them the most highly-developed, best-performing systems available,” Williams said to CarExpert. “However, Euro NCAP requires that the cars tested should be the same as those in production at the time the safety rating is published.”
So far, Euro NCAP has been able to match the chronology of software updates for the Tesla Model 3, Model Y, and Model X. In all cases, Euro NCAP has found that the software updates have aligned with the expected timeline. Specifically, after Euro NCAP randomly selects the cars for testing, Tesla continues to improve the software. The software is frozen and installed on the test vehicles close to the start of Euro NCAP tests.
Results can take up to three months from the start of the tests until publication. It’s plenty of time for Tesla to introduce the development software into production. The improved software is pushed to vehicles via OTA before the Euro NCAP results are published. This means that Tesla used the Euro NCAP test results to improve its cars’ software further before the results were published. It also ensures that the software version used for tests is also on customers’ vehicles when the results are published.
Euro NCAP admits that it couldn’t verify the changes in the development code of the tested vehicles. Nevertheless, Tesla assured them that the safety-related elements in the development code were deployed to customers via software over-the-air updates. Williams confirmed to CarExpert that Euro NCAP would publish software versions for future tests to make the process more transparent.
Resold Tesla computers from lab-based crash tested vehicles contained modified developer firmware. @EuroNCAP investigated and the result wasn't what we expected. Long story short - customers are winners. Full details: https://t.co/JcIUpXjxEQ @greentheonly
— Paul Maric (@PaulMaric) October 12, 2022
Google News