Aurora Labs flags autonomous vehicle software vulnerability

Software code errors seen becoming expensive and a danger to fleet operators

Aurora Labs flags autonomous vehicle software vulnerability
Automated trucking may need to deal with software errors


While the issue of autonomous and platooning truck hacking has lost some steam as a major industry concern, an equally fraught issue, software glitches, is being given air by Israeli IT firm Aurora Labs.

The company, which focuses on predictive software maintenance and has patented what it calls Line of Code Maintenance, argues that the potential for software to cause crashes of a much more violent and deadly variety is rising to the top of the transportation industry’s agenda this year following a series of recalls linked to flawed code.

The increasing number number of glitches with software in vehicles is a good example of how such digital transformations can often spawn new and equally challenging problems, according to Aurora Labs chief executive Zohar Fox.

"You can imagine the downtime of a fleet of autonomous trucks if you have a recall because of software," Fox says.

"This is a topic that’s a new problem in this industry and it will change how we see operations."

Read how truck hacking was dealt with in Australia here

In a report Aurora highlights in US trucking news site, Fox points to US software glitches that led to Fiat Chrysler recalling 229,000 Ram pickup trucks, Toyota’s 65,000 Tundra and Sequoia pickups BMW 11,700 cars and concerningly for the trucking industry there, Paccar’s February recall of 1,938 prime movers over an inaccurate gear engagement display.

That notice involves 2017-2018 Kenworth T680 and T880 trucks, built between November 16, 2016, and December 22, 2017 and Peterbilt 567 and 579 units built between November, 17, 2016 and December 31, 2017. It was the second such this year and affected Eaton Endurant and Paccar AMT transmissions.

Fox asserts that while a passenger vehicle may have 150 million lines of software code, it is possible, generally, for there to be 50 errors in every 1,000 lines of code.

The company’s pitch in the article is that software that is self-repairing is likely to be more effective that attempting to put a fix in remotely, on the road.


You can also follow our updates by joining our LinkedIn group or liking us on Facebook