Uber's attempts at world domination have been slowed by the news that their autonomous vehicles are not as clever as they think. It seems that drivers have to take control more often than was first thought.
The Uber story has been fraught with lawsuits and regulatory bodies but they have persevered with their autonomous trials. Recent reports indicate that the cab-hailing company is also having technical issues with their vehicles.
It seems that Uber's 43-car self-driving fleet in tests in Pennsylvania, Arizona and California have covered many thousands of miles every week, but the vehicles are actually requesting human assistance way too often for autonomy to be truly effective, according to a report in Recode.
Information that allegedly comes from Uber's self-driving division shows that the engineer/drivers whose task it is to evaluate progress in this cutting-edge technology have not achieved the results they hoped. A human intervention was required every 1.45 kilometres during January 2017 trials. Two months later the results were, if anything, fractionally worse.
This is not to say that test drivers were experiencing a potential accident every mile. That particular statistic has actually improved. Critical driver interventions have been much better; up from one every fifty miles to one every 320 kilometres.
Issues and anomalies arose with difficulties navigating lane markings or missing turns, as examples. This demonstrates that autonomous technology is not in itself enough. It also requires a universal improvement in the conditions of roads, unrecognized street furniture and the like.
Rival company Waymo, Alphabet's self-driving company, has been working on its own autonomous technology since 2009. However they too still experience driver interventions. The competition cannot be said to be equal though. The two companies are at very different stages of development. As far as Uber is concerned this technology, once perfected, will be vital to overall profitability.
