MOUNTAIN VIEW, CALIFORNIA • Google's self-driving vehicles are mastering complex situations on public roads, from cars going the wrong way to bicycles darting in front of traffic, as the technology company strives to win the highprofile race to achieve full vehicle automation, executives said on Wednesday.

The Alphabet Inc unit, which has been developing autonomous cars since 2009, said its self-driving vehicles have logged 3.2 million km on public roads. It continues to log about 40,000km of test drives a week.

Google has been a leader in testing autonomous vehicles on public roads, prompting traditional carmakers to step up their self-driving plans to avoid being eclipsed by the technology leader, which has yet to disclose a business strategy for its car project.

The Silicon Valley-based company is focused on making cars fully autonomous, with no need for a driver, which could make driving safer and more efficient and open up transport to the disabled and aged.

It said last year that such cars would be ready for production by 2020.

Its approach stands in sharp contrast to that of many carmakers, including Toyota Motor and Tesla Motors, which are moving towards autonomous driving in incremental steps, now still requiring drivers.

Tesla has said that it has logged more than 160 million km since last October, from drivers using its partially autonomous Autopilot system.

But kilometres driven on predictable highways are easier than navigating busy city streets, said Mr Dmitri Dolgov, head of Google's self-driving technology effort.

Google's cars experience more complex situations than just following a car on the freeway, he added.

"You have to have a deeper understanding of what's on the road and to the side," he told Reuters, following a ride in a self-driving Lexus near Alphabet's Mountain View, California campus, during which the vehicle reacted to individuals in wheelchairs, people pushing strollers and a car making a U-turn. "Every time you drive, it's different."

The Google car can detect whether a moving object is a child or a bicycle and anticipates that both can make fast, unpredictable movements. The car's driving system calculates the probability of such movements and uses the results to determine how the vehicle will react.

Those types of complicated social interactions are the last, most difficult element of autonomous vehicle technology, said Mr Dolgov.

"You get to 90 per cent autonomy in 10 per cent of the time and then spend 90 per cent of your time on the last 10 per cent," he said.

The closest company to rival Google's self-driven distance in California is auto supplier Delphi Automotive, with 26,800 autonomous km, according to a January report filed with state regulators.

REUTERS