(c)iStock/Nancy Nehring

Many drivers, learners in particular, hate complicated maneuvers such as reversing around corners and three-point-turns.

But Google has said in its latest self-driving car update that it’s teaching its cars to handle these sorts of moves.

When manually driving a car, it’s sometimes difficult - especially down narrow streets lined with other cars - to perform a three-point-turn. Not being able to gauge the distance properly means it can sometimes turn into a six or even 10 point turn.

However, Google has made a good point about autonomous vehicles: the connected sensors mean many self-driving cars in the future will be able to calculate the distance down to a few centimetres, see in 360 degrees and calculate the most efficient path for the car.

It’s not quite as simple as that, however. Self driving cars can calculate the quickest option of performing a three-point turn, but Google said it wants the cars to mimic the way we drive naturally.

User experience

This means instead of driving all in reverse, for example, the car copies the way we would manually steer out of a three-point-turn, while still being quick and accurate.

In addition, Google said it had taught its self-driving cars to adapt to “all kinds of variables”, including dead-end streets stacked with parked cars, litter bins and narrow bottlenecks.

The cars are performing 1,000 multi-point turns every week on real city streets, an environment which is necessary for training driverless cars, but which also leaves them more vulnerable to accidents.

The same monthly report on Google’s three-point turns also reported a very small collision between a driverless car and manual-driven one.

When Google’s driverless vehicle stopped to yield to traffic and then started to move forward again, another car bumped it from behind. Both vehicles were going very slow, only sustained minor damage and there were no injuries.