Could You Pass a Driverless Car's Driving Test?

How do robotaxis make decisions? Abby investigates self-driving technology and challenges readers to take a driverless driving test.

Imagine climbing into a car.

There's no driver.

No steering wheel.

And no pedals.

Then the car starts moving.

It sounds like something from a science fiction film, doesn't it?

But when I experienced a Zoox robotaxi, it made me think about a much bigger question.

**How does a car know what to do when there isn't a human driving it?**

### How does a robotaxi see?

A self-driving vehicle doesn't look at the road in quite the same way we do.

Instead, autonomous vehicles can use several types of sensors, including cameras, radar and lidar.

Cameras help recognise visual information such as road markings, traffic lights and surrounding objects.

Radar uses radio waves to detect objects and estimate their movement.

Lidar sends out light pulses and measures their reflections to help create a detailed representation of the surroundings.

Computer systems combine information from these sensors to help the vehicle understand where it is and what's around it.

But understanding the road is only the first challenge.

The next challenge is deciding what to do.

### What happens when things go wrong?

Imagine a robotaxi approaching a pedestrian crossing.

The traffic light is green.

But somebody is still crossing the road.

Should the car continue because it has a green light?

Of course not.

It needs to respond safely to the pedestrian.

Now imagine the robotaxi reaches roadworks.

One lane is closed, and temporary signs direct traffic into another lane.

What should it do?

A human driver would look at the signs, check for other vehicles and follow the temporary instructions.

A robotaxi needs to interpret the situation using its technology and operating rules.

### Can a robotaxi predict the future?

Not exactly.

But it can make predictions about what other road users might do next.

For example, a cyclist looking over their shoulder might be preparing to change lanes.

A car slowing near an intersection might be about to turn.

These are clues that a driving system may need to consider.

But predictions can be wrong.

That's why a self-driving vehicle must be designed to manage uncertainty and avoid unsafe assumptions.

### Are driverless cars safer?

That's a surprisingly difficult question.

Different autonomous systems operate in different places and under different conditions.

Safety depends on the technology, testing, the operating environment and how incidents are measured.

Some driving tasks may be performed very reliably by automated systems. Other situations remain challenging.

We shouldn't assume all robotaxis are equally capable just because they look futuristic.

### Your driving test

Now imagine you are responsible for programming a driverless car.

What should happen when:

- A ball rolls into the road? - A cyclist suddenly swerves? - The traffic lights stop working? - A pedestrian looks like they might step onto a crossing? - Heavy rain makes road markings difficult to see?

These situations make something clear.

Driving isn't simply about moving a car from A to B.

It's about making safe decisions in a world full of surprises.

### My final thought

The part I find most interesting about driverless vehicles isn't the fact that they can move without a steering wheel.

It's how much intelligence, testing and careful engineering is needed to make thousands of small decisions along the way.

### Abby's Big Question

**Should a driverless car be required to prove it is safer than a human driver before being allowed on public roads?**