Could autonomous cars eliminate traffic?

Apart from Tom Cruise in Mission Impossible III, it’s hard to find someone who gets excited about traffic jams. But here’s some (potential) good news – as more driverless vehicles hit the road, a group of researchers believes they could be eliminated completely.

The hypothesis: That AI could be harnessed to develop algorithms to control complex mixed traffic systems, which are defined as having both human-driven and driverless cars on the road simultaneously.

The researchers utilized a branch of AI known as reinforcement learning to develop an algorithm that sets rewards for driverless vehicles to prioritize goals like traffic efficiency or energy consumption. The algorithm allows these robocars to communicate with each other to optimize traffic flow – and because they’re interspersed with cars driven by actual people, all traffic ends up being affected.

In a simulation outlined in a 2023 study:

• The researchers found when driverless vehicles make up just 5% of traffic, traffic jams are eliminated.
• When robot vehicles make up 60% of traffic, traffic efficiency is superior to traffic controlled by traffic lights.

But… We’re a long way away from this reality👆. In China, hundreds of autonomous vehicles deployed by tech giant Baidu are actually causing traffic jams by driving too cautiously. And in San Francisco and Phoenix, two areas with a commercially available driverless-taxi service, autonomous vehicles have driven the wrong way down the road and been involved in multiple crashes (one leading to a Waymo recall of 672 self-driving vehicles).