What Uber's autonomous driving program does

Uber has invested in self-driving car technology through a division called Uber Advanced Technologies Group (ATG), which develops and tests vehicles that can operate without a human driver. These cars use cameras, radar, and lidar (a laser-based sensor) to detect other vehicles, pedestrians, and road conditions. The goal is to eventually run rides without a person behind the wheel, though the program has faced technical challenges and regulatory hurdles that have slowed its rollout.

As of now, Uber's autonomous vehicles are not widely available to regular passengers. The company has conducted limited testing in certain cities and has partnered with other companies to develop the technology. Most Uber rides you take today still require a human driver. Understanding where the technology stands and what it means for the future of rideshare can help you know what to expect.

Key Takeaways

  • Uber's self-driving program uses cameras, radar, and lidar sensors to navigate roads without a human driver controlling the vehicle.
  • The technology is still in testing phases and is not yet available for regular passenger rides in most places.
  • Uber has faced technical setbacks, accidents during testing, and regulatory delays that have slowed the program's progress.
  • Different states and cities have different rules about where and how self-driving cars can be tested on public roads.

How the sensors and technology work together

Self-driving cars rely on multiple types of sensors working at the same time. Cameras mounted on the vehicle capture images of the road, other cars, traffic lights, and pedestrians. Radar sensors detect moving objects and their speed, even in fog or heavy rain when cameras struggle. Lidar shoots out laser pulses and measures how long they take to bounce back, creating a detailed 3D map of everything around the car.

All this sensor data feeds into a computer system that makes decisions about steering, acceleration, and braking. The system has been trained on millions of miles of driving data so it can recognize patterns — a person stepping into the street, a traffic light turning red, a pothole to avoid. However, the technology still struggles with unusual situations, like a police officer directing traffic instead of a traffic light, or a partially blocked road.

Where Uber has tested self-driving cars

Uber has conducted autonomous vehicle testing in several U.S. cities, including Pittsburgh, San Francisco, and Phoenix. In Pittsburgh, the company ran a limited program where passengers could request self-driving rides, though a safety driver sat in the front seat ready to take control if needed. Testing in San Francisco faced regulatory challenges and was eventually suspended. Phoenix saw more extended testing through a partnership with local authorities.

The company has also worked with other organizations and sold or transferred parts of its autonomous technology to other companies. In 2020, Uber sold its self-driving car division to a company called Aurora Innovation, though Uber continues to work with Aurora on autonomous technology. This means Uber's direct involvement in building self-driving cars has shifted, but the company remains interested in using autonomous vehicles for rides in the future.

Regulatory requirements and testing rules

Each state and city sets its own rules about how and where self-driving cars can be tested. Some places require a human safety driver in the vehicle at all times. Others demand that companies report accidents and technical failures to regulators. A few cities have allowed limited public testing where passengers can ride in the vehicle, but these programs are rare and heavily monitored.

The National Highway Traffic Safety Administration (NHTSA) sets federal safety standards, but it does not have a single approval process for autonomous vehicles the way it does for new car models. This patchwork of rules means a company might be able to test in one city but not in another, or might need to modify its system to meet different state requirements. This regulatory uncertainty has slowed Uber's progress and contributed to the company's decision to focus less directly on building the technology itself.

Accidents and safety concerns during testing

Uber's self-driving cars have been involved in accidents during testing. In 2018, an Uber autonomous vehicle struck and killed a pedestrian in Tempe, Arizona — the first death involving a self-driving car in the United States. An investigation found that the vehicle's sensors detected the pedestrian, but the system did not brake in time. The safety driver in the vehicle was also not paying full attention. This incident raised serious questions about whether the technology was ready for public use.

After the accident, Uber suspended its autonomous vehicle testing program for several months and made changes to its safety protocols. The company added more rigorous testing requirements and improved its safety driver training. However, the incident highlighted a core challenge: self-driving cars must perform reliably in unpredictable real-world conditions, and even small failures can have serious consequences. This is one reason why the rollout of autonomous rideshare has been much slower than some early predictions suggested.

What this means for Uber riders today

If you use Uber now, your rides are still driven by people. The company has not launched a widespread autonomous rideshare service, and there is no clear timeline for when that might happen. In most cities, you will continue to request a ride through the Uber app and be matched with a human driver. The autonomous vehicle program remains a long-term research and development effort rather than a service you can currently book.

That said, the technology continues to improve, and Uber's partnerships with companies like Aurora mean the company is still working toward autonomous rides. If you live in a city where testing is happening, you might see self-driving cars on the road, but they are being monitored by researchers and safety drivers, not carrying regular passengers. For now, understanding how the technology works helps you make sense of news about self-driving cars and what the future of rideshare might look like.

How Uber's approach differs from other companies

Other companies like Waymo (owned by Google's parent company Alphabet) and Tesla are also developing self-driving technology, but they have taken different paths. Waymo has focused on building its own vehicles and has launched a limited robotaxi service in Phoenix called Waymo One, where passengers can request fully autonomous rides. Tesla has focused on adding self-driving features to its own cars that owners can purchase, though these features still require driver attention and are not fully autonomous.

Uber's strategy shifted after the 2018 accident and the regulatory challenges. Rather than building its own self-driving cars from scratch, the company decided to partner with and invest in other companies developing the technology. This means Uber is betting on autonomous technology without bearing all the development costs and risks itself. It also means Uber's path to autonomous rideshare depends partly on whether its partners succeed.

Frequently Asked Questions

Can I request an Uber self-driving car right now?

No. Uber does not currently offer autonomous rides to regular passengers in most places. The company's self-driving program is still in testing phases with limited availability in select cities. When you request an Uber, you will be matched with a human driver.

What happened to Uber's self-driving car division?

Uber sold its Advanced Technologies Group to Aurora Innovation in 2020. Uber continues to work with Aurora on autonomous technology, but the company no longer directly builds self-driving cars. This partnership allows Uber to benefit from autonomous technology without managing all the development itself.

Why did Uber stop testing self-driving cars?

Uber suspended testing after a fatal accident in 2018 when one of its autonomous vehicles struck a pedestrian in Arizona. The company paused to review safety protocols and improve its systems. Testing has resumed in limited form, but the incident showed that the technology was not yet ready for widespread public use.

Will Uber rides ever be fully self-driving?

Uber has stated that autonomous rideshare is a long-term goal, but there is no confirmed timeline. The technology continues to improve, and regulatory rules are slowly becoming clearer, but significant technical and legal challenges remain. It may be years before self-driving Uber rides become common.

How do self-driving cars know where to go?

Self-driving cars use GPS for general navigation and detailed maps of city streets. Sensors like cameras and lidar provide real-time information about the road, other vehicles, and obstacles. A computer system processes all this data and makes decisions about steering and speed to reach the destination safely.