Stories from the road: Safety readiness case studies

As Aurora continues to hit milestones and prepare for the commercial launch of our Aurora Horizon autonomous trucking service, we’re keeping our industry-leading safety approach at the forefront of development and deployment. 

Core to this approach are our frequent autonomous trips on major freight routes in Texas, where we navigate a variety of interesting, novel scenarios on the road and use these experiences to continuously improve the Aurora Driver. 

This includes instances in which external factors, like aggressive road users or hazardous debris, compromise the safety of the on-road environment and require the Aurora Driver to respond adeptly and safely—keeping our vehicles and other road users out of harm’s way while maintaining a reliable service for our customers.

Let’s dive into a few recent on-road encounters, examining how the Aurora Driver handled them and how we’re prioritizing safety as we progress toward commercial launch.

Case study #1: Responding to a collision

Operating at scale means navigating more than just the typical scenarios we see daily while hauling commercial loads for our pilot customers—we also need to handle rare situations, including situations where another driver collides with an Aurora Driver-powered semitruck. Even the safest drivers are vulnerable to other drivers’ risky behaviors, so we’re preparing the Aurora Driver to respond accordingly.

A vehicle sideswipes an Aurora Driver-powered truck from the left lane. The Aurora Driver recognizes the collision within milliseconds of impact, and the onboard vehicle operator took control, safely decreased speed, and pulled over.

In this case, a vehicle sideswiped one of our Aurora Driver-powered trucks on the freight route between Fort Worth and El Paso. As the event unfolded, the Aurora Driver detected the incoming vehicle and began to move away from it by entering the shoulder of the road. The vehicle continued to veer toward our truck at over 65 miles per hour, causing a collision. 

The Aurora Driver detected the imminent collision and our onboard vehicle operator took control of the truck, safely decreased speed, and pulled over. When deployed without a vehicle operator, the Aurora Driver will execute this response autonomously while a Command Center Specialist contacts first responders.

After confirming the safety of our team, we shared information about the incident with law enforcement, partners, and regulators. All of these actions align with our organizational preparation for scenarios like this, and as part of our commitment to transparency, we are now sharing information about the event with the public. 

Collisions on the road can be scary and disorienting. Backed by our rigorous approach to safety in the development of our technology and on-road operations, the Aurora Driver has the capabilities needed to respond to collisions and our organization is prepared to handle these events in a safe and responsible way.

Case study #2: Avoiding dangerous debris

Nearly every driver has had to deal with unexpected debris on the road, and the Aurora Driver is no different. We’ve shared in the past how we trained our technology to respond to road debris, ranging from tires to tumbleweeds, safely. 

The Aurora Driver recognizes a ladder in the road over 150 meters away and proactively maneuvers around it.

Recently, one of our autonomous trucks encountered a ladder in the road—a large object that could potentially damage our truck or other vehicles. In this video, the Aurora Driver detected the ladder from a significant distance away and proactively maneuvered around it to avoid a collision. This smooth, natural behavior is consistent with how trained commercial drivers would have handled this scenario.

Case study #3: Sharing the road with aggressive drivers

The Aurora Driver is designed to drive safely and reliably as we transport goods and passengers in autonomous vehicles. However, other drivers don’t always demonstrate the same level of care.

An aggressive vehicle cuts off our autonomous truck, prompting the Aurora Driver to slow down to provide additional room on the road.

In this case, an Aurora Driver-powered truck encountered a driver aggressively speeding and swerving between lanes. In the video above, the other vehicle cuts our truck off with less than 3.5 feet of clearance before nearly colliding with another vehicle the next lane over. The Aurora Driver detects the incoming vehicle and proactively slows down as the vehicle swerves past. 

We’ve trained the Aurora Driver to react quickly to vehicles driving recklessly based on how trained commercial drivers respond to similar scenarios. 

Continuous improvement in simulation

In each of the situations outlined above, our autonomous vehicles responded safely and appropriately. Still, we are committed to continuously improving the Aurora Driver. 

We frequently practice driving encounters like these in our Virtual Testing Suite using simulated camera, lidar, and radar data to evaluate how the Aurora Driver would respond in these scenarios and variations on these scenarios that we haven’t yet encountered on the road. Virtual Testing enables us to test our technology against thousands of permutations of potentially dangerous scenarios with modified variables like speed, the position and type of other actors, and more.

We test variations of on-road scenarios in our Virtual Testing Suite. In this case, we simulated sideswipe variations to test the Aurora Driver’s ability to detect a collision.

Our deep investments in our Virtual Testing Suite enable us to squeeze additional insights out of our real-world encounters, quickly maturing the Aurora Driver’s capabilities. 

Ultimately, as we prepare to deploy autonomous trucks broadly with our partners and customers, the Aurora Driver will encounter sideswipes, debris, and aggressive drivers more often. Handling these situations confidently and responsibly is essential for safe operations at scale.