Follow
Subscribe

How Will Self-driving Vehicles Keep Passengers Safe in Natural Disasters?

Home > News > Content

【Summary】A feasible solution to preparing autonomous vehicles for natural disasters could be virtual testing software. Using this option, developers can recreate a wide range of emergencies for driverless cars to tackle in a controlled environment.

Michael Cheng    May 23, 2019 5:00 AM PT
How Will Self-driving Vehicles Keep Passengers Safe in Natural Disasters?

In the future, self-driving cars will be deployed in locations that frequently experience earthquakes, floods and blizzards. How the vehicles assess such risks and respond to changes to its surroundings will play crucial roles in keeping passengers safe during emergencies.

Auto companies are in the process of developing safety protocols that dictate how driverless cars react to the presence of natural disasters and chaos. Perhaps a simple example could be an autonomous vehicle encountering never-before-seen, temporary construction signs on the road. Will the car slow down? Or will it look for other routes to avoid potential disruptions ahead?

Defensive Strategies in Natural Disasters

Waymo recently provided insights on how autonomous cars are being developed to take on such challenges. The establishment emphasized that its driverless vehicles would implement defensive, cautious actions when facing natural disasters. For instance, instead of attempting to navigate safely through icy roads or floods, the car may reroute itself or pull over and wait for rough conditions to pass.

The company's self-driving fleet continuously learns about its environment, monitoring for changes that could make driving highly risky. Moreover, the information can be sent to nearby Waymo vehicles via real-time notifications over connected networks. This safety feature is particularly useful for avoiding unforeseen road closures, city events, major car accidents and more.

"We're doing parts of the technology development that will help us get there," explained Dmitri Dolgov, Chief Technologist at Waymo, during a press event.

"There are a number of signals we can use to figure out, just like a human driver, what to do in that situation."

Waymo's autonomous vehicles contain external hardware that are capable of closely monitoring changes to the car's surroundings. Every unit is equipped with six LIDAR sensors, which improve 3D visibility in low-light (dark) and extremely bright environments. The driverless cars also use radar for robust detection in foggy conditions and snow. Cameras around the vehicle provide clear images of the environment and road.

Experience through Virtual Testing

According to Waymo's Chief Technologist, its autonomous fleet have already been exposed to a handful of unforeseen elements that occur during emergencies. Such activities, including broken traffic lights (resulting from damaged infrastructure) and car accidents on roads, are very unpredictable as they do not usually appear on maps. Because of this, driverless cars may require functions that are effective in real-time.

"You have to reason through these kinds of new situations in real time, the first time you're seeing them," highlighted Dolgov.

A feasible solution to preparing autonomous vehicles for natural disasters could be virtual testing software. Using this option, developers can recreate a wide range of emergencies for driverless cars to tackle in a controlled environment. Time and cost savings are the advantages of using virtual testing methods to hone the skills of self-driving vehicles. After creating the driving scenario, developers could replay the situation as many times as needed, at different speeds. Lastly, virtual testing can be conducted at any time, as the solution is not dependent on outdoor conditions.

Prev                  Next
Writer's other posts
Comments:
    Related Content