Hit enter to search or ESC to close

Testing the Edge: The Mud Cannon Powers a Down-and-Dirty Solution for Self-Driving

This is the first in our “Testing the Edge” series, focusing on ingenious ways that the Structured Test team at Argo AI tests different components, systems, and scenarios for self-driving cars. 

To prepare for the highly-nuanced world a self-driving car will encounter, engineers who develop and test autonomous vehicle technology are inspired to become very creative. That’s how an innovative and quirky apparatus with a name like “Mud Cannon” comes into being. 

As part of their work to test solutions for the car’s sensor-cleaning systems, Integration and Test Engineers at Argo AI have created a method to throw mud in the “eyes” of autonomous vehicles.  At a closed-course test track outside Pittsburgh, Argo wanted to simulate the experience of mud from a passing vehicle being splashed on one of the car’s sensors, such as lidar or a camera. With nothing available off the shelf to fire the mud, one engineer manufactured a “pneumatic powered projectile device.” Thus, the Mud Cannon was born. Positioned atop the hood of an Argo test vehicle, the device fires a muddy mix (three parts fine filter-testing dust, one part water) at the car’s roof-mounted cameras.

The cannon is made up of a PVC pipe fitted with a reservoir and barrel, plus sprinkler valves and a portable quiet air compressor, all purchased from various local hardware stores. It took about six weeks and multiple prototypes to perfect it, and as the cannon design evolved, pipe fittings were added to control pressure, along with a filter to remove unwanted debris, and suction cups to attach the device to the car. The end result is like a “simple potato gun, with a sprinkler valve.”  

The Mud Cannon Test

Because sensor performance and sensor cleaning are critical to safe and reliable self-driving, a long list of sensor-related test requests is sent to the Argo Structured Test team. These engineers work at the Argo test track, designing and carrying out repeatable tests to explore the performance and limits of the Argo self-driving system (SDS).

“We get a number of requests from software developers and engineers who want a specific camera lens to be obstructed, and we test all of these edge cases on an autonomous vehicle to make sure it’s safe to operate on public roads,” says Test Technician Jutaporn Huesman. 

Where, when, and how fast the vehicle is driving all affect the level of soiling and its impact on a sensor. Among the most challenging scenarios is water and mud sprayed from passing vehicles on the road. Once the day’s testing has been scheduled, the team loads up the Mud Cannon with the mixture, pressurizes it, and sends the vehicle out for the test.

As the vehicle reaches the given point on the test track, the right seat operator, either a Test Technician Associate or Test Technician, pulls the trigger to fire the mud at the appropriate camera lens. The response of the SDS is carefully recorded and relayed to a number of Argo engineering teams, including those who develop obstruction-detection algorithms, and the labelling team, whose work helps the SDS identify lens obstruction.

A Mucky Business

Of course, working with mud is a dirty business, and it’s also surprisingly difficult to fire mud accurately at a specific sensor on a moving vehicle. Test engineers have to be cognizant that the pressure of the delivery isn’t too high because it might damage the lens.

The “mud”—a carefully developed mix of water and a specific test dust chosen because its particle size doesn’t scratch the lenses of sensors—is fired at the car at approximately 4 mph. Interestingly, the lenses are coated in an oleophobic coating which repels dirt and oil, so if they were to get the mud mix wrong, it would just slide right off. The team has dubbed the right mix “Snack Pack Pudding” because of its texture.

The Mud Cannon has become an essential tool for the Argo Structured Test team, and the results it produces help engineers understand the effects of sensor soiling and further improve its sensor-cleaning systems. It has also inspired Argo engineers to consider other devices, including a larger version to shoot leaves and feathers.

Perhaps most surprising, however, is that no one on the team has been pranked with a mud cannon shower. 

Choose your lane

How Autonomous Vehicles Distinguish Between Bicycles and People Who Ride Them

How Autonomous Vehicles Distinguish Between Bikes and People

When it comes to how autonomous vehicles see the world, humans come first, literally. Autonomous vehicles (AVs), like the kind operated by Pittsburgh-based Argo AI, use Machine Learning to detect and classify the objects in their surroundings, identifying people...
Why The League of American Bicyclists is optimistic about autonomous vehicles

Why a Leading Cycling Advocacy Group Is Optimistic About Autonomous Vehicles

As autonomous vehicle use grows, AV companies and the League of American Bicyclists are collaborating on how to ensure cyclists and motorists can share the roads safely, even if the “motorist” is artificial intelligence software. As part of the...

Self-Driving Is Arriving Right On Time. Just Like Ice Cream Did

Seven years ago, I was a self-driving skeptic. Not of the technology. Of all the “experts” promising autonomous vehicles would be everywhere by 2020. You didn’t need to be Nostradamus to know that was ridiculous. All you needed was...
Illustration of a futuristic parking deck turned into a mixed-use space, with AVs driving by

How Autonomous Vehicles Could Help Transform Parking Lots

Researchers say it’s likely that autonomous vehicles (AVs) can help reduce the need for parking lots, opening more room for grass and trees and other elements of nature. It may not seem like it when you’re circling the block...
An illustration of an Argo autonomous vehicle in teal against a backdrop of buildings, a bicyclist, and research papers

7 Big Breakthroughs From Argo Research at CVPR 2022

The 2022 Conference on Computer Vision and Pattern Recognition (CVPR 2022) is nearly here. Thousands of computer scientists, software engineers, and researchers from around the globe will gather in New Orleans to review and discuss their latest work in...

Researchers Predict the Future With Lidar Data

Researchers at Carnegie Mellon University’s Argo AI Center for Autonomous Vehicle Research, a private-public partnership funded by Argo for advancing the autonomous-vehicle (AV) field, say they have come up with a way to use lidar data to visualize not...

Must Reads