50 Best Workplaces of the Year 2019

Argo AI, a Pittsburgh-based Technology Company, is Building Self-Driving Technology to Improve the Way the World Moves

thesiliconreview-bryan-salesky-ceo-argo-ai-19“We’ve been building self-driving technology to provide a safer, more affordable, convenient, and accessible way for everyone to get around.”

Today’s roadways are congested and the cost of car ownership is rising. And while new technologies are improving the driving experience, human errors and choices – such as distracted driving, drunk driving, and speeding – continue to claim lives every day.

In light of the foregoing, we’re thrilled to present Argo AI.

Argo operates as an artificial intelligence company. The Company develops applications with computer science and robotics for self-driving vehicles. Argo serves customers in the States of Pennsylvania, Michigan, and California.

The company is developing ‘self-driving’ technology according to the strictest interpretation of what the Society of Automotive Engineers (SAE) calls “Level 4” automation: a system that can drive a vehicle — under specific conditions — that does not require passenger supervision.

Argo AI was incorporated in 2016 and is headquartered in Pittsburgh, Pennsylvania.

Argo AI: Synopsis

Factors that Makes Argo AI Stand Out

A Commitment to the Best in Engineering: It takes commitment to high-quality, disciplined engineering in order to make this technology available at scale. Argo is designing and building an automotive grade self-driving system from the ground up, suitable for commercial operation.

Extensive Testing: Through simulation and real-world testing on closed courses and public roads, Argo puts its technology through the world’s most rigorous driving school every day.

Prioritizing Experience: No one likes a bumpy ride! Argo is paying close attention to the experience it delivers by designing a system, which is safe and confident. No abrupt stops – unless needed to ensure safety. The company drives with the flow of traffic.

The Testing Process –

While testing on public streets gets the bulk of the attention, it’s just one part of Argo’s process. On-road testing doesn’t begin in any city until the company has completed rigorous development in the lab, via computer simulation, and on closed courses. A culture of safety means the process never ends. Argo calls it ‘Continuous Testing’. Here’s what that looks like:

Development Testing: You can’t build a self-driving system without first knowing how individual parts work, so the first place Argo tests anything is in a lab. From the radar, camera and lidar sensors to the computer hardware and software running on it, everything is individually tested, then tested as a system.

Simulation Testing: Here Argo creates a virtual world where it tests out a wide variety of scenarios. Argo simulates environments as small as a single street to as large as an entire city, into which the company put virtual cars running its software to the test.

Because Argo can run multiple simulations at once, it can test tens of thousands of scenarios in the time it would take to plan, set up, and test just one scenario in the real world.

If Argo makes changes to its hardware or software, it re-simulates to make sure those changes don’t undo earlier improvements. This is called regression testing, and it’s absolutely essential to developing a self-driving system robust enough for human passengers.

Closed Course Testing: Once its self-driving software has passed simulation testing, Argo takes it to a private track staffed by trained professionals the company calls Test Specialists. A closed course allows Argo to safely test whether software behaves as it did in simulation.

The company uses all kinds of tools to help replicate what it may encounter on the road, from inflatable pedestrians and fake dogs to remote-controlled skateboarders and baby strollers. If Argo observes behavior it doesn’t like, the software is back to development and simulation testing before returning to the track to try it all over again.

Once the software passes closed course testing, it’s time to begin testing on the street.

Real World Testing: Why do we test on public roads? It’s not just to try out the software to see how it behaves — that’s what closed course testing is for. It’s because the world is more complicated than any simulation or track can ever be, and real products have to be tested in the real world.

Testing on public roads is a privilege Argo takes very seriously. It starts by abiding by all applicable laws, regulations and guidelines in the cities where the company operates. Argo tests in multiple cities because every city is different. A person who got their license in Tulsa, Oklahoma, has a lot to learn about driving in New York City traffic. Traffic laws change, infrastructure like roundabouts and crosswalks differ, and cultural differences can radically alter the relationship between cars, bicyclists and pedestrians.

Luckily, self-driving cars have two huge advantages over people: the technology remembers everything it sees and learns, and lessons learned are shared across the fleet.

When Argo decides to test in any given city, Test Specialists begin by manually driving through the city, gathering data to build 3-dimensional maps of the streets where the company intends to operate. Once the map is ready, Argo can begin testing in autonomous mode.

The more information Argo collects, the more it can solve, and the safer its next iteration of software becomes.

In time, Argo’s self-driving system — always vigilant, always learning, never forgetful — will become a better driver than any human.

Bryan Salesky: A Formidable Leader

Bryan is passionate about incorporating promising robotics technology into products and systems that will improve safety and productivity while enhancing people’s lives. While at Carnegie Mellon University’s National Robotics Engineering Center (NREC), Bryan managed a portfolio of the center’s largest commercial programs, including autonomous mining trucks for Caterpillar. In 2007, Bryan led software engineering for Tartan Racing, Carnegie Mellon’s winning entry in the DARPA Urban Challenge. Bryan departed NREC and joined the Google self-driving car team in 2011 to continue the push toward making self-driving cars a reality. While at Google, Bryan was responsible for the development and manufacture of their hardware portfolio, which included self-driving sensors, computers and several vehicle development programs. Bryan graduated from the University of Pittsburgh with a bachelor’s degree in computer engineering in 2002.

“We believe we can significantly reduce the number of incidents on roads today by replacing the driver behind the wheel.”