Todd Jochem on the evolution of autonomous vehicles, the state of self-driving technology and Carnegie Mellon’s lasting impact
Todd Jochem (SCS 1993, 1996) has been in the driver’s seat of autonomous vehicles for decades. Long before “self-driving car” became a household phrase, he was helping to demonstrate the technology could work beyond the lab.
In 1995, Jochem helped complete a 2,850-mile journey from Pittsburgh to San Diego, known as “No Hands Across America,” that was no typical summer road trip. For more than 98% of the drive, the vehicle steered itself. The driver was a computer system called Rapidly Adapting Lateral Position Handler (RALPH) operating inside Navlab 5, part of a series of autonomous vehicles developed at Carnegie Mellon University’s Robotics Institute beginning in 1986.
Now, armed with a deep understanding of how autonomous vehicle technology has evolved, Jochem envisions two-seater robotaxis, driverless buses, and cars without steering wheels or pedals. He believes self-driving vehicles could expand mobility for millions if the industry can navigate the technical and societal challenges ahead.
Jochem recently released “Driven,” a memoir about his work on self-driving cars in which he revisits early breakthroughs in autonomy. In the conversation below, he reflects on the early days of autonomous vehicle research, the conditions of self-driving technology today, what continues to drive him and how CMU’s longstanding commitment to building technology with real-world impact continues to shape the field.
What still feels like a challenge in autonomous driving?
In the cities where these systems operate, where they know the environment and have been tested extensively, I would say they’re on par with human driving. That’s not to say they’re better, but they’re certainly not a lot worse. To me, the bigger challenge is societal. There’s a natural pace to technological adoption. Young people are probably much more willing to hop in and try it. People my age might say, “This will never work.” That hesitation isn’t really about autonomy. It’s about how humans react to new technology.
How do you respond when someone expresses distrust in autonomous vehicles?
My elevator pitch is simple: I was once in your boat. Five years ago, I would have said there’s no way I’d hit go in an urban environment. I knew all the hard problems that had to be solved. But over time, watching these systems work, I changed my mind. I tell people: Try it. Get in, see how it feels. It won’t drive the way you drive, but put aside the assumption that your way is the only safe way. There are many ways to be safe. Once most of my friends actually ride in a Waymo, they come back saying, “Oh, this is pretty cool.” And then they’re sold.
Does a city like Pittsburgh, with its infamous narrow roads and complex layout, offer particular challenges?
Pittsburgh is actually a great “next up” city because it’s modern. The roads are in good shape, but you have this ancient road network with five-way intersections and weirdness. There’s a downtown with a dense building population that cuts off GPS, then you move a mile and it’s fine.
It’s not San Francisco, but it has its own quirks. Even small cultural things — like people putting chairs in parking spots during the winter to save them –– are the kinds of unique edge cases that systems might not have encountered elsewhere. You and I understand them instantly, but a vehicle has to learn it. Those last 0.00001% edge cases are often the hardest to learn.
You recently published “Driven,” where you reflect on your early career and the groundbreaking autonomous drive across the U.S. What compelled you to document that era, and what does the book capture that might otherwise have been lost?
I’m thrilled with how the launch is going. My original goal was just to sell 100 copies, and we’re well on our way. But I didn’t write it to tell the world about Navlab or the early days of autonomous driving. I wrote it for my kids. I wanted them to understand what I did before they were born or before they were old enough to understand it. It just so happens that the rise of self-driving vehicles was a big part of that era, and I had access to the material, so that became the focus of this volume.
What I’m proud of is that there’s now a record of what we did before social media and TikTok. There’s a tendency to think if something isn’t online, it didn’t happen. A lot of foundational work happened before that documentation culture existed. I wanted to capture it in one place. Personally, writing has engaged my brain in new ways. At this stage of life, that’s incredibly rewarding.
You note in the book that you initially felt out of place at CMU. How did that shape the researcher you became at the Robotics Institute?
When I arrived at CMU from a small school in Indiana, I absolutely felt like an outsider. My reaction was that I had to prove I belonged. And if you want to do that, you can’t do the minimum. Everything I did had to be solid and better than average.
That outsider syndrome was a forcing function. I wanted to make my family proud. I wanted to show that the kid from a college in Indiana could compete here. I was also fortunate to choose the right research area and to have great mentors like Takeo Kanade and Chuck Thorpe. And CMU gave us tremendous freedom. The attitude was: we do things in the real world, so go do it. That trust enabled us to take risks and attempt things that might not have been possible elsewhere.
What do you think the next major breakthrough in self-driving technology will be?
I personally don’t think we’re going to see another massive leap in core capability. Instead, I think the breakthroughs will be domain-specific. You may see things like two-seater robotaxis, since 90% of U.S. trips are one or two people. You may see unmanned buses on fixed routes. Smaller, autonomous vehicles could use routes that aren’t economically viable with human drivers, increasing overall mobility. There could be improved access for people with disabilities. Even vehicle design could change if you don’t need steering wheels and pedals –– the things humans need to drive. Cars very well may become less expensive.
So, I see the next wave less as a technical leap and more as integration into new domains and business models. That is, assuming society continues to accept the technology.
If you were advising an RI grad student who is passionate about autonomous systems, what would you tell them to focus on?
I think the time when you could get a Ph.D. on a narrow slice of autonomy, like lane changes or intersection handling, has largely passed. Those problems are increasingly being subsumed by rapid advances in AI. If you’re a technical student, I’d suggest looking at higher-level system optimization. Once you have fleets of autonomous vehicles, you can think about fleet-wide planning, charging logistics and coordination –– problems humans can’t solve at scale. That’s more of a planning and optimization issue than a perception problem.
Jochem earned both his Master’s degree and Ph.D. at CMU
And if you’re in psychology, sociology or communications, there’s enormous opportunity. How do we get this technology accepted? How do we design it so people trust and understand it? Those challenges are huge and just as pressing.
Does CMU deserve its reputation as being years ahead in robotics?
Yes, because they’re not doing what everyone else is doing. One thing CMU has always been good at is the next thing. You don’t go to CMU to replicate existing work. You go there to build what’s coming next.
What sets CMU apart from other institutions is its deep-rooted culture of hands-on, real-world impact. The Robotics Institute was founded to do real-world robotics. There’s always been an appreciation for research that translates into practical impact. At CMU, you’re not just an academic. You’re expected to connect what you build to reality. I still think that’s a fundamental advantage.
What drives you now?
Legacy. I want to leave something for my kids so they can understand what I did and who I was during the years before them. I’ve been fortunate that each phase of my life has brought a new passion: sports, academia, entrepreneurship, learning to be a dad, and now writing. Writing stimulates a completely different part of my brain, and I enjoy that growth. At this stage, the process matters more than the outcome. If the book sells 30 copies, I’m happy. It doesn’t have to win a Pulitzer. The joy is in creating something meaningful and continuing to learn.
For More Information: Aaron Aupperlee | 412-268-9068 | aaupperlee@cmu.edu