All Segments
Chris Urmson: Self-Driving Cars at Aurora, Google, CMU, and DARPA | Lex Fridman Podcast #28
29 segments available
Segments Timeline
Meet Chris Urmson: A Pioneer in Autonomous Vehicles
In this segment, Lex Fridman introduces Chris Urmson, the CTO of Google's self-driving car team and CEO of Aurora Innovation. Urmson's background includes significant contributions to the DARPA Grand Challenges and leadership in autonomous vehicle technology. This introduction sets the stage for a deep dive into the challenges and innovations in self-driving technology.
"the following is a conversation with Chris Urmson he was a CTO of the Google self-driving car team a key engineer and leader behind the Carnegie Mellon University autonomous vehicle entries in the DAR..."
Lessons from DARPA Challenges
Chris Urmson reflects on his experiences with the DARPA Grand Challenge and Urban Challenge, emphasizing the belief that autonomous driving could be achieved. He discusses the initial skepticism and the excitement of overcoming seemingly impossible challenges, highlighting the importance of naivete in tackling hard problems.
"you were part of both the DARPA Grand Challenge and the DARPA urban challenge teams at CMU with red Whittaker what technical or philosophical things have you learned from these races I think the the h..."
The Pain Points of Autonomous Driving
Urmson outlines the major challenges faced in autonomous vehicle development, including mechanical issues, sensor integration, and software algorithms for mapping and localization. He emphasizes the complexity of the field and the necessity of mastering various components to succeed in autonomous driving.
"possible did you from the very beginning did you personally because you're one of the lead engineer you actually had to do a lot of the work yeah I was the technical director there and did al huddle t..."
The Evolution of Technology in Autonomous Vehicles
In this segment, Chris discusses the technological advancements from the DARPA challenges to today. He highlights the significance of HD mapping and multi-beam lidar as pivotal innovations that transformed the capabilities of autonomous vehicles, allowing for better navigation and understanding of the environment.
"sensors hardware software algorithms for mapping localization just general perception control what the hardware soft first of all I think that's the joy of this field is that it's all hard and that's ..."
Navigating Uncertainty in Autonomous Driving
Chris shares insights on the unpredictability of real-world driving scenarios compared to the controlled environments of the DARPA challenges. He discusses the challenges of dealing with unknown routes and the importance of adaptability in autonomous vehicle systems.
"uncertainty of exactly what they wanted us to do was was part of the challenge as well right you didn't actually know the track heading in you know approximately but you know it didn't actually know t..."
Leadership Lessons from Red Whittaker
Reflecting on his mentor Red Whittaker, Urmson shares key leadership lessons, including the importance of empowering others and recognizing potential in individuals. He emphasizes the value of taking on difficult challenges and fostering a culture of trust and responsibility.
"kind of leading that effort in terms just broadly speaking so you're a leader now what have you learned from reading about leadership I think there's a couple things one is you know go and try those r..."
From Challenges to Real-World Applications
Chris discusses the transition from the DARPA challenges to real-world applications of autonomous vehicles. He highlights the advancements in technology and the need for vehicles to operate in unpredictable urban environments, addressing the complexities introduced by various road users.
"fast-forward through the history can you maybe talk through the technical evolution of autonomous vehicle systems from the first to Grand Challenges to the urban challenge to today are there major shi..."
The Role of Perception in Autonomous Driving
In this segment, Urmson explains the evolution of perception technologies in autonomous vehicles. He discusses the importance of localization and the advancements in sensor technologies that have enabled vehicles to better understand their surroundings and predict the behavior of other road users.
"of the that was the big step there for the urban challenge you know one of the big technological innovations there was the multi beam lidar and being able to generate a high resolution you know mid to..."
Mapping and Localization Challenges
Chris delves into the intricacies of mapping and localization in autonomous driving. He explains the challenges of matching high-resolution maps to real-world coordinates and the importance of accurate localization for safe navigation in complex environments.
"naively trusting GPS I and s before that and and again like lots of work had been going on in this field certainly this was not doing anything particularly innovative in slam over in localization but ..."
The Unpredictability of Real-World Driving
Urmson discusses the unpredictability of real-world driving compared to the controlled environments of the DARPA challenges. He emphasizes the need for autonomous vehicles to adapt to a wide variety of road users and scenarios, highlighting the complexities of urban driving.
"much work was the perception doing there so how how much were you relying on localization based on maps without using perception to register to the maps and how I guess the question is how advanced wa..."
The Future of Autonomous Driving
In the concluding segment, Chris reflects on the future of autonomous driving, discussing the ongoing challenges and the potential for innovation in the field. He emphasizes the importance of real-world testing and the need for continuous improvement in technology to achieve safe and reliable autonomous vehicles.
"cities to the urban environment yeah I think the the biggest thing is that the you know the the actors are truly unpredictable that most of the time you know the drivers on the road the other road use..."
The Role of Lidar in Self-Driving Technology
Chris Urmson discusses the critical role of lidar in the sensor suite for autonomous vehicles. He argues that while lidar is essential, cameras and radars are also necessary for a robust system. He addresses Elon Musk's controversial view that lidar is a crutch, emphasizing the importance of using the best available technology to enhance safety.
"and you know it's just a it's a different game so how important he said leiter came into the game early on and it's really the primary driver of autonomous vehicles today as a sensor so how important ..."
Technology and Road Safety
Urmson highlights the urgent need for technology that can reduce road fatalities, citing the alarming statistic of 37,000 deaths in the U.S. annually. He advocates for the use of any technology that can accelerate the development of self-driving cars to improve safety, questioning the arbitrary rejection of certain technologies like lidar.
"them and and so there's an existence proof that you can drive using you know passive fission no doubt can't argue with that in terms of sensors yeah so yes maybe sensors right so like there's there's ..."
Cost vs. Safety in Sensor Technology
In this segment, Urmson discusses the balance between cost and safety in developing sensor suites for autonomous vehicles. He argues that while reducing costs is important, the effectiveness of the sensors must not be compromised, as safety is paramount in the deployment of self-driving technology.
"be using and it feels just arbitrary to say well you know I'm I'm not okay with using lasers because that's whatever but I am okay with using an 8 megapixel camera or a 16 megapixel camera you know li..."
The Future of Lidar and Economic Viability
Urmson explores the future of lidar technology, suggesting that while it may remain more expensive than cameras, there is potential for cost reduction. He emphasizes the need for economically viable sensor suites that can support the scalability of autonomous vehicles while ensuring safety.
"matter of time and I think really maybe the I would talk to the question you asked about you know the cheapest set certainly I don't think that's actually what you want what you want is a sensor suite..."
Human Factors in Autonomous Driving
Chris Urmson addresses the human factors involved in level two autonomous systems, highlighting the risks of over-trust in technology. He discusses the importance of clear communication about the capabilities of these systems to prevent misuse and ensure safety.
"costs down to zero it's the basic economics you've talked about your intuition that level to autonomy is problematic because of the human factor of vigilance that command complacency over trust and so..."
The Divergence of Technology Paths
In this segment, Urmson discusses the divergence between level two driver assistance systems and the technology needed for fully autonomous vehicles. He warns that the economic pressures on level two systems may hinder the development of truly self-driving cars, emphasizing the need for responsible marketing and communication.
"capability set that that enables and the trust they should have in it and that is where I you know I kind of I am actually incremental II more you know concerned around level three systems and you kno..."
Trust and Understanding in Autonomous Vehicles
Urmson concludes by reflecting on the public's understanding of autonomous vehicle technology. He argues that even with proper communication, the inherent trust people place in technology may lead to dangerous situations, underscoring the need for ongoing education and awareness about the limitations of current systems.
"doing so quickly enough to where it doesn't cost safety critical safety concerns you believe that the it needs to diverge at this point in two different basically different routes and really that come..."
Trust and Experience with Autonomous Vehicles
In this segment, Urmson explores the psychological aspects of trust in Level 2 autonomous vehicles. He explains how personal experience can lead to overconfidence in the technology, despite a lack of comprehensive risk understanding. Urmson warns that even with a perfect understanding of the system, users may drift into misplaced trust based on their daily experiences, which could lead to safety concerns.
"communication and this probably applies though for vehicles as well the marketing and a communication of what the technology is actually capable of how hard it is how easy it is all that kind of stuff..."
Demonstrating Safety in Autonomous Vehicles
Chris Urmson outlines the challenges of demonstrating the safety of autonomous vehicles. He discusses the need for a combination of thorough engineering processes and evidence collection to build public confidence. Urmson emphasizes the importance of working with regulatory bodies to ensure that safety metrics are credible and that the public understands the capabilities of autonomous systems.
"that as this technology moves from what I say it's kind of the more technology savvy ownership group to you know the mass market you may be able to have some of those folks who are really familiar wit..."
Metrics for Assessing Autonomous Vehicle Safety
In this segment, Urmson addresses the inadequacies of current metrics used to assess autonomous vehicle safety, such as disengagement rates. He argues for the development of a more comprehensive set of metrics that reflect human performance in various driving tasks. Urmson believes that these metrics can help demonstrate that autonomous systems can outperform human drivers in safety.
"place that means you work with people that are essentially experts at safety to try to discuss and show do you think the answer is probably no but just in case do you think there exists a metric so cu..."
Winning Public Trust for Autonomous Technology
Chris Urmson discusses the importance of public perception in the adoption of autonomous vehicles. He acknowledges the skepticism surrounding new technology but emphasizes the potential benefits, such as improved safety and mobility. Urmson shares personal experiences demonstrating autonomous vehicles to skeptics, illustrating how firsthand experience can shift perceptions and foster acceptance.
"most popular topic about autonomous vehicles in the public is the trolley problem formulation right which has let's not get into that too much but is misguided but in many ways but it speaks to the fa..."
The Future of Autonomous Vehicle Deployment
In this forward-looking segment, Urmson speculates on the timeline for large-scale deployment of autonomous vehicles. He identifies the moment when driverless vehicles operate continuously on public roads without safety drivers as a significant milestone. Urmson discusses the transition from engineering projects to commercialization and the importance of creating a user-friendly experience for widespread adoption.
"so putting this technology in the hands of people speaks to scale the deployment all right so what do you think the the dreaded question about the future because nobody can predict the future yeah but..."
Urban vs. Highway Deployment
Urmson shares insights on where autonomous vehicles are likely to first be deployed, suggesting urban and suburban environments will be the initial focus. He contrasts the complexities of urban driving with the perceived simplicity of highway driving, emphasizing the need for safety and learning opportunities.
"real in the world what do you think that deployment looks like where do we first see the inkling of no safety driver one or two cars here and there is it on the highway is it in specific roads in the ..."
Learning from Urban Driving Challenges
Chris Urmson discusses the advantages of deploying autonomous vehicles in urban settings, where lower speeds and frequent interactions allow for quicker learning and adaptation. He highlights the importance of safety in these environments to build public trust.
"that intuition is pretty good except we don't really care about most of the time we we care about all of the time and when you're driving on a freeway with a truck say 70 70 miles an hour and you've g..."
Accelerating Autonomous Vehicle Development
Urmson reflects on the technological breakthroughs needed to accelerate the development of autonomous vehicles. He emphasizes the importance of advanced perception and forecasting capabilities to enhance safety and efficiency in self-driving systems.
"urban environment so ten years and then scale twenty thirty year I mean who knows if if it's sufficiently compelling experience is created it can be faster and slower do you think there could be break..."
Addressing Vulnerable Road Users
Chris Urmson expresses his concerns about the safety of vulnerable road users, such as pedestrians and cyclists, in the context of autonomous vehicles. He discusses the challenges of ensuring safety while navigating complex interactions with these users.
"pretty dramatically how you in terms of staying up at night are you mostly bothered by cars pedestrians or cyclists so I I worry most about the vulnerable road users about the combination of cyclists ..."
The Game of Pedestrian Interactions
Urmson delves into the complexities of how autonomous vehicles will interact with pedestrians, particularly in crowded situations. He discusses the need for algorithms that can effectively manage these interactions while ensuring safety.
"them and I've heard this and I don't really believe it because if I'm driving down the road and somebody steps in front of me I'm going to stop right like a even if I'm annoyed I'm not going to just d..."
Competing in the Autonomous Vehicle Space
Chris Urmson shares insights on how Aurora approaches competition in the autonomous vehicle industry. He emphasizes the importance of a strong leadership team, a supportive culture, and significant investments in infrastructure to accelerate their progress.
"competitor to Arora today and how do you out-compete them in the long run so we really focus a lot of what we're doing here I think that you know I've said this a few times that this is a huge difficu..."