The Role of Lidar in Autonomous Vehicles: A "Crutch" or Essential for Robustness and Safety?

The discussion centers on the critical role of Lidar in autonomous vehicles, directly addressing Elon Musk's controversial assertion that Lidar is merely a "crutch." Chris Urmson, an expert in the field, firmly believes that Lidar is a central component of a robust sensor suite, alongside cameras and radar. He argues that a composition of data from multiple sensor types is essential for achieving reliable and safe autonomous driving, emphasizing that the goal is to accelerate the deployment of self-driving technology to save lives, rather than arbitrarily rejecting effective tools. Urmson acknowledges the "existence proof" that humans can drive using passive vision (cameras), validating the premise that Lidar isn't strictly *necessary* in all contexts. However, he distinguishes between what is *possible* and what is *optimal* for safety and scalability. He frames Lidar as a technology that, while potentially superseded in the distant future, is currently vital. He draws an analogy to the combustion engine being a "crutch" on the path to electric vehicles, suggesting that current "crutches" are necessary steps for progress. The core nuance is that the objective is not the cheapest sensor suite, but an "economically viable" one that *works* reliably. The practical insight is that the selection of sensor technology should be driven by efficacy and the ability to solve the problem of road safety, not by arbitrary technological preferences or a singular focus on minimal cost. Urmson recommends utilizing "the best technology from the tool bin" to accelerate the deployment of self-driving cars. He also highlights that while Lidar may be more expensive than camera-based systems due to manufacturing processes, its costs can be substantially driven down, and a robust business model can absorb higher Bill of Materials if the technology provides superior value and functionality, leading to a sustainable and scalable solution. The broader implications touch upon the ethical imperative of reducing road fatalities, citing the tragic statistic of 37,000 American deaths annually. The debate over Lidar versus camera-only approaches is not just technical but also economic and societal, impacting the speed at which life-saving autonomous technology can be brought to market. The discussion underscores the tension between technological purity (e.g., vision-only) and pragmatic engineering solutions that prioritize safety and rapid deployment, ultimately shaping the future of transportation and public safety.

leiter came into the game early on and it's really the primary driver of autonomous vehicles today as a sensor

I think it's I think it's a central you know I believe it but I also believe is the cameras are essential and I believe the radars is essential

you really need to use the composition of data from from these different sensors if you want the thing to to really be robust

lidar is a crutch that is the kind of I guess growing pains and that's much of the perception tasks can be done with cameras

there's an existence proof that you can drive using you know passive vision no doubt can't argue with that

the combustion engine was a crutch on the path to an electric vehicle the same way that you know any technology ultimately gets replaced by some superior technology in the future

any technology that we can bring to bear that accelerates the this techno you know self-driving technology coming to market and saving lives is technology we should be using

what you want is a sensor suite that is economically viable and then after that everything is about margin and driving cost out of the system