Announcing the Autonomous Vehicle Summit Breakouts
Photo by Erin BeachWe're so excited to see everyone on June 14 in San Francisco at our Autonomous Vehicles Summit, focused on the latest developments in the field. An important part of each Summit is the breakout session: your opportunity to hear from the innovators in the industry and to ask them all your burning questions.
If you're registered for the Summit, you'll receive an email from [email protected] requesting your top choices among the breakout sessions. Not a subscriber but interested in attending this event? Subscribe here and get access to the event.
There will be three sets of breakouts, with four sessions happening at once. You’ll be able to attend two or three breakout topics during the event (space is limited). You’ll receive your breakout assignments the morning of the Summit. Here are the breakouts:
Greasing the (18) Wheels: What stands in the way of automated truck driving on highways
Alex Rodrigues, CEO, Embark
Paul Konasewich, Business Development Director, Paccar
Shoaib Makani, CEO, KeepTruckin
The path to getting self-driving trucks on highways may be shorter than for self-driving cars in dense cities, but high-speed driving requires high-speed tech, some of which isn’t available yet. Alex Rodrigues, who runs a self-driving truck software startup; Paul Konasewich, whose company Paccar supplies trucks to developers including Waymo, Embark, and Nvidia; and Shoaib Makani, whose startup KeepTruckin helps trucking companies manage and track their fleets, will explain how the field will develop. They will highlight the technical and regulatory challenges facing traditional truck makers such as Mercedes and Silicon Valley players like Waymo and Uber, and walk through the changing economics of trucking and logistics.
Ready to Roll: How to measure technical progress and validate what's "safe" for public use
Alisyn Malek, COO, May Mobility
Axel Gern, VP Autonomous Driving, Mercedes/Daimler
Self-driving vehicle developers often don’t define what level of safety is necessary to take drivers out of cars. Even defining “human-level” performance is fraught with uncertainty. How are the various developers approaching these questions? Can the industry or regulators help create realistic standards for developers to follow? Will the weakest players who cause the most fatalities taint the rest of the industry? Two practitioners—one from Mercedes, the other a former longtime GM manager now running an autonomous vehicle startup—discuss how various companies approach these issues.
Smart Money: The autonomous vehicle value chain and the multitude of business models that can win
Anne Widera, consultant, formerly at Waymo & Uber ATG
Christopher Heiser, CEO, Renovo
Current autonomous car prototypes are driving the first mile of a marathon. In this session, a strategist who has worked at Waymo and Uber and a startup CEO working with multiple developers will discuss the timeline and business models that will shape how such vehicles proliferate. They’ll also consider the economic impact on cities, retail, real estate and other adjacent industries. Which companies will create lasting businesses in this industry and which ones won’t?
Where's Everybody Going? predicting the movement of people, vehicles and other objects around a self-driving car
Alexandre Haag, CTO, Audi Autonomous Intelligent Driving
Forrest Iandola, CEO, DeepScale
It’s hard to anticipate the trajectory of lots of moving objects at the same time. But doing so may be necessary to make automated driving work in dense cities. Prediction practitioners are debating things like how much to rely on machine learning and whether it makes sense to determine a single “most probable path” for each object and react to that. The technical chief of Volkswagen/Audi’s automated driving subsidiary and the CEO of DeepScale, a firm that processes raw data from a vehicle’s sensors, explore the latest work in this field and how prediction evolves.
Look Ma’, No Hands: Triumphs and failures of today’s “advanced driver-assistance” features.
Kay Stepper, VP/Head of Driver Assistance, Bosch
Mark Rosekind, Chief Safety Innovation Officer, Zoox
Semi-autonomous driving tech already is saving lives, but existing and forthcoming hands-free features are unproven—and possibly dangerous. A key executive of one of the biggest suppliers of car components will evaluate results from today’s providers (Bosch, GM, Mobileye, Tesla and others). He’ll discuss what’s coming next and how such features will—or won’t—contribute to fully automated driving. He’ll be joined by the former head of the National Highway Traffic Safety Administration, Mark Rosekind, who has studied human fatigue issues for decades. Dr. Rosekind will talk about why he has safety concerns around some of the new features and how major developers can allay them.
Supplier & Demand: Which hardware & software components will be built in-house and which will be outsourced?
Boris Shulkin, VP, R&D, Magna International
Sasha Ostojic, former Cruise engineering executive
Mike Volpi, General Partner, Index Ventures
Companies like Cruise and Waymo are trying to develop as much software and hardware in-house as possible, with the belief that it’s the only way to create a reliable system. Companies with business models that depend on automakers being able to mix and match different systems hope to keep the supply chain status quo. Startups appealing to the latter camp offer everything from simulators and sensors to maps and “annotation” to help train computer vision models. In this session, Boris Shulkin, who evaluates new technologies at one of the biggest auto components suppliers, Sasha Ostojic, a former head of engineering at Cruise, and Mike Volpi, who has invested in auto-related startups (and also sits on the boards of Aurora Innovation and Fiat Chrysler), discuss the opportunities and dead ends for startups trying to ride the autonomous vehicle craze.
The Sims: What simulation systems can—and cannot—teach autonomous vehicles
Sertac Karaman, President, Optimus Ride
Justyna Zander, Automotive Manager, Nvidia
Simulators are a critical tool for running autonomous vehicle software through billions of potential scenarios. But they may be difficult to develop, and developers don’t always see eye to eye on how much to rely on them to train the software. This session will explore a number of related topics. These include the different ways that developers use simulators; whether there's room for standalone simulation vendors that work with multiple developers; how to simulate human behavior; the limitations of using videogame software as the basis for autonomous vehicle simulation; and how regulators may create simulators of their own to test companies’ software. Justyna Zander previously founded a company focused on simulation and has extensive experience working with car manufacturers. She worked at Intel before joining Nvidia, which offers simulation software to developers. Sertac Karaman, an aeronautics professor at the Massachusetts Institute of Technology, has experience with simulators there as well as at his autonomous vehicle startup.
Hardware Feelings: Which on-board computers & sensors will win the day
Helen Pan, Director of Hardware, Baidu Intelligent Driving Group
Jennifer Haroon, VP, Strategy & Business Ops, Nauto
Helen Pan spent more than three years working on hardware at Waymo/Google, after developing wireless technologies at Intel Labs for more than 10 years. And she sits on the board of lidar leader Velodyne. That experience gives her a prime vantage point for evaluating the vast array of car sensor and computing packages being sold to autonomous vehicle developers. She can also speak about the combinations of GPU, CPU, and AI accelerator microchips that will proliferate in cars of the future, and the kind of limitations they may have in trying to power the autonomous vehicle software.
Ms. Pan’s opinions matter because, as a Baidu executive, she may play a big role determining which suppliers will succeed in China, the largest auto market in the world. She will be joined by the former head of Waymo’s business operations, Jennifer Haroon, who is now works at Nauto, which develops software to detect if drivers are distracted.
Deep Thoughts: Robotics versus DNNs
Oliver Cameron, CEO, Voyage
Sertac Karaman, President, Optimus Ride
In the ongoing argument over whether to rely more on robotics or deep neural networks to make a self-driving car, passions run high. Two developers of robotaxis examine the various technical approaches being pursued by longtime autonomous vehicle developers and newer entrants. Which developers have an edge and which don’t? The speakers will explain where they fall on the spectrum of this industry-wide debate, how/why many different approaches could succeed, and when we’ll know who’s right.
Living on the Edge Case: How autonomous vehicles can deal with an extremely messy world
Alex Wang, CEO, Scale API
Adrien Gaidon, Machine Learning Lead, Toyota Research Institute
Mike Volpi, General partner, Index Ventures
No matter how many miles they travel virtually in a simulator, autonomous cars in the real world will encounter a nearly infinite array of edge cases, or unusual circumstances they haven't previously experienced. That means an autonomous driving system will be filled with uncertainty about never-before-seen objects detected by its sensors; whether it can classify them as a pedestrian, animal, inanimate object or something else; and how to predict if or where they might move. How will the system account for such uncertainties, aside from constantly braking as a precaution? And how can it tell the difference between a “false positive,” or object that don't pose a danger, and one that does? In this session, Alex Wang, CEO of a startup that helps autonomous vehicle developers like GM/Cruise, Embark, and Voyage to train their computer vision algorithms; Adrien Gaidon from Toyota; and Mike Volpi an investor who is a board director at Aurora Innovation and Fiat Chrysler, discuss how these systems can navigate the “long tail” of scenarios.