Hardware Podcasts

134. Can We Trust Self-Driving Cars? 🚗💥

The Next Byte Podcast · 8 Aug 2023 · 23m

What this episode covers

Hosts Daniel and Forbode discuss a proposal by MIT researchers to bridge the gap between autonomous vehicle technologists and regulators by introducing a hybrid safety model. They analyze how remote human supervisors can assist self-driving cars with complex tasks like highway merging, achieving 99.999% reliability when 30% of vehicles are cooperative. The episode highlights the scalability challenges of this approach, noting that while feasible for early adoption, it requires significant labor resources compared to fully autonomous systems.

Most relevant to

makers and hobbyists

Topics

Mentioned

Companies
MIT, Tesla, Starship Robotics, George Mason University, University of Michigan

Related episodes from other shows

Other shows on the same ground, closest first.

  • EE Times Current · 41m

    Voice-Activated Everything on the Cheap | Jets & Hoverboards & Reliability | AVs: How Safe is Safe?

    EE Times editors Brian Santo, Junko Yoshida, and George Leopold discuss the tension between reliability and time-to-market, using the Boeing 737 Max crashes and the hoverboard safety crisis as case studies. They analyze the shift in autonomous vehicle safety from driver-assist systems to full autonomy, featuring insights from Phil Koopman of Edgecase Research on the engineering challenges of prediction and required reliability levels. The episode also covers the emergence of edge-based voice processing technologies from companies like Pico Voice and Pietro Electric as alternatives to cloud-based solutions.

    Debugging & failure analysisAI & machine learning
    5 chapters

  • Moore's Lobby · 43m

    Ep. 33 | From Autonomous Golf Carts to Semis: The Journey to the Self-Driving Truck

    Brandon Moak, CTO and co-founder of Embark Trucks, explains how he and co-founder Alex Rodrigues scaled an autonomous driving system from a student golf cart project to a $5 billion public company. Moak details the technical stack, including the fusion of LIDAR, radar, and cameras, and the strategic decision to focus on hub-to-hub highway trucking to avoid the complexity of urban passenger vehicles. He discusses the operational challenges of scaling, including adapting to COVID-19 restrictions and mitigating chip shortages through risk-buying, while emphasizing Embark's safety-first approach and software-as-a-service business model.

    Robotics & autonomySensors & measurement in products
    6 chapters

  • Robot Talk · 21m

    Episode Twenty-Eight: Amanda Prorok

    Professor Amanda Prorok from the University of Cambridge discusses coordination challenges in multi-robot systems, focusing on multi-agent path finding and the transition from caged environments to mixed human-robot spaces like Amazon warehouses. She explains how collaborative driving algorithms, inspired by ant behavior, can increase traffic throughput by 30% in lab tests, though she argues that infrastructure augmentation is key to widespread autonomous adoption. The conversation also covers the use of drones and underwater robots for environmental monitoring, highlighting the limitations of GPS in dense forests and the potential of sensor networks on public transport.

    Robotics & autonomyEmbedded firmware
    6 chapters

  • The Robot Report Podcast · 1h 06m

    Kodiak Robotics & autonomous trucks, Anthony Levandowski going to jail

    Daniel Goff, head of policy at Kodiak Robotics, explains the company's 'middle mile' autonomous trucking strategy, focusing on highway operations between Dallas and Houston to simplify the technical challenge compared to robotaxis. Tyler Mantle, co-founder of the Ventilator Project, discusses the rapid development of the 'Era' low-cost ventilator, highlighting the use of 3D printing for rapid prototyping and partnerships with firms like Mass Robotics and Latham & Watkins to navigate FDA regulations and supply chain scaling. The hosts also cover the sentencing of former Google executive Anthony Levandowski for stealing Waymo trade secrets and a remote teleoperation demo of Boston Dynamics' Spot robot.

    Robotics & autonomyValidation & product-market fit
    8 chapters

  • EE Times Current · 40m

    AV & ADAS: Shall the Twain Ever Meet? | AI & Academia — a Fitful Fit | SiC ‘Em

    EE Times editors Brian Santo and Junko Yoshida discuss the divergent development paths of autonomous vehicles (AV) and advanced driver-assistance systems (ADAS), noting that cost structures and liability models prevent a linear transition from one to the other, as seen in Daimler's scaling back of Robotaxi plans. The episode then explores the rise of silicon carbide (SiC) in power electronics, highlighting partnerships between companies like Cree, STMicroelectronics, Soitec, and Applied Materials to overcome manufacturing challenges. Finally, George Leopold and Loring Morbel analyze the 'fitful fit' between AI and academia, arguing that the 'black box' nature of deep learning makes it difficult for universities to meet the quantifiable deliverables required by government and industry funders.

    Mechanical engineeringMaterials & manufacturing processes
    6 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by The Next Byte Podcast. The episode itself belongs to its publisher.