Hardware Podcasts

208: What If You Had a Machine Do It

Embedded.fm · 27 Jul 2017 · 49m · Episode 208

What this episode covers

Elecia White discusses her project to build a typing robot named 'Ty' using an Nvidia Jetson TX2 and a MeArm kit, explaining the intersection of machine learning and robotics. She details the challenges of inverse kinematics, the Law of Cosines, and the decision to abandon Robot Operating System (ROS) due to its complexity and dependency issues. White also covers the use of OpenCV for laser tracking, the computational limits of retraining neural networks, and her plans for future reinforcement learning integration.

Most relevant to

makers and hobbyists

Topics

Mentioned

Companies
Hackaday, SupplyFrame DesignLab, LA CrashSpace
Products
MeArm
Chips
Nvidia Jetson TX2, Adafruit PCA9685
Tools
Robot Operating System, OpenCV, TensorFlow, Weka
Also
Andrew Ng, Patrick Pilarski

Related episodes from other shows

Other shows on the same ground, closest first.

  • The Robot Report Podcast · 1h 18m

    Ted Larson from OLogic shares robot development stories

    Ted Larson, co-founder and CEO of OLogic, discusses the complexities of developing commercial robotics, emphasizing the need to focus on solving specific market problems rather than just building robots. He shares insights on the transition from ROS to ROS 2, the importance of design for manufacturing, and the emerging role of Edge AI in applications like mobile manipulation and connected fitness. Larson also highlights the potential of large language models to democratize robot programming through natural language interfaces.

    Embedded firmwareAI & machine learning
    8 chapters

  • The Amp Hour · 1h 18m

    Robo Fry Guy

    Dave Jones and Chris Gammell discuss the feasibility of humanoid robots in low-end service jobs, citing high costs and the 'jack of all trades, master of none' design philosophy, while Jones details Tesla's dense supercomputer power supply failures caused by MEMS oscillator resonance. The hosts also explore practical hardware engineering, including Gammell's use of DALL-E for AI image generation, his 'Lego prototyping' with off-the-shelf sensor boards, and Jones's audit of household phantom power using specialized meters. Finally, they debate the challenges of sourcing standard plastic enclosures and the logistics of remote hardware training via browser-based containers.

    AI & machine learningRobotics & autonomy
    8 chapters

  • Being an Engineer · 36m

    Jacob Evink | From CAD to Mechanical Curiosity: Student Engineer Spotlight

    Jacob Evink, a mechanical engineering student at Western Michigan University, discusses his journey from high school manufacturing classes to working in a high-volume injection molding factory. He details the design and construction of a four-axis robotic arm using an ESP32 microcontroller and stepper motors, highlighting the trial-and-error process of electronics integration. Evink also shares his perspective on using AI tools like ChatGPT for coding assistance and the importance of documenting failures publicly on Instagram to build an engineering portfolio.

    Prototyping & iterationMicrocontrollers & dev platforms
    9 chapters

  • Robot Talk · 24m

    Episode 109: Building robots at home - Dan Nicholson

    Dan Nicholson, founder of MakerForge.tech, discusses his journey from a software engineer to building a large-scale open-source bipedal companion robot. He explains how he transitioned from a simple cube design to a complex, 3D-printable robot using hobby servos, Raspberry Pi, and cloud-based CAD tools to make robotics accessible to non-experts. Nicholson highlights the challenges of balancing cost and functionality in hobby robotics, the importance of rapid prototyping with 3D printing, and the value of open-source collaboration in accelerating development and community engagement.

    Robotics & autonomyPrototyping & iteration
    9 chapters

  • Built 2 Scale · 1h 09m

    The Robotics Rundown: AI Agents, Secondary Markets & the Future of Construction

    Hosts Matt Perrott (Build Pass) and Scott Wilcox (W Build) discuss the accelerating robotics landscape, highlighting Figure AI's rise on secondary markets and Nvidia's Omniverse platform for robot training. They analyze the impact of AI agents like Manus AI and Sesame AI on software development and daily life, while also sharing personal strategies for balancing high-performance entrepreneurship with fatherhood. The episode concludes with Wilcox outlining his vision for Elysium, an autonomous luxury home-building venture inspired by the Golden Gate Bridge's construction history.

    Robotics & autonomyAI & machine learning
    8 chapters

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