Understanding how to build and implement gesture control in robotics is an essential step in developing smart applications.
In this two-part webinar series, deep learning models are trained, deployed, and integrated into robot operating systems (ROS2) using state-of-the-art, cost-optimized programmable logic and system-on-chip (SoC) solutions. how to do it. Each 60-minute webinar uses fun robotics applications to demonstrate the development steps necessary to create basic designs and how to enhance these designs with hardware accelerators.
Each technical presentation includes Q&A for the audience.
Part I: Training and deploying a classification model
Tuesday, May 9, 2023 at 2:00 PM US EDT
In this session, we will discuss how to train a machine learning model for gesture recognition, using American Sign Language classification as an example. Through a series of easy-to-follow Jupyter notebooks, participants will learn to analyze and understand datasets, train a classification model, and deploy on his ZUB1CG programmable logic board using the Vitis-AI framework.
Part II: Integrate the model into ROS2
Tuesday, May 23, 2023 at 2:00 PM US EDT
This session introduces ROS2, shows how to integrate the model trained in Part I into a ROS2 graph, and explores how hand gestures control a simple robot simulation.
speaker:
Mario Bergeron, Machine Learning Specialist, Avnet
Mario Bergeron is a machine learning specialist at Avnet, specializing in embedded vision and AI at the edge. He started his career as an ASIC designer with a Bachelor’s degree in Computer Engineering from Laval University in his City of Quebec. Mario has over 30 years of experience in DSP and FPGA based embedded design. He currently focuses on Deep He Learning, He Platform, Reference He Design, and Customer Training.
Bryan Fletcher, Product Line Manager, AMD
Bryan Fletcher has been interested and fascinated by programmable logic devices since working on Lattice CPLDs and Hewlett Packard’s high speed test cards in the 1990s. He is currently a product line manager at AMD, responsible for his SoC cost-optimized products including the Zynq-7000 and Zynq UltraScale+ families. Bryan said he worked at Avnet and Memec for 22 years, first as his FAE dedicated to Xilinx, then in technical marketing, working on the Spartan 2 low-cost evaluation board, Spartan-3A DSP starter kit, ZedBoard, and The focus was on development boards such as the Ultra96-V2. .
Moderator:
Amanda Hosey, Editor, SAE Media Group
By registering for this webinar, you agree to receive product announcements and special offers from SAE International via email. By submitting your personal information, you acknowledge that SAE Media Group, SAE International Company and our carefully selected partners sponsors of this content may contact you and that you have read and agree to our Privacy Policy. You agree (CA Residents see Section D of the Privacy Policy). policy). You can opt out of communications at any time.you can contact us This email address is being protected from spambots. JavaScript must be enabled to view..
