Tesla AI Day 2021: FSD Autopilot Progressing Under the Hood

2021/10/12

Summary

About This Report

  This report was written by VSI Labs (VSI) for MarkLines' portal users. VSI is a technology research company that provides industry with deep insight and analysis on the enabling technologies used for active safety and automated driving.

 

Introduction

  Tesla hosted its first ever AI Day on August 19, giving updates on its progress with AI technologies for the Autopilot system, and unveiling an AI training supercomputer - Dojo and an upcoming humanoid Tesla Bot teaser. Overall, Tesla unveiled a lot of proprietary technologies with impressive details on what they have developed, where they are at now, and where they are going.

  This Tech Brief will focus on Tesla’s new perception neural network (NN) architectures and SW stack rewrite (both perception and path planning), and data annotation. It will also round up the current deployment status of a variety of Tesla Autopilot systems, as Tesla is going through unprecedentedly complex development efforts in software and hardware. By covering both the AI Day content and public deployment status, the Brief will provide insights on how Tesla is uniquely executing the progress of semi-autonomous systems.

 

Material from: Danny Kim and Matthew Linder, Tesla AI Day 2021: FSD Autopilot Progressing Under the Hood, Technology Brief [August 26, 2021] [the websites of VSI Labs AUTONOMOUS SOLUTIONS PORTAL] reproduced with permission of VSI Labs.

 

VSI Labs


VSI Labs reports:
The Automated Drive South - Highlights & Observations (Jul. 2021)
Cameras Taking a Step into Making the Unseen Visible (Apr. 2021)
CES 2021 Round-up: Industry Leaders Still Stood Out at Digital CES (Jan. 2021)
The Role of Big Tech in AV Development (Dec. 2020)
Is Tesla an Auto Company or a Tech Company? (Nov. 2020)
Tesla Bets on Urban ADAS: Traffic Light & Stop Sign Control (Aug. 2020)

 

This report is for paid members only. Remaining 3 chapters remaining.
Free membership registration allows you to read the rest of the article for a limited time.