BUILDING AUTONOMOUS VEHICLES USING DRIVE PX 2 Shri Sundaram, May 8, - - PowerPoint PPT Presentation

▶
building autonomous vehicles using drive px 2
SMART_READER_LITE
LIVE PREVIEW

BUILDING AUTONOMOUS VEHICLES USING DRIVE PX 2 Shri Sundaram, May 8, - - PowerPoint PPT Presentation

Oct 14, 2022 268 likes •475 views

BUILDING AUTONOMOUS VEHICLES USING DRIVE PX 2 Shri Sundaram, May 8, 2017 Autonomous Vehicles Artificial Intelligence NVIDIA AT THE CENTER OF AV REVOLUTION 2 DRIVE PX 2 DRIVING AV AI Launched CES 2016 Spike in AV AI engagements after we

slide-1
SLIDE 1

Shri Sundaram, May 8, 2017

BUILDING AUTONOMOUS VEHICLES USING DRIVE PX 2

slide-2
SLIDE 2

2

Autonomous Vehicles Artificial Intelligence

NVIDIA AT THE CENTER OF AV REVOLUTION

slide-3
SLIDE 3

3

Launched CES 2016 Spike in AV AI engagements after we powered on the discrete GPU late last year

DRIVE PX 2 DRIVING AV AI

slide-4
SLIDE 4

4

DRIVE PLATFORM TODAY FOR LEVEL 3/4

DRIVE PX (Xavier)

slide-5
SLIDE 5

5

BUILDING BLOCKS

Autonomous Driving SDK Compute – APIs and Libraries Compute Engines & IOs

  • DriveWorks
  • CUDA
  • CuDNN
  • Tensor RT
  • NVMedia
  • GPU
  • Accelerators
  • Sensors
slide-6
SLIDE 6

6

BUILDING BLOCKS

Autonomous Driving SDK Compute – APIs and Libraries Compute Engines & IOs

  • DriveWorks
  • CUDA
  • CuDNN
  • Tensor RT
  • NVMedia
  • GPU
  • Accelerators
  • Sensors
slide-7
SLIDE 7

7

COMPUTE & IO

System

IOs: 70+ Gbps Compute: 20+ DL TOPS SW: Linux, RFS, BSP, OTA

Sensors/Actuators

Camera, Lidar, Radar DBW, CAN interface + More incl. IMU, GPS, V2X modules

slide-8
SLIDE 8

8

BUILDING BLOCKS

Autonomous Driving SDK Compute – APIs and Libraries Compute Engines

  • DriveWorks
  • CUDA
  • CuDNN
  • Tensor RT
  • NVMedia
  • GPU
  • Vision Processors
  • Accelerators
slide-9
SLIDE 9

9

CUDA

DRIVE PX 2 has latest CUDA-capable GPU – 4x! Runs latest CUDA 8. Includes CUDA accelerated DL SDK

TensorRT CuDNN CuBLAS.

slide-10
SLIDE 10

10

TENSOR RT

GPU Inference Engine

slide-11
SLIDE 11

11

NVMEDIA

NVMedia: APIs to control cameras as well data NVMedia abstracts:

  • Image capture
  • Image Signal processor
  • 2D Engine
  • Encoding
  • Interop with other SW components
slide-12
SLIDE 12

12

BUILDING BLOCKS

Autonomous Driving SDK Compute – APIs and Libraries Compute Engines

  • DriveWorks
  • CUDA
  • CuDNN
  • Tensor RT
  • NVMedia
  • GPU
  • Vision Processors
  • Accelerators
slide-13
SLIDE 13

13

DRIVEWORKS

Input Perceive Localize Visualize Vehicle Control APIs

Sensors IMU GPS Image Processing DNN interface CV Modules HD Maps interface CAN/DBW Kit interface

Tools

  • Calibration and sensor registration

Vehicle rig configuration Data Logging/Acquisition

Samples/ Previews

Sensor visualization DriveNet LaneNet Several CV examples Occupancy Grid *PilotNet

slide-14
SLIDE 14

14

DRIVE PLATFORM

DriveWorks CUDA, CuDNN, TensorRT NVMedia DRIVE PX 2 — System Software DRIVE PX 2 — System Hardware Vehicle Control DNNs and Autonomous Driving Applications Sensors HD Maps

slide-15
SLIDE 15

15

PX2: ON WHEELS

HW

DRIVE PX 2 nicely mounted in the trunk of a car, pre-wired for cameras, and other sensors Sensors - Your choice or NVIDIA’s configuration.

SW

All of what comes with DRIVE PX2

Photo Courtesy: AutonomouStuff

slide-16
SLIDE 16

16

DRIVE PLATFORM

HW

ZF Pro AI – built on DRIVE PX 2 Auto Cruise Bosch – building on DRIVE PX using Xavier Xavier: 30 DL TOPS + ASIL-C for Level 4

SW

QNX - Common arch across Linux and QNX Common APIs across CUDA, TensorRT, cuDNN, NvMedia + Support for POSIX APIs

For production

slide-17
SLIDE 17

17

OTHER RESOURCES

Tools

Application Development tools Developer Zone (developer.nvidia.com/DRIVE) Developer Forums (devtalk.nvidia.com)

Schools

Deep Learning Institute – Autonomous Driving Udacity Nanodegree

slide-18
SLIDE 18

18

DRIVE Platform

DRIVE Platform – Open AI Autonomous Vehicle platform One architecture, from L2 to L3 to L4 to AI Car DRIVE PX 2 for L3, L4 development today

slide-19
SLIDE 19