& Co Cogn gnition System em Proprietary & Confidential | - - PowerPoint PPT Presentation

co cogn gnition system em proprietary confidential 1
SMART_READER_LITE
LIVE PREVIEW

& Co Cogn gnition System em Proprietary & Confidential | - - PowerPoint PPT Presentation

Int nteg egrated Sen ensin ing & Co Cogn gnition System em Proprietary & Confidential | 1 Proprietary & Confidential | 1 The last generation of drivers Proprietary & Confidential | 2 Proprietary &


slide-1
SLIDE 1

Proprietary & Confidential | 1 Proprietary & Confidential | 1

&

Int nteg egrated Sen ensin ing Co Cogn gnition System em

slide-2
SLIDE 2

Proprietary & Confidential | 2 Proprietary & Confidential | 2

The last generation

  • f drivers
slide-3
SLIDE 3

Proprietary & Confidential | 3 Proprietary & Confidential | 3

Company Overview

⁺ Addressing the heart of the autonomous vehicle challenge; Understanding the changing environment with the most advanced 3D sensing and cognition algorithms ⁺ Our offering:

⁺ AI sensor fusion platform ⁺ “Point and Shoot” LiDAR

⁺ Founded in Feb 2016 ⁺ Raised seed in July 2017

AI Deep learning

Computer

vision 3D modeling

System architecture

slide-4
SLIDE 4

Proprietary & Confidential | 4 Proprietary & Confidential | 4

Autonomous Vehicles

Driving Control Driving Decisions Cognition/Fusion Sensors

slide-5
SLIDE 5

Proprietary & Confidential | 5 Proprietary & Confidential | 5

Driving Control Driving Decisions Cognition/Fusion Sensors

Autonomous Vehicles

Int ntegrat grated ed Se Sens nsing ing & Cog

  • gni

nitio tion n Sy System em

slide-6
SLIDE 6

Proprietary & Confidential | 6 Proprietary & Confidential | 6

⁺ Sensors: Camera, Lidar Radar ⁺ Basic fusion

⁺ High accuracy ⁺ Low latency ⁺ High reliability ⁺ Affordable

Main challenge for L3 and higher autonomous cars

Solutions need to be: But current solutions are lacking:

slide-7
SLIDE 7

Proprietary & Confidential | 7 Proprietary & Confidential | 7

Legacy Fusion Solutions

Driving Decisions & Control Basic Fusion/Cognition

Camera cognition SW LiDAR cognition SW RADAR cognition SW

Camera LiDAR RADAR Environment Model

slide-8
SLIDE 8

Proprietary & Confidential | 8

slide-9
SLIDE 9

Proprietary & Confidential | 9 Proprietary & Confidential | 9

Raw Data Fusion Combined Algorithms for a rich environment model Point & Shoot LiDAR and RADAR Driving Decisions & Control Raw Data Fusion/Cognition

Camera LiDAR

Point & Shoot

RADAR

Point & Shoot

Rich Environment Model

slide-10
SLIDE 10

Proprietary & Confidential | 10

Data Fusion Raw

slide-11
SLIDE 11

Proprietary & Confidential | 11

Raw Data Fusion

⁺ Integrating sensor data on a pixel level ⁺ Overcome sensor faults ⁺ Optimize detection ⁺ Simplify decision making

Legacy Fusion Solutions VAYAVISION Raw Data Fusion

slide-12
SLIDE 12

Proprietary & Confidential | 12

& Shoot LiDAR Point

slide-13
SLIDE 13

Proprietary & Confidential | 13

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

& Shoot LiDAR Point

+

+ +

+

+ + +

+ +

+ +

+ + + + + + + + + + + + +

+ +

+ + + + +++

+++ ++

+ + + + +++

+++

+ + ++ ++ ++ + ++ ++ + ++ ++ + +++ ++ + + + ++ ++ + + ++ ++ ++ ++ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + ++ ++ ++ + ++ ++ + ++ ++ + +++ ++ + ++ ++ ++ ++ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

slide-14
SLIDE 14

Proprietary & Confidential | 14

Point and Shoot LiDAR

Requirements ⁺ 150m Range @ 10% reflectivity ⁺ 120° x 25° Field of View ⁺ <0.1° Horizontal resolution ⁺ 0.1° Vertical resolution ⁺ <6cm accuracy ⁺ 60 Frames per second ⁺ 5000 points per frame

slide-15
SLIDE 15

Proprietary & Confidential | 15

for a Rich Environment Model

Combined Algorithms

slide-16
SLIDE 16

Proprietary & Confidential | 16

Rich Environment Model

U1

Free space Lane Car Pedestrian Unidentified Obstacle Traffic light Sign Bicycle SLAM

slide-17
SLIDE 17

Proprietary & Confidential | 17 Proprietary & Confidential | 17

Combined algorithmic approach

⁺ SLAM ⁺ High Resolution 3D Model ⁺ Free space detection

slide-18
SLIDE 18

Proprietary & Confidential | 18 Proprietary & Confidential | 18

slide-19
SLIDE 19

Proprietary & Confidential | 19 Proprietary & Confidential | 19

slide-20
SLIDE 20

Proprietary & Confidential | 20 Proprietary & Confidential | 20

slide-21
SLIDE 21

Proprietary & Confidential | 21 Proprietary & Confidential | 21

Overwhelming amount of data

slide-22
SLIDE 22

Proprietary & Confidential | 22 Proprietary & Confidential | 22

Stronger processor

slide-23
SLIDE 23

Proprietary & Confidential | 23 Proprietary & Confidential | 23

What we need

⁺Strong CPUs and GPUs ⁺Development Environment ⁺Scalability ⁺Market acceptance

slide-24
SLIDE 24

Proprietary & Confidential | 24 Proprietary & Confidential | 24

What we need

⁺Real Time System

⁺CPUs ⁺Interfaces ⁺Peripherals

⁺Algorithms

⁺Strong GPU acceleration ⁺Development cycle

slide-25
SLIDE 25

Proprietary & Confidential | 25

Best platform

slide-26
SLIDE 26

Proprietary & Confidential | 26

Summary

Change pics

slide-27
SLIDE 27

Proprietary & Confidential | 27

VAYAVISION’s AI Sensor Fusion Platform

⁺ Raw data fusion:

⁺ Better cognition algorithms ⁺ Can operate with VAYAVISION LiDAR and 3rd party sensors (Cameras, LiDARs and RADARs)

⁺ Point and Shoot LiDAR

⁺ Better resolution and higher frame rate ⁺ Affordable

⁺ Richer environment model

Fusion/ Cognition ADAS/AD

LiDAR Point & Shoot Camera RADAR Point & Shoot

GPS V2X

Map Cellular Vehicle Control

HMI

DRIVE PX 2

slide-28
SLIDE 28

Proprietary & Confidential | 28 Proprietary & Confidential | 28