Self-Healing Maps for Autonomous Driving Sanjay Sood GPU - - PowerPoint PPT Presentation

▶

Aug 19, 2023 375 likes •568 views

Self-Healing Maps for Autonomous Driving Sanjay Sood GPU Conference| May 10, 2017 HERE T Technologies 7,000+ 7, Countries mapped Millions Mi ons of 200 20 changes made to the Employees in 56 countries map eve very day focused on

SLIDE 1

Self-Healing Maps for Autonomous Driving

Sanjay Sood GPU Conference| May 10, 2017

SLIDE 2

HERE T Technologies

20 200

Countries mapped

4of

In-car navigation systems in Europe and N America use HERE maps

30 30+

Years of experience transforming mapping technology

Today, HERE provides services to nearly all global OEM brands as well as many market leading brands such Microsoft, Amazon, Facebook etc.

7, 7,000+

Employees in 56 countries focused on delivering the world’s best map and location services

Mi Millions

ns of

changes made to the map eve

very day

SLIDE 3

We are making sense of the world through the lens of location, helping people achieve better outcomes through data-driven location solutions HERE T Technologies

SLIDE 4

Maps for cars Maps for cars by cars

SLIDE 5

SLIDE 6

HERE’s HD Live Map solves several major problems

Precise p positi tioning

for lateral & longitudinal control Planning of vehicle control maneuvers

beyond s sensor v visibility ty Enhanced s sensor f functi tionality ty

for contextual awareness

f the environment

Local knowledge of the

rules of the road

SLIDE 7

The industry has evolved, providing new possibilities for cars

Connected Intelligent

SLIDE 8

Building the high definition map

Proprietary cameras LiDAR Novatel GPS/INS GPS Antenna

96.4 MP every 6 meters

41.3º

700,000 Points Per Second

SLIDE 9

Harness crowd- sourced sensor data For near real- time updates

High definition map

A high definition, continuously updated map requires the crowd

SLIDE 10

Mach chine learning will be cr critica cal for self-driving ca cars

Map

SLIDE 11

The self-healing map process

Map update Pu Publish Aggregation Localization Change detection Observations Features Internal HD map SE SENSO SORIS HD HD Live Map Points

(signs, signals, pavement markings…)

Polylines

(lane markings, road edges…)

Many to one SD SDRI (C (Campaign ma manageme ment) t) OEM Cloud Drive paths HD map tiles OEM sensor observations

1

Sensor collection & ingestion 2 Aggregation in the cloud 3 Creation of a new feature 4 Update map & publish

SLIDE 12

The self-healing map process

Map update Pu Publish Aggregation Localization Change detection Observations Features Internal HD map SE SENSO SORIS HD HD Live Map Points

(signs, signals, pavement markings…)

Polylines

(lane markings, road edges…)

Many to one SD SDRI (C (Campaign ma manageme ment) t) OEM Cloud Drive paths HD map tiles OEM sensor observations

SLIDE 13

The self-healing map process

Map update Pu Publish Aggregation Localization Change detection Observations Features Internal HD map SE SENSO SORIS HD HD Live Map Points

(signs, signals, pavement markings…)

Polylines

(lane markings, road edges…)

Many to one SD SDRI (C (Campaign ma manageme ment) t) OEM Cloud Drive paths HD map tiles OEM sensor observations

SLIDE 14

The self-healing map process

Map update Pu Publish Aggregation Localization Change detection Observations Features Internal HD map SE SENSO SORIS HD HD Live Map Points

(signs, signals, pavement markings…)

Polylines

(lane markings, road edges…)

Many to one SD SDRI (C (Campaign ma manageme ment) t) OEM Cloud Drive paths HD map tiles OEM sensor observations

SLIDE 15

The self-healing map process

Map update Pu Publish Aggregation Localization Change detection Observations Features Internal HD map SE SENSO SORIS HD HD Live Map Points

(signs, signals, pavement markings…)

Polylines

(lane markings, road edges…)

Many to one SD SDRI (C (Campaign ma manageme ment) t) OEM Cloud Drive paths HD map tiles OEM sensor observations

NDS Protobuf

SLIDE 16

Localization Change detection

Cloud and edge analytics: What to do where, best?

Map update Pu Publish Aggregation Localization Change detection Observations Features Internal HD map SE SENSO SORIS HD HD Live Map Points

(signs, signals, pavement markings…)

Polylines

(lane markings, road edges…)

Many to one SD SDRI (C (Campaign ma manageme ment) t) OEM Cloud Campaign mngt. & fleet requests Drive paths HD map tiles OEM sensor observations

NDS Protobuf

SLIDE 17

Summary Vehicles today are connected and becoming more intelligent

1

Intelligent cars need a high definition map

2

An HD map needs the ability to heal itself via the crowd

3

Self-healing maps require large amounts of vehicle sensor data

4

SLIDE 18