Introduction

Accurate maps are an essential part of a reliable autonomousdriving system, millions of kilometers needs to be efficiently mapped. Currentprocess is slow and expensive. So we see an opportunity to develop asystem that can Automatically extract road network map information from(Image, LIDAR, GPS, IMU, CAN, …) data, which will be used for autonomousdriving applications. We are using state of the art technologies like (Deeplearning, probabilistic inference, reinforcement learning). Our target is toreduce the time and cost for creating a precise road network map from (40hours/ kilometer, 1000$) to (10 minutes / kilometer, 100$)

Accurate design map

 Difference between ground truth and created map at the centimeter level.

Secure

 Providing encrypted maps with secure cloud services.

Smart Mapping Tools

 Use state of the art AI technology to accelerate map creation and update processes.

Updated Maps

 Continuous updates of the existing maps with new available data.

Rich Details

 Multilayer map support for different types of Autonomous applications.

Evaluated Results

 Automatically testing and comparing maps using simulation software and real vehicles if necessary.