Yandex Introduced In-House LiDARs For Their Self-Driving Vehicles

Yandex, a technology company creating intelligent products and services based on machine learning has just entered the business of making LiDAR (Light Detection and Ranging) systems for self-driving vehicles. Yandex is currently testing their new LiDAR systems and has planned to use these sensors on their delivery robots too.

 

Yandex LiDAR systems under test consists of a 360 degree LiDAR on the top and two solid-state LiDARs below

 

Yandex is working on two types of LiDARs, solid-state LiDAR with 120-degree FOV (Field Of View) for use on the front of the vehicle and a 360-degree LiDAR that can generate the model of the vehicle’s entire environment. The products have been improved based on the extensive three years and over 1.5 million miles of testing. The new LiDAR sensors offer better flexibility that enables us to tune the scanning patterns on the fly. The custom parameters provided with the new versatile software allows the LiDAR to adapt to different driving conditions, such as freeways, dense city streets and diverse weathers. In real world, the Yandex LiDAR sensors can focus on an object while adjusts its scanning pattern to further determine if it is a pedestrian, a bicycle or something else from 200 meters away.

Surrounding image created by Yandex LiDAR sensors

Compared to other LiDAR sensors, Yandex engineers receive more information about the vehicle’s surroundings from their own LiDAR sensors from which they can access the raw data. Because of this advantage, Yandex self-driving cars can create a better 3D view of the surroundings via their LiDARs that are synchronized with the data collected by the cameras and RADARs on board. Yandex LiDAR sensors are much cheaper than others. The current prototypes are already half the prices of existing devices. It’s expected to cut the cost of sensors up to 75% when Yandex LiDARs transition to mass production.

Read more at: https://sdc.yandex.com/#id2

Share What You’ve Learned

Leave a Reply