Multi-Sensor LiDAR Solution Utilizes Deep Learning

By Mike Ball / 14 Dec 2021
Follow UST

LiDAR perception system for autonomous vehicles

Seoul Robotics has introduced Endeavor, the latest iteration of the company’s plug-and-play LiDAR perception system. The all-in-one deployment kit includes SENSR, the company’s proprietary perception software that uses deep learning to analyze and understand 3D data, as well as LiDAR sensors and a GPU (Graphical Processing Unit).

Endeavor can connect up to eight 32-beam rotating sensors at one time, capturing a surface area of up to 50,000 square meters. The system provides high-accuracy object detection, tracking, and classification for a variety of autonomous vehicle applications, and can mix and match different makes and models of LiDAR. This versatility allows the system to be configured to specific user requirements.

​By providing all key components in one solution, Endeavour shortens the time to deployment and makes the technology more accessible to industries that often face bandwidth and cost constraints. Key capabilities of the system include:

  • Centralized computing hardware with multiple network ports so the system does not require any additional hardware to attach multiple sensors, thus ensuring easier maintenance, less computing power, and simpler data presentation.
  • Power over Ethernet, making it cost-effective for both indoor and outdoor applications as there is no need for an external power supply to LiDAR sensors.
  • Reduction of raw sensor data bandwidth by over 95%, enabling the system to be deployed in both rural and urban environments regardless of broadband capabilities.
  • Enhanced customizations at request, including a deep learning model and either four- or eight-port configurations.

The backbone of the solution is SENSR, a 3D perception software package with the ability to leverage deep learning. SENSR can detect more than 500 objects at one time within a four-centimeter radius at distances of up to 200 meters, and predict up to three seconds in advance. SENSR is equipped with weather-filtering AI, enabling it to detect and track fast-moving, partially obstructed, or clustered objects even in heavy snow or rain.

Jerone Floor, VP of Products & Solutions at Seoul Robotics, commented: “Countless companies and public organizations can benefit tremendously from having access to 3D insights, but affordability and flexibility were historically barriers to entry. We’re disrupting this landscape by democratizing access with a cost-effective, all-in-one deployment system that is ready to install out of the box. With Endeavor, companies can tap this unparalleled awareness to create more efficient businesses and cities, improve safety, and fuel mobility.”

Posted by Mike Ball Mike Ball is our resident technical editor here at Unmanned Systems Technology. Combining his passion for teaching, advanced engineering and all things unmanned, Mike keeps a watchful eye over everything related to the unmanned technical sector. With over 10 years’ experience in the unmanned field and a degree in engineering, Mike’s been heading up our technical team here for the last 8 years. Connect & Contact