The main focus of autonomous vehicle (AV) technology development has been focused on vehicles and their systems. However, one company is taking a unique approach that uses the infrastructure and environment around a vehicle to achieve Level 5 autonomy.

At CES 2022, Seoul Robotics unveiled its Level 5 Control Tower (LV5 CTRL TWR), a mesh network of computers in the infrastructure that guides vehicles autonomously without requiring that sensors be placed on individual vehicles. The company also revealed that it is in the early stage of commercial deployment of its LV5 CTRL TWR technology with BMW to automate last-mile fleet logistics at automaker’s manufacturing facility in Munich.

With an eye on several commercial fleet applications, including vehicle distribution centers to car rental companies and trucking logistics, Seoul Robotics said it developed its new AV solution to overcome the cost, safety, and flexibility challenges presented by traditional approaches to vehicle autonomy. By placing sensors equipped with 3D perception software around vehicles, such as traffic lights, buildings, and highway overhangs, the system can fully capture an environment and communicate with other sensors and the 4/5G systems that come standard on vehicles today. It collects 3D data and then automates vehicles accordingly using V2X communications.

“Level 5 mobility has been proven to be more challenging to achieve than expected—until now,” said HanBin Lee, CEO of Seoul Robotics. “LV5 CTRL TWR has massive potential to fuel autonomous mobility. Ultimately, these systems will be deployed in additional public and commercial settings, powering aspects of our everyday lives, such as autonomously navigated parking and public transit.”

The company’s collaboration with BMW, which it said is the largest-ever implementation of this type of AV technology, leverages hundreds of connected LiDAR and 3D sensors on infrastructure to automate newly manufactured vehicles within factories and vehicle distribution centers, without any human involvement. By making this process autonomous, Seoul Robotics believes OEMs like BMW can increase operational efficiencies and safety within automotive logistics.

According to Seoul Robotics, LV5 CTRL TWR is able to automate vehicles from multiple vantage points such as behind a truck and around corners. It can also predict trajectories to eliminate blind spots, which is a current challenge for on-vehicle LiDAR systems. The company says this awareness of a vehicle’s environment and surrounding activity can reduce collisions and create a more reliable process. The system can also handle the movement of hundreds of vehicles simultaneously without any added cost, ensuring vehicles can drive slower or take longer, safer routes to avoid accidents.

As with any AV technology, software is at the heart of the new Seoul Robotics solution. LV5 CTRL TWR is powered by SENSR, Seoul Robotics’ 3D perception software platform that is both sensor-agnostic and able to leverage deep learning AI for enhanced tracking, detection, and prediction capabilities with centimeter accuracies. The system is designed to mix and match different makes and models of 3D sensors, providing it with the most accurate environmental understanding, according to the company.

The story of Seoul Robotics is an interesting one. The company was founded in 2017 in Seoul, Korea, by a group of engineers with backgrounds in 3D data processing. The group met in an online class as they studied computer vision together. The students teamed up for a competition to develop a software system for self-driving cars, relying on data generated only by LiDAR sensors rather than a combination of camera sensors, which was unheard of at the time. The team placed 10th out of more than 2000 teams. After the team’s success, the Seoul Robotics company was launched.

Seoul Robotics calls itself a 3D perception company with a specialty in LiDAR, specifically focused on software. The company’s website also states a bold ambition: Seoul Robotics aims to use SENSR to provide the base operating system for LiDAR around the world, for all applications, as Microsoft Windows is to desktop computers.

In addition to SENSR, Seoul Robotics develops additional applications for LiDAR, including autonomous-driving software, autonomy through infrastructure, and object tracking and recognition solutions.