LAVER LABS

NAVER LABS has been internally developing mapping devices for digital twin data production. To name a few, we’ve created the on-road MMS (mobile mapping system), R series, and indoor mapping robots, M series. The device being introduced now, T2-B, belongs to the T series. “T” here stands for “transformable.” As the name suggests, mapping devices with various shapes belong in the T series. Among them, T2-B is a backpack-type device for spaces that are difficult to scan with wheel-based mapping robots. Like the sub-initial “B” in this version, many different types of mapping devices such as T2-H and T2-D may come out in the future. Of course, their shape and functions are still confidential.

Creating and improving our own device and using it in our solutions has a huge advantage. Compared to using third-party solutions or equipment, we can maintain high data quality in various environments and purposes because algorithm modifications or equipment upgrades are quick and easy. In fact, we find technical perfection in the process of direct communication between engineers in each area, while identifying areas of improvement and better optimized methods.

History

In the past, T2-B was introduced to the public under the project name COMET. One of the goals of the COMET team at the time was vertical data connection. In other words, it started with the idea that if there was a device that can smoothly match both stair & indoor data in a building, we could quickly create digital twin data of the indoor space through the SLAM algorithm. Those ideas have now been broadened, expanding the data collection area to outdoor sidewalks. Following the demonstration of the first version, the current T2-B has greatly improved accuracy, scalability, and operational convenience.

T2-B is actively being used in the Incheon Ieum Project and collaborations with the National Museum of Korea, as well as in NAVER’s second headquarters which is designed as the world's first robot-friendly building. The characteristics of T2-B are as follows.

Flexible Modular Configuration

• SLAM module: due to its backpack-type feature, it is difficult to use specific sensors to measure the distance moved. To overcome this, we enabled visual inertial odometry algorithm using an IMU and camera, and used RTK GPS for absolute coordinate data.
• VL module: collects vision-based localization data.
• Main Control module: manages the collected data. It also manages the current status of sensors and devices, and interconnects with external systems.
• 5G module: enables data interconnection and remote monitoring between operator and assistant.

Excellent Scalability

• Separate UIL (universal interlink) design method: one of NAVER LABS' internally developed design methods that enable physical separation of each module. In scenarios where the operator has to carry the device him/herself for mapping purposes, this allows easy and reliable disconnection, movement, and connection of devices.
• Common module: each HW / FW / ME / SW area is designed to be flexible to structural change. For example, the core modules of T2-B are quite identical to the mapping robot M2. The data collecting function also works with the same code. Since the upgraded common module can easily be applied, it provides greater flexibility to system expansion and change.

Sensor Features and Improvements

• It eliminates errors in adjusting the illuminance and increases the size of the image sensor to acquire image data with improved quality.
• The GPS reception rate has been improved by resolving the mutual interference between frequencies within the equipment. By minimizing EMI through sensor modularization and main controller, the operating efficiency of each sensor has also been improved.

Wearable, the Idea of Human Usage

• Operator convenience has been significantly improved. The previous version, COMET, had implemented the basic purpose of mapping devices. However, it had work to do in terms of “consideration for humans” who actually carry the device around to map different areas. In response, T2-B has enhanced both efficiency and convenience, with the addition of voice notifications on the device status, color-based device status display, battery monitoring, low-center of gravity design, device lightweighting, and an interface that condenses operation and status information into a single button.

ARC System Interconnection (Powered by NAVER CLOUD PLATFORM)

• ARC is a multi-robot intelligence system developed by NAVER LABS, and consists of ARC brain which controls multi-robots and interconnects with the building infrastructure, and ARC eye which provides digital twin data and localization algorithms. T2-B is interconnected with ARC eye, which automates the workflow of data collection, data uploading & processing, and the creation of serviceable data. Securing such a stable system maximizes the efficiency of creating digital twin data in wide areas.