NVIDIA press release (2024.11.13) NVIDIA Laputa Robotics uses NVIDIA accelerated computing to speed up its flexible automated warehouse “Laputa ASRS” The “automated warehouse” market is expanding against the backdrop of ever-growing e-commerce
https://prcdn.freetls.fastly.net/release_image/12662/486/12662-486-7c204181db3d2e4bbc89168fce7af794-567×268.jpg
warehousing, storage, and shipping operations, but you can also utilize height and aisle width, making it possible to make the most of your warehouse space. Global Information predicts that the market size for automated warehousing systems (ASRS) will reach USD 8.7 billion in 2023, with a compound annual growth rate of 7.9%, reaching USD 12.7 billion in 2028. Laputa Robotics’ flexible automated warehouse “Laputa ASRS” Laputa Robotics is one of the startups that is attracting attention in the field of logistics robots. Unique cloud robotics platform “ rapyuta.io ” development, implementation, and operational support for robot solutions. The company’s strength is its “group control” technology, which efficiently controls the movements of large numbers of robots. Laputa Robotics has been releasing the flexible automated warehouse “Laputa ASRS” since August 2023. This is an automated warehouse that can be constructed without anchors, and has the flexibility to be designed to suit the customer’s needs. It uses a resin frame jointly developed with Mitsui Chemicals, and has a block structure that requires no anchor work and allows for free layout. After installation, you can expand or reduce the scale or relocate according to business growth with minimal cost. It is easy to adapt to unusual warehouse shapes and can be installed across fire doors. There are various types of automated warehouses, but Laputa Robotics’ “Laputa ASRS” uses warehouse space more efficiently than other companies’ automated warehouses, and is also superior in terms of power consumption and speed. It can be doubled.” So far, it has been operating at the existing warehouse of HobbyLink Japan, which handles models and plastic models. In addition, the warehouse of Nippon, a major publishing agency, is scheduled to start operating soon.
https://prcdn.freetls.fastly.net/release_image/12662/486/12662-486-71dcd31be97d94249aaaceae25f66b9d-567×319.jpg Video: Flexible automated warehouse Laputa ASRS concept video ( https://www.youtube.com/watch?v=2ZhZwHQizNQ ) A small, thin robot moves in an automated warehouse and carries cargo “Laputa ASRS” consists of four parts. It has a multi-floor structure made of resin frames for storing cargo, a special container called a “bin” with four legs, and a vehicle that crawls under the “bin” and lifts up to carry the cargo as it travels around the warehouse. This is a small, thin robot, and an elevator that allows robots to move up and down between floors. The small robot with a thickness of 80mm was also independently developed. The wheels are special wheels called “mecanum wheels” that can move in any direction without turning, and no turning time is required. A white line is drawn on the floor, and the robot moves using this as a guide.
https://prcdn.freetls.fastly.net/release_image/12662/486/12662-486-a0f4b9738b379c350f03bdcd8173aab2-1430×807.png A robot running inside an automated warehouse NVIDIA Jetson Orin Nano System-on-Module enables AI to process camera data to enable autonomous robot navigation This robot is equipped with cameras on the front, back, left and right sides. The floor has recesses in it so that the legs of the bottle fit snugly together. Normally, your feet fit into these indentations to prevent them from shifting, but on rare occasions they may slip out of place. Therefore, the robot uses a camera to recognize the orientation and posture of the bottle, as well as surrounding obstacles, and navigates to avoid collisions. For edge AI processing of camera images, NVIDIA Jetson Orin Nano (
https://www.nvidia.com/ja-jp/autonomous-machines/embedded-systems/jetson-orin/ ) module is used. Prior to development on NVIDIA Jetson, robots were unable to perform camera image processing due to a lack of onboard AI processing power. The Jetson Orin Nano module allows real-time processing of both front and side camera footage. Laputa ASRS currently implements video processing for one camera, but further functional expansion is planned in the future.
https://prcdn.freetls.fastly.net/release_image/12662/486/12662-486-ce592509a9af6000b905a86167583121-1430×805.png robot camera view Run inference faster with NVIDIA TensorRT The recognition AI was independently developed by Laputa Robotics and uses NVIDIA TensorRT ( https://developer.nvidia.com/tensorrt ) is used to run it. When executing a trained model of a deep neural network, the model is automatically converted to the optimal form according to the type of GPU that will actually be executed. Users can significantly reduce inference time without having to think about the details. In addition, for learning recognition AI, NVIDIA RTX 5000 Ada generation GPU (
https://www.nvidia.com/ja-jp/design-visualization/rtx-5000/ ) is used. Picking stations support worker productivity with AI Loading and unloading operations for the Laputa ASRS are carried out at fixed points called stations. Robots carry the bottles, so workers hardly have to walk. Up to 7 storage/shipping bins can be independently controlled around the operator. By having multiple robots on standby, the waiting time for workers to pick can be reduced to zero. Guidance for picking and loading is provided using a large screen monitor and a projector. The light is projected from the projector, so all you have to do is follow that light. Even untrained workers can start working right away. NVIDIA’s accelerated computing is also used to ensure worker safety. Use this to check whether robot transport within automated warehouses is being carried out safely and correctly.
https://prcdn.freetls.fastly.net/release_image/12662/486/12662-486-6988269cf5e2b545437484049cb2530d-1430×1039.jpg picking station Calculation performance requires margin Laputa Robotics has installed NVIDIA Jetson in its “Laputa ASRS” robot, and has installed powerful NVIDIA Isaac (
https://developer.nvidia.com/ja-jp/isaac/sim ) The company decided to adopt NVIDIA accelerated computing, such as its robotics platform, because of the need for more AI processing at the edge. Customers always demand “higher performance.” Automated warehouses are a developing technology and can be made even faster. However, in order to continue to improve the performance of robots and the entire automated warehouse system, we need some margin in computing performance. Even if additional development is required, if there is not enough computational performance, it will be necessary to start from scratch. With the future in mind, Laputa Robotics chose to utilize NVIDIA Jetson. Aiming for comprehensive efficiency with digital twin So what will happen in the future? In promoting automation of logistics warehouses, what is really important is “total optimization” rather than partial optimization. To do this, it is necessary to handle the movements of various material handling equipment and workers all at once. Laputa Robotics is developing the automatic warehouse “Laputa ASRS”, the autonomous mobile robot “PA-AMR” that works with humans, and the automatic forklift “Laputa AFL”. This is an automatic forklift truck that can move and transport flexibly without having to attach markers or the like to walls or floors. It is already being used at four warehouses in Japan. If we are to truly improve efficiency, it is important not only to automate parts of the system, but also to handle all of them in one system and make them work together. It optimizes and streamlines the entire flow, including unloading from the warehouse’s truck berth, unloading, transferring to temporary storage, and entering and exiting the warehouse. Omniverse, a platform that includes application programming interfaces, software development kits, and services for developers to easily integrate Universal Scene Description (OpenUSD) jp/omniverse/” target=”_blank” rel=”nofollow ugc noopener”>
https://www.nvidia.com/ja-jp/omniverse/ ) and NVIDIA Isaac Sim, the reference robot simulation platform for building and testing AI-enabled robots, provides an effective way to simulate robot systems before building them and improve them once they are in production. Makes it easier to continue the cycle. It also allows all parties involved to reach a common understanding and take on the challenge of making improvements. You can also use NVIDIA Omniverse to create photorealistic, physically accurate virtual environments for presentations during customer visits. Omniverse allows you to create highly detailed virtual representations of your automated warehouse and work processes, helping your customers form a clear and consistent picture of what their automated warehouse will look like. . In the future, Laputa Robotics will continue to utilize NVIDIA’s accelerated computing in its solutions to further improve logistics efficiency. About NVIDIA Since its founding in 1993, NVIDIA (
https://www.nvidia.com/ja-jp/ ) (NASDAQ: NVDA) is a pioneer in accelerated computing. Invented by the company in 1999, the GPU is fueling the growth of the PC gaming market, redefining computer graphics, and igniting the modern AI era while helping to digitize industries. NVIDIA is a full-stack computing company with data center-scale products that are currently reshaping the industry. For more information, follow this link: https://nvidianews.nvidia.com/