This vision-based (camera-based) self-position estimation technology (Visual SLAM technology) enables highly accurate and robust self-position recognition for conveyance automation in factories and warehouses. Operation doesn’t require magnetic tape, helping to improve maintainability while also reducing equipment startup lead times.
An embedded computer (DX-U2200 Series) equipped with an NVIDIA Jetson Orin NX module is used as an autonomous mobile robot (AMR) controller. GPUs and AI accelerators allow for high-accuracy, high-speed recognition of 3D environments for self-position estimation and autonomous operation.
igital Media Professionals Inc.’s ZIA MOVE software, developed for autonomous driving, allows for the development of state-of-the-art AMRs. Using vision-based (camera-based) information input enables three-dimensional space recognition, resulting in greater self-position estimation accuracy than 2D-LiDAR. Meanwhile, using AI to detect people that would otherwise cause self-positioning loss, and excluding them from the feature values, improves the robustness of self-position estimation.
AMRs are used not only for inter-process conveyance in the auto and manufacturing industries but also for automated transfers in the logistics industry. To meet the need for non-stop, extended operation, a DX-U2200 Series fanless computer is used as the computing unit. Equipped with NVIDIA’s high-performance GPU and AI accelerator, the DX-U2200 Series is ideal for running ZIA MOVE, developed by Digital Media Professionals Inc. Moreover, this computer is designed to be not only high-performance but also compact and operable with a DC power supply—perfect for integration into an AMR, where battery operability and a space-saving design are preferred.
To PageTop