AIIM Human Detection

AIIM obstacle detection

AIIM delivers an advanced scene understanding solution that enhances the autonomy of the forklift AMR (Autonomous Mobile Robot) throughout the operation in dynamic, unpredictable indoor warehouse environments where the AMR collaborates with human operators and avoids collision with obstacles in the intended path of the AMR.

Traditional automated systems make use of legacy systems including sensing the environment with (multiple) 2D-LiDARs. These systems typically have limited capabilities that result in AMR OEMs having to install individual, dedicated systems to complete separate tasks. Due to the limited amount of data points that these systems capture from the environment, it is extremely challenging for these systems to provide a reliable understanding of the scene. AIIM enables the AMR to understand the scene and solve the problem. AIIMs standardized product capabilities can easily be combined into a holistic solution to scene understanding where separate tasks do no longer require individual, dedicated hardware.

During operation, the AIIM human detection software finds and detects human operators in the warehouse. The AIIM solution determines which operator is of most interest and provides the relative pose of that operator to the AMR system. AIIM enables the AMR to communicate to the human operator that he or she is detected by the AMR, and to which angle relative to the AMR. Communicating with human operators in the warehouse improves collaboration between humans and AMRs by reducing the perceived risk of human operators in the warehouse.

The AIIM obstacle detection software runs in parallel with the AIIM human detection software. The obstacle detection software enables the AMR to detect obstacles in the intended path of the AMR. Obstacles are detected beyond the capability of the mandatory 2D-safety LiDAR. The AIIM obstacle detection software is not limited to a single scan plane but covers a configurable 3D space in front of the AMR and, because of that, the risk of a collision with obstacles that are outside the scan plane of the mandatory 2D-safety LiDAR is reduced.

The standardized product capabilities of AIIM on human detection and obstacle detection are deployed quickly and effectively to best fit the customer requirements. The solution will be delivered and deployed in a hardware reference design and seamlessly integrated into the customer mobile robotics system. The hardware reference design is based on a NVIDIA Jetson processing unit and an Intel Realsense D455 sensor.

CTO F3 Design: "With the AI-enabled advanced scene understanding solution of AI in Motion, we can further enhance the autonomy of our latest generation of Nippers. The AIIM solution will make our Nipper AGV more collaborative, communicating with human operators in the warehouse. On top of that, obstacles are successfully detected along the intended path of the Nipper AGV. The AIIM solution enhances the autonomy of our Nippers while at the same time it allows us to simplify the sensor suite."