Algorithm
Amicro's in-house perception and decision-making engine covers SLAM, real-time path planning, multi-object detection and tracking. Deeply optimized for our own chips, these algorithms deliver accurate, low-latency, low-power edge inference, enabling reliable autonomous navigation and intelligent interaction in complex environments.
Simultaneous localization and mapping using LiDAR–vision fusion enables high-precision real-time mapping and centimeter-level localization in large, complex environments, providing reliable environmental perception for robots.
Multilevel path planning combines global planning with local obstacle avoidance and supports real-time replanning in dynamic environments so robots reach their destinations safely and efficiently.
Lightweight deep-learning object detection enables real-time multiclass recognition and tracking on the NPU, including pedestrians, obstacles and furniture.
Precise mapping and path planning for comprehensive cleaning coverage, with intelligent obstacle avoidance around furniture.
SLAM mapping for large spaces, with multi-floor navigation and real-time avoidance of moving obstacles.
Localization and navigation in complex indoor environments, with multi-object recognition and grasping guidance for efficient delivery.
Powering Your
Next Robotics Product
Interested in our solutions?
We would be happy to help!
