TM
PERCEPTION ENGINE
Breakthrough Visual Perception System
RGo's intelligent vision and AI system, Perception Engine, provides mobile robots with artificial perception capabilities, enabling them to understand complex surroundings and operate autonomously just like humans. Perception Engine is a full-stack software platform that runs on multiple hardware reference designs.
FEATURES
Layers of 3D Perception Data
Spatial and perception data are collected, analyzed and presented as layered maps, allowing easy and intuitive utilization of information over the API. Perception Engine provides all the layers of information a robot needs to be able to move in complex, real life environments.
3D Localization
Perception Engine provides the precise position and orientation of the robot (x, y, z, yaw, pitch, and roll) while generating and updating a user-friendly map. Pose information is accompanied with confidence level and dynamic driving feedback to the navigation stack.
Intelligent Live 3D Mapping
As the robot travels throughout the space, Perception Engine creates and updates maps continuously. The live, 2D and sparse 3D maps allows for user-friendly, intuitive maps that require very little compute. Changes in obstacles or map boundaries can be intelligently assessed. Customer-defined logic will update the permanent map or alert the user or a supervisory system. Maps are designed to require low memory footprint enabling large area coverage. Maps can be shared globally if a network is available.
Global Map Alignment
With Perception Engine's Global Map Alignment tool, the system is easily programmed to output localization and object position data in the same coordinate frame and units as the facility map - or with any transform that is desired.
Indoor & Outdoor Operation
Perception Engine was designed to be robust in a wide range of conditions including outdoor environments. Direct sunlight, precipitation, and uneven terrain are all supported. In low light environments or for night operation, an optional illumination module can be integrated into the system.
3D Obstacle Detection & Free Space Detection
RGo uses a unique, patent-pending approach for robust obstacle and free space detection. Obstacle detection data include distance, bearing angle, and height of obstacles in the robot path with cm-scale accuracy. Free space analysis provides the information needed for the path planning stack to calculate the optimal path to the destination.
Object Recognition & Scene Understanding (Coming Soon)
Perception Engine brings unprecedented understanding to robots. AI and additional machine learning techniques achieve human-level understanding of objects and space including Object-Type Labeling, Alerts, and Contextual Understanding. This enables the robot to behave more intelligently in its surroundings.
Collective Perception (Coming Soon)
When integrated with cloud- or server-side software, real-time data from Perception Engine enables sharing live maps and obstacles across a fleet of robots. Traffic-aware routing, human recognition and awareness of other mobile equipment improves safety and allows more collaborative deployments in environments that include both people and robots.
Intelligent Behavior & Path Planning (Coming Soon)
With all the information provided by the Perception Engine, we can enable smarter robot behavior in dynamic environments, including advanced path planning to enable robots to operate next to humans and other moving machines.
Tracking & ADAS
Real-time data from Perception Engine brings advanced driver assist features including collision warning, speed monitoring, and route guidance and optimization to manned industrial vehicles including forklifts. Its object recognition and barcode-reading capabilities can be leverage for additional capabilities including inventory mapping, row / shelf guidance and pick verification.