AMR (Outdoor, Indoor-Outdoor mix)

Use Case Description

Various types of AMRs which mainly operate at outdoor or indoor-outdoor mix environments such as factory, construction site, office building, shopping mall and hospital, especially with dynamic objects and scenery changes.

Overview

Visual/3D-Lidar based SLAM enables operating AMR at following environments/usage, which existing GNSS/GPS based outdoor AMR couldn’t handle:

GNSS/GPS denied or unstable route such as high office building areas, passing through warehouses, running along the buildings
Can be operated at indoor-outdoor mix environment
Can be started from anywhere on the map even at GNSS/GPS denied area (doesn’t need to set initial position in advance)

Kudan’s Visual/3D-Lidar SLAM can be offered independently, or integrated to leverage both camera and Lidar. Also, it can be fused with other sensors flexibly (e.g. IMU, 2D-Lidar, Wheel Odometry, GNSS/GPS, Marker), depending on use case requirement.

Kudan’s unique map-handling features also helps to reduce costs in deployment and operations of autonomous applications at broader scale, rather than the limited area for PoCs.

Furthermore, Kudan is capable to support full navigation system in addition to SLAM.

The benefits of Kudan software

More accurate and robust localization with Visual SLAM and/or 3D-Lidar SLAM against scenery changes and dynamic objects
Faster time-to-market compared to internal development or open-source based approach
Productivity improvement due to faster operation speed
Overall hardware cost reduction through lower grade sensors

Customer Reference Case (Selected)

Visual and 3D-Lidar fused AMR at indoor-outdoor mix shopping mall (China, Yours Technologies)
3D-Lidar based multi purpose robot at indoor-outdoor use cases (China, Whale Dynamic)
GNSS and 3D-Lidar fused AMR at construction site (US)

Relevant Demos

Kudan & NVIDIA Collaboration: Integration of Kudan Visual SLAM with NVIDIA Isaac Perceptor
KdVisual in action: Overcome Typical Challenges of Real-Time Forklift Positioning
KdVisual in action: Multiple camera SLAM