3D visual obstacle avoidance for mobile robots

Mobile robots operating in warehouses, hospitals, and service environments must detect both static and moving obstacles on their paths. 3D vision provides depth and RGB data for obstacle detection, classification, and avoidance planning.

​

Perception limits of 2D LiDAR

2D LiDAR normally scans a single plane, which creates two limitations for obstacle avoidance:

• No height information: The scan can miss low or overhanging obstacles outside the scanning plane.
• Limited scene information: A planar contour does not describe an obstacle’s full spatial shape or texture.

​

Solution overview

For mobile-robot obstacle avoidance, MRDVS provides the RGB-D S Series multimodal industrial obstacle-avoidance cameras. The cameras output 3D spatial and RGB texture data for detecting low, overhanging, and other obstacles. Supporting algorithms use depth and RGB data for object recognition and classification.

​

Application scenarios

• Autonomous forklifts, stackers, and under-ride AMRs
• Bin-transport and industrial cleaning robots
• Medical and service robots

​

Solution highlights

​

Operating mode