Moving Toward High-Information Navigation Maps: Efficient Camera and Lidar Fusion. Ottonomy @NVIDIA GTC
23rd March 2022, 7:30 PM – 7:55 PM IST
Autonomous robots are entering various spheres of life. Recently, we’ve seen a huge increase in robots operating in regular human environments such as malls, airports, and sidewalks. Most new applications require robots to operate both indoors and outdoors. While the use cases are growing rapidly, the datasets available for these use cases are very limited. This results in a need for the following:
- a balance between robust classical approaches and latest learning techniques,
- low-compute, real-time fusion of multimodal sensors, especially 3D lidars and RGB cameras, and
- navigation maps that combine both metric and semantic information of the world. We’ll describe some of the fusion techniques and NVIDIA tools our delivery robots use to achieve these objectives while operating safely and efficiently in crowded indoor and outdoor environments.
Industry Segment: Automotive / Transportation
Primary Topic: Autonomous Machines – Robotics
