Currently, the locations for sensor installations within a room or indoor space are based on broad assumptions about physical spaces, which can result in sensor locations that provide suboptimal coverage in spaces that do not conform to the assumptions. This disclosure describes techniques to add intelligent sensor placement capabilities to products and services for installing and managing sensors in indoor spaces. With user permission, suitable placement of various sensors in a given room can be determined automatically by a trained neural network based on various criteria such as room coverage with the sensor field-of-view (FoV), number of sensing devices of a given type, etc. A coarse 3D representation of a room, e.g., obtained using simultaneous localization and mapping (SLAM) techniques is fused with floor plans for the indoor space to obtain a refined map. A trained neural network takes as input the refined map and generates sensor placement locations as output. Custom recommendations can alleviate the burden manually determining custom sensor placements and help achieve optimal sensing coverage.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Shin, D, "Custom Recommendations for Optimal Sensor Placement in Indoor Spaces", Technical Disclosure Commons, (November 22, 2023)