1000realities

CLOUDSlam is our own implementation of SLAM algorithm allowing positional tracking of people and machines equipped with simple, low cost hardware. We’re using standard-definition video stream as main source of data, which is analysed in a cloud to build a point-cloud representing any interior the user might be in. The point cloud is used to locate this user within that space. Additional information like IMU data (device orientation and acceleration) or WiFi signals can be used to further improve positioning performance.
Thanks to moving processing to the server/cloud, end user equipment can be a simple, low-cost and power efficient device in variety of form factors. We’re using light smart glasses to enable AR maintenance use-cases (backend system can precisely detect user’s location an provide augmenting data to whatever the user is looking at). Another scenario assumes company-wide deployment of small end-user devices in a form of smart badges, which help to track position of all employees and in parallel detect people accessing certain area without authorization.
In general, our application is an enabler for variety of use-cases that require precise positioning of people and machines. Without need of using sophisticated, expensive end-user hardware, mass deployments become possible. Another advantage over „classic” indoor localization systems is complete lack of necessity to install any sensors. Pure connectivity and access to low latency computing is sufficient, which is cost-effective and also mandatory in many industrial applications.