A High Precision Aerial Indoor AI-Based Inspection System

The project entitled High Precision Aerial Indoor AI-Based Inspection System - PROBE aims to address the limitations of current drone navigation systems by developing an improved inertial navigation solution for indoor applications. In recent years, inertial odometry has emerged as a promising approach to enhance navigation performance, particularly in GNSS-denied environments such as indoor or obstructed areas. Its ability to model complex sensor dynamics and correct error drift makes it an ideal candidate to support or replace traditional positioning aids. The core innovation of PROBE is located in leveraging inertial odometry to compensate for the drift typically associated with pure inertial sensors. The use of advanced Artificial Intelligence based algorithms to improve traditional navigation configurations is a groundbraking field, even if propoer data collection and data management procedures are needed to guaranteed the generalization of the achieved results. Specifically, the presented paper describes a methodology for system configuration development and data acquisition aimed at the training process of Artificial Intelligence based algorithms acquiring an input-output dataset based on a custom set of Commercial Off The Shelf sensors to improve the inertial navigation estimates.