Over the past two years, the research and development team has been developing a method to identify faulty motors on a drone without interrupting a mission (automated or manual). This is of high interest as it will give the pilot more information to ensure the successful recovery of the system in the event of a fault. The team of engineers has developed a Machine learning (using Artificial Neural Networks) framework that superimposes the dynamics on the controls in order to detect and locate a fault in a rotor without compromising the mission. The graph above shows simulate a rotor fault once the drone reaches 9 m/s and the fault identification system (FIS) detects a fault 1-second later. Once the other rotors are analyzed, the Rotor1 is identified as having a fault and this information is sent back to the pilot. It's important to know that even though a major fault has occurred, the drone a capable of flying for some time until instability grows and the drone becomes unc...

So one of the main objectives of my PhD research was to achieve the difficult task of developing a learning algorithm for machine learning ( RBF neural networks to be more specific) applications, that would enable the prediction of drone propeller damage in real-time AND without altering the bought-out flight controller ( DJI Naza , APM , Pixhawk , etc...). The only way it could achieve that was by analyzing the outputs of the flight controller sensors and learn when a fault would occur. Well, I believe I'm getting closer to this objective (submission is Nov 2019). I've decided to include the two figures below which illustrates the training process (0.2 sec on desktop) and prediction time (0.008 sec) and the accuracy to the true dynamics of the quadcopter drone. In this case the pitch dynamics are being predicted. Although noise hasn't been introduced, it's quite clear from the graphs, that the learning algorithm has enabled the RBF network to accurately capture th...