Intelligent acoustic rotor speed estimation for an autonomous helicopter

Acoustic sensing to gather information about a machine can be highly beneficial, but processing the data can be difficult. In this work, a variety of methodologies have been studied to extract rotor speed information from the sound signature of an autonomous helicopter, with no a-priori knowledge of its underlying acoustic properties. The autonomous helicopter has two main rotors that are mostly identical. In order to identify the rotors’ speeds individually, a comparative evaluation has been made of learning methods with input selection, reduction and aggregation methods. The resulting estimators have been tested on unseen training data as well as in actual free-flight tests. The best results were found by using a genetic algorithm to identify the important frequency bands, a centroid method to aggregate the bands, and a neural-network estimator of the rotor speeds. This approach succeeded in estimating individual rotor speeds of an autonomous helicopter without being distracted by the other, mainly identical, rotor. These results were achieved using standard, low-cost hardware, and a learning approach that did not require any a-priori knowledge about the system's acoustic properties.