Intelligent Flight State Identification of a Self-Sensing Wing through Neural Network Modelling

Abstract

With the development of micro-fabrication techniques, multi-functional sensor networks have been created with many micro-sensor nodes for various functions. An UAV wing is integrated with a stretchable sensor network including piezoelectric transducer, strain gauge and resistance temperature detector embedded near the wing surface and this novel design intends to equip the wing with the self-sensing capability similar to bird feathers. One of the main challenges of the state-of-the-art research is how to make the wing aware its flight states in real-time through large amount of sensing signals. Features are extracted and further selected for neural network modelling. Both classic neural network and deep belief network are proposed respectively to map the relationship from the sensing data to physical flight states and compare the identification accuracy with each other. The simulation results show a relatively high identification accuracy of the proposed methods, enabling new perspectives in developing intelligent self-awareness capabilities for next generation smart wings.

Flight State Identification of a Self-Sensing Wing via an Improved Feature Selection Method and Machine Learning Approaches

“…We have demonstrated the effectiveness of establishing the mapping relationship from the feature space to the flight state space through neural networks modelling [ 27 ]. This paper significantly improves the previous work by creating a much larger feature pool and considering the co-linearity among various features.…”