Flight evaluation of deep model reference adaptive control

This thesis presents flight test results for a new neuroadaptive architecture: Deep Neural Network based Model Reference Adaptive Control (DMRAC). This architecture utilizes the power of deep neural network representations for modeling significant nonlinearities while marrying it with the boundedness guarantees that characterize MRAC based controllers. Through experiments on a real quadcopter platform, it is shown that DMRAC can outperform state of the art controllers in different flight regimes while having long-term learning abilities. This makes DMRAC a highly powerful architecture for high-performance control of nonlinear systems