Fig. 4: Experiment results and analysis for the wing in pre-stall condition with an 8° angle of attack.
From: Super-Turing synaptic resistor circuits for intelligent morphing wing
a The drag-to-lift force ratios, \({s}_{1}={F}_{D}/{F}_{L}\), b the magnitude of the fluctuation of the drag-to-lift force ratio, \({s}_{2}\), and c the objective function, \(E=\frac{1}{2}{{{\bf{s}}}}^{2}\) (arbitrary unit), d \(E\) and \(\frac{\partial E}{\partial {y}_{n}}\) (arbitrary unit, Methods) for the wing controlled by a synstor circuit (left), a human (middle), and an ANN (right) are displayed versus time \(t\) (gray and black lines), where \(t\) represents concurrent inference and learning time in synstor and human neurological circuits, and the cumulative time for sequential inference and learning in the ANN. The average values of \({s}_{1}\) and \(E\) are also shown versus time \(t\) (red lines). The \(E-t\) curves are best fitted by \(E\left(t\right)=\left(E\left(0\right)-{E}_{e}\right)\,{e}^{-t/{T}_{L}}+{E}_{e}\) (green lines).
