Improved learning in an evolvable hardware hover controller for an insect scale Flapping-Wing Micro Air Vehicle

In previous work, a single Degree Of Freedom (DOF) controller for altitude of a Flapping-Wing Micro Air Vehicle (FW-MAV) has been presented which meets the various needs of such a device. Unique challenges posed by the small size of such a device include accounting for process variation during the manufacturing of small scale wings and effectors, damage accumulation during the operation of the device, as well as weight and power restrictions limiting the computational power available on-board the device have been overcome with a cycle-average controller which preserves mathematical rigor while still allowing computational power for evolvable hardware on-board the device. This paper will describe an attempt to optimize altitude control learning times on a controller which has been shown to support single DOF flight with the intent to shorten learning times of a multiple DOF flight controller which is much more computationally intense.