Flapping flight for biomimetic robotic insects: System modeling and flight control in hover

The Berkeley Micromechanical Flying Insect project aims to design and develop a micro air vehicle (MAV) the size of a typical housefly which is capable of autonomous flight using a pair of independent flapping wings. The goal of this dissertation is to (1) develop a complete system dynamic model; (2) analyze and design flapping flight control algorithms for the MFI.; The first part of this dissertation presents a systematic analysis and mathematical modeling of flapping flight for inch-size MAVs. In particular, the development and implementation of a Virtual Insect Flight Simulator (VIFS) is presented. It is a simulation tool intended for modeling true insect flight mechanisms and for testing the flight control algorithms for the MFI. The VIFS includes dynamic models that have several elements which greatly differ from those with either larger rotary, or fixed wing MAVs. It consists of a quasi-steady state flapping flight aerodynamic model at low Reynolds number regime, a body dynamic model, a linearized wing-thorax dynamic model, and a biomimetic sensory system consisting of ocelli, halteres, magnetic compass and optical flow sensors. Each of these models are developed based on biological principles, mathematical analysis and experimental data, and they are designed in a modular structure for quick upgrading. These models are integrated together to give a realistic simulation for MFI flapping flight. The VIFS is intended to serve as a simulation tool to evaluate the performance of the MFI flight control unit with an accurate low-level modeling of dynamics, actuators, sensors and environment.; The second part of this dissertation presents the analysis and development of flight control algorithm for flapping flight MAVs. Inspired by the sensory feedback and neuromotor structure of biological flying insects, a similar top-down hierarchical architecture to achieve high performance despite the MFIs' limited computational resources is proposed. Flight stabilization problem is formulated as high frequency periodic control of an underactuated system. In particular, a methodology to approximate the time-varying body dynamics caused by the aerodynamic forces with time-invariant dynamics using averaging theory and a biomimetic parametrization of wing trajectories is presented. This approximation leads to a simpler dynamical model that can be identified using experimental data from the onboard sensors and the input voltages to the wing actuators. Moreover, the overall control law is a simple periodic proportional output feedback. Simulations, including sensor and actuator models, demonstrate stable flight in hovering mode. Moreover, the controller design methodology developed here is not limited to the mathematical models of aerodynamics considered in this dissertation, and it can also be easily adapted to the MFI experimental data as it becomes available.