Biomimetic sensor modeling and simulations for flight control of a micromechanical flying insect

Inspired by the exceptional flight capabilities of flying insects, the UC Berkeley Micromechanical Flying Insect (MFI) project entails the development of a flapping wing micro aerial vehicle that will be capable of sustained autonomous flight. It has a target size of 25mm wing span and a mass of 100mg. Biological principles are used for actuation, transmission, sensing, and control of the MFI so that the device will exhibit the same performance merits as those observed in real flying insects.; This dissertation first describes the sensing mechanisms used by real insects and then presents formal models of biologically inspired sensory systems including optic flow sensors, angular position sensors, and angular rate sensors. The analysis and simulations of the proposed sensor models suggest the feasibility of using these biomimetic sensors for flight control of a robotic flying insect. It is also demonstrated that a number of insect flight behaviors can be reproduced using simple control schemes based on these sensors. An example of attitude stabilization is given in which a proportional control law using the outputs from the ocelli and halteres as feedback is able to steer the insect toward the upright posture from any initial body orientation and angular velocities.; This dissertation also includes the design of an optic flow sensor, an ocelli sensor, a haltere sensor, and a magnetic field sensor for use on the MFI. Preliminary experimental results of these prototype sensors show promising performance. Compared to existing commercial micro sensors, these devices have the advantages of simple structure, easy implementation, simple signal processing, and low power consumption. Therefore, these sensors are particularly appropriate for micro robotic platforms that have limited computational resources, little power budget, and small payload capacity.