Jinhong Qu

Position estimation for a capacitively-sensed magnetoelastic rotary microstage using an extended Kalman smoother

An estimation scheme is described for measuring angular rotation of a magnetoelastic rotary micro-stage with capacitive sensing electrodes. Experimental measurements of micro-stage motion are used to create a simulation model for translational dynamics of the stage rotor. A capacitive sensing electrode geometry providing stable peak capacitance measurements at specific rotation angles is then introduced. Application of an extended Kalman smoother to the measurement of rotation angle in the simulated system is shown to produce high precision estimates of angular position near peak capacitance locations.

Dynamics of millimeter-scale hexapod microrobotics with PZT-polymer micro-actuators

This paper examines the dynamics of a type of millimeter-scale hexapod microrobot based on piezoelectric actuation. These robots [1], having a 5mm × 2 mm footprint, are formed from integrated thin-film lead zirconate titanate (PZT) and high-aspect-ratio parylene-C polymer structures. The in-chip dynamics of the microrobots are measured when actuated with tethered electrical signal, to demonstrate the resonant behavior of different parts of the robot. Out-of-chip robot motion is then actuated by external vibration after the robot has been detached from its silicon tethers to external power. A dynamic model for robot and ground interaction is presented and validated to explain robot locomotion in the vibrating field using the in-chip measurements of actuator dynamics and certain additional design information. The model accounts for the interaction between the robot and ground, for multiple resonances of the robot leg, and for rigid robot body motion in 3 degrees of freedom. The dynamic model with first vertical and lateral resonant modes of leg shows a good match with experimental results for the motion of the robot on a vibrating surface actuated within low-frequency range.

Estimators for micro-robot leg dynamics with uncertain jump conditions and infrequent sampling

This paper examines estimation of micro-robotic leg dynamics across uncertain ground impact conditions. A model for impact of compliant micro-structures with unknown ground geometries is adapted to a switched system format having explicit random parameters associated with uncertain impact conditions. Estimation schemes for dynamic states of the robot leg are derived based on Kalman filtering, including a novel ad hoc method for incorporating inter-sample impact events based on linear interpolations. Model and estimation schemes are tested against experimental motion of a meso-scale compliant piezoelectric robot test bed. The impact model improves on prior modeling techniques in capturing mean impact time and variability of step height, though impact time variation is overstated in open-loop. Impact time estimates are much improved using the estimator with inter-sample events, though estimation error over the entirety of leg motion is inferior to that of a more conventional hybrid Kalman filter.