| The attraction of the interdisciplinary field of nanotechnology is the promise to build smaller functional units with tuned properties that surpass any material we encounter in the macro domain. A natural requirement to utilize these advances is building micro/nanorobotic systems to manipulate matter in these small scales. This task, however, is not a very straightforward mission to embark on. It is now well-known that physics in small scales have unexpected and nonlinear behavior due to the change in the dominance of forces. Due to a difference in scaling, prominent forces in the macro domain become more and more negligible in the micro/nano domain, which is dominated by adhesive or capillary interactions. Theoretically modeling and experimentally investigating to better understand this new sticky environment for micro/nanomanipulation is the first step one must take.Atomic force microscope (AFM) is a suitable tool for micro/nanomanipulation. It offers the benefit of locally and directly interacting with small scale phenomena on the order of a few nanometers. This ability provides precision and repeatability in micro/nanorobotics. Using an AFM for micro/nanomanipulation has its drawbacks as well. Since manipulating matter from the bottom up with a single end-effector is a serial process, it mainly suffers from a limitation on speed. Moreover, having no real-time visual feedback decreases the reliability of operation in the nanoscale. Due to these problems, future micro/nanomanipulation tools need to make automation of teleoperation possible, in contrast to current manipulation systems mostly being manually operated by a human being.In this work, we demonstrate two control approaches to micro/nanomanipulation with an AFM namely, teleoperated and automated control to increase speed and reliability. In both cases, a combination of micro/nanoscale physical theories, mathematical transformations, and control theory is utilized to achieve our objectives. We use fast and robust particle detection algorithms to detect positions of spherical micro/nanoparticles on a substrate from visual or topographical feedback, respectively. We demonstrate through experimental results that it is possible to push and pull particles on a flat surface into defined patterns, autonomously using an AFM probe tip, and with an error less than the particle diameter, and with success rates as high as 87%. By detecting contact loss through either visual or force feedback for micro and nanomanipulation, we bring control to the conventional, blind push-and-look approach. Patterns or assemblies of particles can be made using a commanding task planner that orders individual manipulation operations based on a minimization of trajectory blockage metric.For teleoperation, we focus on transparency and impedance reflection issues on scaled bilateral control for 1-D vertical motion of the tip. This motion is also animated using a general contact mechanics model to augment visual information with the force feedback. We use a modified passivity controller for stability. We also demonstrate 3-D force feedback from the micro/nanoscale for stable and transparent teleoperation. For this task, we propose a force decoupling algorithm that accounts for local surface orientation and solves an empirical friction model, in real-time during the experiments. This friction model is adaptive, so no initial calibration is necessary.In addition, we propose methods to correct two common problems in AFM based manipulation systems. First is the cross-talk of the two deflection signals measured by the optical system due to geometric, mechanical, and/or electrical reasons. We propose a method that is applicable to most AFM systems, to estimate and compensate the effects of cross-talk on the measurements using two initial experiments on a flat surface. Secondly, due to a thermal inequilibrium inside the AFM chamber, significant drift can occur between the tip and the substrate. We utilize a particle filter to localize the tip position with respect to the sample due to drift, using small linear observations of topography. |
