Teaching an Old Robot New Tricks: Learning Novel Tasks via Interaction with People and Things

As AI has begun to reach out beyond its symbolic, objectivist roots into the embodied, experientialist realm, many projects are exploring different aspects of creating machines which interact with and respond to the world as humans do. Techniques for visual processing, object recognition, emotional response, gesture production and recognition, etc., are necessary components of a complete humanoid robot. However, most projects invariably concentrate on developing a few of these individual components, neglecting the issue of how all of these pieces would eventually fit together. The focus of the work in this dissertation is on creating a framework into which such specific competencies can be embedded, in a way that they can interact with each other and build layers of new functionality. To be of any practical value, such a framework must satisfy the real-world constraints of functioning in real-time with noisy sensors and actuators. The humanoid robot Cog provides an unapologetically adequate platform from which to take on such a challenge. This work makes three contributions to embodied AI. First, it offers a general-purpose architecture for developing behavior-based systems distributed over networks of PC's. Second, it provides a motor-control system that simulates several biological features which impact the development of motor behavior. Third, it develops a framework for a system which enables a robot to learn new behaviors via interacting with itself and the outside world. A few basic functional modules are built into this framework, enough to demonstrate the robot learning some very simple behaviors taught by a human trainer. A primary motivation for this project is the notion that it is practically impossible to build an "intelligent" machine unless it is designed partly to build itself. This work is a proof-of-concept of such an approach to integrating multiple perceptual and motor systems into a complete learning agent.

Interaction and Intelligent Behavior

We introduce basic behaviors as primitives for control and learning in situated, embodied agents interacting in complex domains. We propose methods for selecting, formally specifying, algorithmically implementing, empirically evaluating, and combining behaviors from a basic set. We also introduce a general methodology for automatically constructing higher--level behaviors by learning to select from this set. Based on a formulation of reinforcement learning using conditions, behaviors, and shaped reinforcement, out approach makes behavior selection learnable in noisy, uncertain environments with stochastic dynamics. All described ideas are validated with groups of up to 20 mobile robots performing safe--wandering, following, aggregation, dispersion, homing, flocking, foraging, and learning to forage.

Investigation of aspect ratio of hole drilling from micro to nanoscale via focused ion beam fine milling

Holes with different sizes from microscale to nanoscale were directly fabricated by focused ion beam (FIB) milling in this paper. Maximum aspect ratio of the fabricated holes can be 5:1 for the hole with large size with pure FIB milling, 10:1 for gas assistant etching, and 1:1 for the hole with size below 100 nm. A phenomenon of volume swell at the boundary of the hole was observed. The reason maybe due to the dose dependence of the effective sputter yield in low intensity Gaussian beam tail regions and redeposition. Different materials were used to investigate variation of the aspect ratio. The results show that for some special material, such as Ni-Be, the corresponding aspect ratio can reach 13.8:1 with Cl₂ assistant etching, but only 0.09:1 for Si(100) with single scan of the FIB.

Plasma polymerization of C[subscript 4]F[subscript 8] thin film on high aspect ratio silicon molds

High aspect ratio polymeric micro-patterns are ubiquitous in many fields ranging from sensors, actuators, optics, fluidics and medical. Second generation PDMS molds are replicated against first generation silicon molds created by deep reactive ion etching. In order to ensure successful demolding, the silicon molds are coated with a thin layer of C[subscript 4]F[subscript 8] plasma polymer to reduce the adhesion force. Peel force and demolding status are used to determine if delamination is successful. Response surface method is employed to provide insights on how changes in coil power, passivating time and gas flow conditions affect plasma polymerization of C[subscript 4]F[subscript 8].