Perceiving,learning, and exploiting object affordances for autonomous pile manipulation

Autonomous manipulation in unstructured environments will enable a large variety of exciting and important applications. Despite its promise, autonomous manipulation remains largely unsolved. Even the most rudimentary manipulation task—such as removing objects from a pile—remains challenging for robots. We identify three major challenges that must be addressed to enable autonomous manipulation: object segmentation, action selection, and motion generation. These challenges become more pronounced when unknown man-made or natural objects are cluttered together in a pile. We present a system capable of manipulating unknown objects in such an environment. Our robot is tasked with clearing a table by removing objects from a pile and placing them into a bin. To that end, we address the three aforementioned challenges. Our robot perceives the environment with an RGB-D sensor, segmenting the pile into object hypotheses using non-parametric surface models. Our system then computes the affordances of each object, and selects the best affordance and its associated action to execute. Finally, our robot instantiates the proper compliant motion primitive to safely execute the desired action. For efficient and reliable action selection, we developed a framework for supervised learning of manipulation expertise. To verify the performance of our system, we conducted dozens of trials and report on several hours of experiments involving more than 1,500 interactions. The results show that our learning-based approach for pile manipulation outperforms a common sense heuristic as well as a random strategy, and is on par with human action selection.     

Electrospinning pure protein solutions in core–shell fibers

Electrospinning of protein‐loaded fibers faces many challenges, e.g. burst release owing to segregation of the protein on the fiber surface, loss of activity due to electrospinning conditions, limitation of loading capacity etc. Core–shell electrospinning provides an effective way to electrospin fibers wherein the core can be loaded with bioactive molecules in friendly conditions of a compatible polymer solution, thereby protecting the molecules from the electrostatic field and organic solvent of shell solutions. The shell polymer, after the electrospinning, acts as a barrier to control the release of the loaded molecules. However, the limitation of loading capacity still remains due the prerequisite of using an additional polymer as additive to achieve the minimum viscosity of the core solution required for viscous drag by the shell solution being drawn by the electrostatic force. The work reported here aims to alleviate the need of a polymer additive by using aqueous protein solutions of very high concentration. High concentrations of protein solutions were successfully electrospun as the core of the protein–poly(lactide‐co‐glycolic acid) core–shell fibers. A partitioning effect was seen in the controlled release of hydrophilic proteins as they were retained in the aqueous core for longer times. Using lysozyme as a model protein, it was shown that the activity is significantly retained after electrospinning, compared with electrospinning in monolithic fibers. Moreover, the lysozyme activity was also comparable with the lysozyme released from core–shell fibers spun using poly(vinyl acetate) as additive in the core. Copyright © 2012 Society of Chemical Industry     

Importance of viscosity parameters in electrospinning: Of monolithic and core–shell fibers

Electrospun polymeric fibers are attractive candidates in the development of scaffolds for the tissue engineering and for providing new systems for delivery of bioactive molecules. Co-axial fibers have emerged as an efficient tool to protect the core material from the adverse conditions of electrospinning process, to spin difficult-to-process fluids and to generate fibers with much more control of the delivery of encapsulated bioactive molecules. Currently, there is very little reported work on the optimization of the processing parameters of electrospinning, especially core–shell electrospinning. This study extends the understanding of the role of solution viscosity as a vital material parameter for electrospinning of fibers. The spinning solutions were characterized for viscosity and optical imaging of the compound Taylor cone for spinnability, and the fibers were imaged by Scanning Electron Microscopy (SEM). Our experimental results, using PLGA as the model polymer, confirm that the solution concentration be above the entanglement concentration (C_e) to obtain uniform beadless monolithic fibers; for core–shell fibers, the shell solution must fulfill the above criterion for spinnability and, further, the ratio of the viscosities of core and shell solutions (η_core/η_shell) has to be greater than a threshold value to get a stable compound Taylor cone and therefore to obtain uniform beadless core–shell fibers. Addition of surfactant led to reduction of the threshold η_core/η_shell (from 0.55 to 0.18) for the PVA–PLGA system.     

Understanding e‐Government portal use in rural India: role of demographic and personality characteristics

Electronic government (e‐Government) is one of the most important ways to bridge the digital divide in developing countries. We develop a model of e‐Government portal use. We use various individual characteristics, namely demographics and personality, as predictors of e‐Government portal use. Specifically, our predictors were (1) gender, age, income and education; (2) the Big Five personality characteristics, i.e. extraversion, neuroticism, conscientiousness, agreeableness and openness to experience; and (3) personal innovativeness with information technology. We conducted a field study in a village in India. We collected data from over 300 heads of household. We found support for our model, with most variables being significant and explaining 40% of the variance in e‐Government portal use.     

Defect Characterization Using Pulsed Thermography

Infrared Thermography is one of the advanced NDE methods that is becoming attractive due to its ability to inspect non invasively large areas in short times and provide full field images in a non contact nature. While initially the method started as a qualitative technique for defect detection alone, with the advent of lock in and pulsed techniques, quantitative defect detection was made possible. While these techniques have been applied in case of ceramics and composites for defect quantification, defect sizing and quantification using these methods have not been systematically attempted in case of stainless steels. A systematic study has been undertaken to evaluate the accuracy of these empirical Pulsed Thermography (PT) techniques for defect size and depth estimation in type 316 L austenitic stainless steels. Theoretical modelling based on finite difference analysis using Thermo Calc 6L software has also been done and the results compared.     

Effect of dietary fat on diabetes-induced changes in liver microsomal fatty acid composition and glucose-6-phosphatase activity in rats

Experimental diabetes may manifest itself in a defect in liver microsomal fatty acid desaturation and increased activity of glucose-6-phosphatase (G-6-Pase). The present study was designed to determine whether these changes could be normalized by a change in the dietary fat consumed. Control and streptozotocin-induced diabetic rats were fed nutritionally adequate diets which varied in fatty acid composition. Fatty acid analysis of liver microsomal phospholipids revealed that non-diabetic control animals fed saturated fat (beef tallow) or a diet high in ω3 fatty acids (fish oil) exhibited a significantly higher level of 18∶2ω6 and a lower level of 20∶4ω6 in the phosphatidylcholine and phosphatidylethanolamine fractions compared with diabetic animals. Control and diabetic animals fed the high linoleic acid diet had similar levels of 18∶2ω6 in the microsomal phosphatidylcholine and phosphatidylserine fractions. Microsomal G-6-Pase activity was higher in diabetic than in control animals. Activity of G-6-Pase was lower in microsomes of control animals fed the soybean oil or the fish oil diet, but was not significantly reduced in diabetic animals fed high polyunsaturated fats. Blood glucose levels were similar in control groups fed the different diets, but the plasma hemoglobin A1c level was lower in diabetic animals fed the soybean oil diet. Cholesterol and triglyceride levels were lower in diabetic animals fed the fish oil-based diet. The results suggest that dietary fat manipulation has the potential to change at least some of the abnormalities in the microsomal membrane in experimental diabetes.