摘要： Two decades of research have helped show that government agencies can be innovative under certain conditions. We test hypotheses about the adoption and use of robotics as a key emerging leading-edge technology as advanced economies undergo the latest technological revolution. We focus on the case of U.S. crime laboratories as a core component of the "evidence assembly process" in the U. S. justice system. Using data from the census of crime labs, we show that the adoption of robotics depends on familiar "push-and-pull" factors: the push of agency professionalism, the pull of agency task environments, and the supporting capability of resources. Together these findings suggest that agencies can be early adopters of robotics as advanced technologies if they have the capacity (and need) to do so.

关键词： chord recognition   multi-task learning   guitar-playing robot   convolutional neural network   data augmentation  

摘要： In recent years, many researchers have shown increasing interest in music information retrieval (MIR) applications, with automatic chord recognition being one of the popular tasks. Many studies have achieved/demonstrated considerable improvement using deep learning based models in automatic chord recognition problems. However, most of the existing models have focused on simple chord recognition, which classifies the root note with the major, minor, and seventh chords. Furthermore, in learning-based recognition, it is critical to collect high-quality and large amounts of training data to achieve the desired performance. In this paper, we present a multi-task learning (MTL) model for a guitar chord recognition task, where the model is trained using a relatively large-vocabulary guitar chord dataset. To solve data scarcity issues, a physical data augmentation method that directly records the chord dataset from a robotic performer is employed. Deep learning based MTL is proposed to improve the performance of automatic chord recognition with the proposed physical data augmentation dataset. The proposed MTL model is compared with four baseline models and its corresponding single-task learning model using two types of datasets, including a human dataset and a human combined with the augmented dataset. The proposed methods outperform the baseline models, and the results show that most scores of the proposed multi-task learning model are better than those of the corresponding single-task learning model. The experimental results demonstrate that physical data augmentation is an effective method for increasing the dataset size for guitar chord recognition tasks.

关键词： Social robotics   Education   Engineering   Culture   Cultural background  

摘要： This study investigates the level of familiarity and interest of students towards social robotics through a survey conducted with the Electrical Engineering and Computer Science Students at Heriot-Watt University in the UK and Electrical Engineering Students at Xidian University in China. The results indicate that whereas there is no significant difference in the level of familiarity within the three groups of students and no significant difference in the level of interest between the Electrical Engineering and Computer Science Students at Heriot Watt University, there is a statistically significant difference in the level of interest towards social robotics between the Heriot-Watt University and Xidian University Students. Xidian University Students demonstrate a higher level of interest towards social robotics. The qualitative analysis shows that many of the Xidian University Students are willing to perceive and have robots as companions whereas none of the Heriot-Watt University Students show such or similar tendency. This observation indicates that cultural background plays a significant role in interests and preferences of the students towards social robotics.

关键词： nonholonomic constraint   rubber rolling   unbalanced ball   rolling on a plane  

摘要： The motion of a spherical robot with periodically changing moments of inertia, internal rotors and a displaced center of mass is considered. It is shown that, under some restrictions on the displacement of the center of mass, the system of interest features chaotic dynamics due to separatrix splitting. A stability analysis is made of the upper equilibrium point of the ball and of the periodic solution arising in its neighborhood, in the case of periodic rotation of the rotors. It is shown that the lower equilibrium point can become unstable in the case of fixed rotors and periodically changing moments of inertia.

关键词： Sliding mode control  

摘要： Tethered Space Net Robot (TSNR) is proposed as a hopeful solution for space debris capturing. An unideal case of net capture is studied in this paper, in which the net does not contact the target in the center of the net. For this case, the net cannot naturally close to wrap around the target completely after collision due to the asymmetric net configuration. Therefore, an integral adaptive Super-Twisting sliding mode control (IASTSMC) is designed to close the net. The control algorithm consists of a nonlinear integral sliding surface and adaptive Super-Twisting sliding mode control (ASTSMC), which not only has the property of "saturate the large error, magnify the small error", but also provides a positive chattering alleviation effect. Compared with the nonlinear integral sliding mode control (ISMC) with equivalent control technique and the original Super-Twisting algorithm (STA), the proposed control method shows a better performance to close the net in this unideal case. (C) 2020 The Franklin Institute. Published by Elsevier Ltd. All rights reserved.

关键词： Robotic palpation   Tactile sensing   Physical sensing   Sensory-motor coordination  

摘要： This paper proposes a framework to investigate the influence of physical interactions to sensory information, during robotic palpation. We embed a capacitive tactile sensor on a robotic arm to probe a soft phantom and detect and classify hard inclusions within it. A combination of PCA and K-Means clustering is used to: first, reduce the dimensionality of the spatiotemporal data obtained through the probing of each area in the phantom;second categorize the re-encoded data into a given number of categories. Results show that appropriate probing interactions can be useful in compensating for the quality of the data, or lack thereof. Finally, we test the proposed framework on a palpation scenario where a Support Vector Machine classifier is trained to discriminate amongst different types of hard inclusions. We show the proposed framework is capable of predicting the best-performing motion strategy, as well as the relative classification performance of the SVM classifier, solely based on unsupervised cluster analysis methods.

关键词： Optimal control   Mobile robots   Task analysis   Dynamical systems   Training   Real-time systems   Autoassembly mobile robot   Lyapunov techniques   path following   path integral   reinforcement learning (RL)  

摘要： Reinforcement learning (RL) combined with deep neural networks has led to a number of great achievements for robot control in virtual computer environments, where sufficient data can be obtained without any difficulty to train various models. However, thus far, only few and relatively simple tasks have been accomplished for practical robots, which is mainly caused by the following two reasons. First, training with real robots, especially with dynamic systems, is too complicated to be fully and accurately represented in simulations. Second, it is very costly to obtain training data from real systems. To address these two problems effectively, in this article, a path-integral-based RL algorithm is proposed for the task of path following of an autoassembly mobile robot, wherein three kernel techniques are introduced. First, a generalized path-integral-control approach is proposed to obtain the numerical solution of a stochastic dynamical system, wherein the calculation of the gradient and kinematics inverse is avoided to ensure fast and reliable training convergence. Second, a novel parameterization method using Lyapunov techniques is introduced into the RL algorithm to ensure good performance of the system when directly transferring simulation results into practical systems. Third, the optimal parameters for all discrete initial states are first learned offline and then tuned online to improve the generalization and real-time performance. In addition to the optimization control for the mobile robot, the proposed method also possesses general applicability for a class of nonlinear systems such as crane systems. Simulation and experimental results are included and analyzed to illustrate the superior performance of the proposed algorithm.

关键词： Swarm robotics   Foraging   Probabilistic timed supervisory control theory   ptSCT  

摘要： The development of flexible swarm robotics systems capable of adapting to the task and environmental changes is a serious challenge. The main motivations of Swarm robotics are decentralized control, stability, adaptivity, and flexibility. Usually, ad-hoc approaches are employed to design a controller capable of meeting the problem specifications. However, these methods cannot be easily verified, and in some cases, it is not even shown that they meet the specifications. Moreover, the controller source code has to be developed separately, primarily when formal methods are employed;As a result, it cannot be guaranteed that the implementation matches the design. This paper proposes a new method - probabilistic timed supervisory control (ptSCT) - to formally design a controller from systems specifications. The proposed ptSCT has several advantages: 1) the automatic generation of the controller source code utilizable in ARGoS platform, 2) formal designing capability using the implemented software tool, 3) set of powerful design components like probabilistic decisions and time constraints, and 4) the reusability of formally designed modules among different scenarios and multiple robotic platforms. Two case studies are considered to investigate various aspects of the proposed system. Firstly, the synchronization case study is implemented for a comparison between SCT and ptSCT in terms of design capabilities and memory consumption. Secondly, the foraging case study as a complex and medium-sized problem is modeled using ptSCT step by step. More than 2400 experiments with a varying number of obstacles, targets, and robots are executed in ARGoS platform in order to show the performance of the automatically generated source code.

关键词： autonomous surgery   machine learning   robotic surgery   surgical assessment   surgical training  

摘要： Purpose of review The increasing use of robotics in urologic surgery facilitates collection of 'big data'. Machine learning enables computers to infer patterns from large datasets. This review aims to highlight recent findings and applications of machine learning in robotic-assisted urologic surgery. Recent findings Machine learning has been used in surgical performance assessment and skill training, surgical candidate selection, and autonomous surgery. Autonomous segmentation and classification of surgical data have been explored, which serves as the stepping-stone for providing real-time surgical assessment and ultimately, improve surgical safety and quality. Predictive machine learning models have been created to guide appropriate surgical candidate selection, whereas intraoperative machine learning algorithms have been designed to provide 3-D augmented reality and real-time surgical margin checks. Reinforcement-learning strategies have been utilized in autonomous robotic surgery, and the combination of expert demonstrations and trial-and-error learning by the robot itself is a promising approach towards autonomy. Robot-assisted urologic surgery coupled with machine learning is a burgeoning area of study that demonstrates exciting potential. However, further validation and clinical trials are required to ensure the safety and efficacy of incorporating machine learning into surgical practice.