A multitude of flexure-based displacement amplifiers have been developed to amplify piezoelectric actuators for achieving both high-precision motion and large output stroke. Single-stage amplifiers are compact but provide only amplification of several times, while multi-stage amplifiers are able to achieve amplification of dozens of times but are generally bulky in structure. In this work, a displacement amplifier with only a single stage displacement amplifier, which shows the capability of obtaining a very large amplification ratio, is proposed. The displacement amplifier contains two semi-bridge mechanisms with a slight geometric difference between them. This slight difference makes the amplifier require a very small input while achieving a large displacement at the output, leading to a very large amplification ratio. A kinetostatic model considering the nonlinearities in the deflections of both the flexure hinges and the links for the amplifier is developed, based on which the parameters of the amplifier are optimized to maximize the amplification ratio, resulting in an amplifier exhibiting an amplification ratio of 107. The optimization results were validated by those of a finite element model, proving the effectiveness and correctness of the proposed amplifier and the kinetostatic model. The finalized design was prototyped and the measured amplification ratios in a bilateral output mode and a unilateral output mode are 98.10 and 88.42, respectively. A neural network PID controller was designed for the displacement amplifier, with a maximum trajectory tracking error less than 4.7% of the displacement amplifier was achieved. Note to Practitioners—This paper was motivated by the need to extend the actuation stroke of piezoelectric actuators, which are crucial in automation systems requiring high-precision displacement and large motion range. Traditional single-stage amplifiers are compact but provide limited extension. Multi-stage amplifiers offer greater extension, but their bulkiness limits the practical use. We have developed an innovative single-stage amplifier with very large amplification ratio that incorporates a unique arrangement of semi-bridge mechanisms. This design significantly extend the actuation stroke of piezoelectric actuators with a compact structure. The integration of a neural network-based PID controller improves the accuracy and efficiency of the positioning control of the amplifiers system. The principles and design approach we used could also be applied to fields where high precision and large displacement are needed, for example, micro/nano manufacturing and aerospace applications. This could open up new avenues for enhancing the efficiency and capability of devices in these sectors.

Piezoelectric actuators and giant magnetostrictive actuators are widely used in micropositioning and micromanipulation devices. To achieve sub-nanometer resolution with these actuators, a compact flexure-based displacement reducer, which shows the capability of obtaining a very large reduction ratio so as to achieve motion resolution of sub-nanometer, is proposed. It incorporates both bridge-type and lever-type mechanisms, arranged such that the reducer’s output equals the differential displacement between the two mechanisms. Additionally, a kinetostatic model for the reducer is developed. The parameters of the reducer are optimized to minimize the variation of the reduction ratio. The optimization results are validated by those of a finite element model, proving the effectiveness and correctness of the proposed reducer and the kinetostatic model. A prototype is fabricated and valuated by open-loop control and closed-loop control. The average reduction ratio can reach 206.4. By adopting a neural network-based H∞ robust controller, the reducer achieves satisfactory tracking performance. This comprehensive study affirms the reducer’s potential in enhancing the precision of micropositioning and micromanipulation devices.

A soft bistable actuator can generate high-speed motion between two prescribed stable positions, which is very useful for boosting the actuation of soft robots. Generally, the stroke of such an actuator is completely determined once the design is finalized, which prohibits its applications in robots that perform multiple tasks. In the current work, a bistable actuator with adjustable characteristics is proposed by exploring its strain energy landscape, in which the energy barrier is manipulatable via electroactive twisted and coiled polymer fibers. As such, the actuator can operate in either bistable or postbistable mode, both of which exhibit adjustable stiffness. A kinetostatic model that combines the chained beam constraint model and the mechanics of electroactive materials is established to characterize the actuator design. Experimental results validate the kinetostatic model and the behaviors of the actuator. As a robotic demonstration, a gripper that is formed by two actuators is prototyped, and it exhibits an adjustable load capacity (up to 6.5 times its weight under a 3 V voltage).

Multiple-view stereo has potential applications in robotic operations and autonomous driving (unstructured environment construction, visual servo). With assisted depth information, inertial navigation systems can achieve precise navigation. It is, especially suitable for GPS failures in complex environments. Accurate depth estimation is a challenge in low-textured or occluded regions. To alleviate the inference of incorrect depth, a multi-stage pixel-visibility learning-based stereo network is presented in this paper. Its improvements are as follows: 1) a new content-adaptive cost volume aggregation mechanism based on neighboring pixel-wise visibility is designed to effectively produce more accurate and smoother depth map predictions in the object boundary. 2) global convolution block and boundary refinement block are developed to regularize its cost volume, they can learn the inherent constraints of feature matching correspondence and effectively mitigate the depth estimation uncertainty in low-textured regions. 3) a new loss function is designed to measure the uncertainty of predicted probability distribution and enhance the reliability of depth map inference. Experimental results on the indoor DTU datasets and the outdoor Tanks & Temples datasets indicate that our method can achieve superior performance and has a powerful generalization ability, which is comparable to state-of-the-art works. Note to Practitioners—Multiple-view stereo (MVS) can estimate dense 3D representations of scenes, which is widely used in autonomous driving, robotic navigation, virtual reality (VR), and augmented reality (AR). Aiming at the problem of incorrect depth inference in low-textured or occluded regions, this work proposes a novel multi-stage depth prediction method based on neighboring pixel-wise visibility. Our method cannot only achieve accurate depth estimation for robot perception but also make no concession to real-time performance. It is clear that the proposed method has good potential in 3D reconstruction, robotic navigation, and VR/AR fields to provide accurate depth estimation in real-time with limited memory consumption.

Yuanhe Tian, Guimin Chen, Yan Song, Xiang Wan. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Aspect-level sentiment analysis (ASA) has received much attention in recent years.Most existing approaches tried to leverage syntactic information, such as the dependency parsing results of the input text, to improve sentiment analysis on different aspects.Although these approaches achieved satisfying results, their main focus is to leverage the dependency arcs among words where the dependency type information is omitted; and they model different dependencies equally where the noisy dependency results may hurt model performance.In this paper, we propose an approach to enhance aspect-level sentiment analysis with word dependencies, where the type information is modeled by key-value memory networks and different dependency results are selectively leveraged.Experimental results on five benchmark datasets demonstrate the effectiveness of our approach, where it outperforms baseline models on all datasets and achieves state-of-the-art performance on three of them. 1 * Equal contribution.

Aspect-based sentiment analysis (ABSA) predicts the sentiment polarity towards a particular aspect term in a sentence, which is an important task in real-world applications.To perform ABSA, the trained model is required to have a good understanding of the contextual information, especially the particular patterns that suggest the sentiment polarity.However, these patterns typically vary in different sentences, especially when the sentences come from different sources (domains), which makes ABSA still very challenging.Although combining labeled data across different sources (domains) is a promising solution to address the challenge, in practical applications, these labeled data are usually stored at different locations and might be inaccessible to each other due to privacy or legal concerns (e.g., the data are owned by different companies).To address this issue and make the best use of all labeled data, we propose a novel ABSA model with federated learning (FL) adopted to overcome the data isolation limitations and incorporate topic memory (TM) proposed to take the cases of data from diverse sources (domains) into consideration.Particularly, TM aims to identify different isolated data sources due to data inaccessibility by providing useful categorical information for localized predictions.Experimental results on a simulated environment for FL with three nodes demonstrate the effectiveness of our approach, where TM-FL outperforms different baselines including some well-designed FL frameworks. 1 * Equal contribution.

It is popular that neural graph-based models are applied in existing aspect-based sentiment analysis (ABSA) studies for utilizing word relations through dependency parses to facilitate the task with better semantic guidance for analyzing context and aspect words.However, most of these studies only leverage dependency relations without considering their dependency types, and are limited in lacking efficient mechanisms to distinguish the important relations as well as learn from different layers of graph based models.To address such limitations, in this paper, we propose an approach to explicitly utilize dependency types for ABSA with type-aware graph convolutional networks (T-GCN), where attention is used in T-GCN to distinguish different edges (relations) in the graph and attentive layer ensemble is proposed to comprehensively learn from different layers of T-GCN.The validity and effectiveness of our approach are demonstrated in the experimental results, where state-of-the-art performance is achieved on six English benchmark datasets.Further experiments are conducted to analyze the contributions of each component in our approach and illustrate how different layers in T-GCN help ABSA with quantitative and qualitative analysis.1

End-to-end aspect-based sentiment analysis (EASA) consists of two sub-tasks: the first extracts the aspect terms in a sentence and the second predicts the sentiment polarities for such terms.For EASA, compared to pipeline and multi-task approaches, joint aspect extraction and sentiment analysis provides a one-step solution to predict both aspect terms and their sentiment polarities through a single decoding process, which avoids the mismatches in between the results of aspect terms and sentiment polarities, as well as error propagation.Previous studies, especially recent ones, for this task focus on using powerful encoders (e.g., Bi-LSTM and BERT) to model contextual information from the input, with limited efforts paid to using advanced neural architectures (such as attentions and graph convolutional networks) or leveraging extra knowledge (such as syntactic information).To extend such efforts, in this paper, we propose directional graph convolutional networks (D-GCN) to jointly perform aspect extraction and sentiment analysis with encoding syntactic information, where dependency among words are integrated into our model to enhance its ability to represent input sentences and help EASA accordingly.Experimental results on three benchmark datasets demonstrate the effectiveness of our approach, where D-GCN achieves state-of-the-art performance on all datasets.1