检索结果-内蒙古大学图书馆

FAST-U2++: FAST AND ACCURATE END-TO-END SPEECH RECOGNITION IN JOINT CTC/ATTENTION FRAMES

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Liang, Chengdong Zhang, Xiao-Lei Zhang, BinBin Wu, Di Li, Shengqiang Song, Xingchen Peng, Zhendong Pan, Fuping School of Marine Science and Technology Northwestern Polytechnical University Xi’an China Horizon Robotics Beijing China WeNet Open Source Community China

Recently, the unified streaming and non-streaming two-pass (U2/U2++) end-to-end model for speech recognition has shown great performance in terms of streaming capability, accuracy and latency. In this paper, we present fast-U2++, an enhanced version of U2++ to further reduce partial latency. The core idea of fast-U2++ is to output partial results of the bottom layers in its encoder with a small chunk, while using a large chunk in the top layers of its encoder to compensate the performance degradation caused by the small chunk. Moreover, we use knowledge distillation method to reduce the token emission latency. We present extensive experiments on Aishell-1 dataset. Experiments and ablation studies show that compared to U2++, fast-U2++ reduces model latency from 320ms to 80ms, and achieves a character error rate (CER) of 5.06% with a streaming setup. Copyright © 2022, The Authors. All rights reserved.

关键词： Speech recognition

Mobile robot navigation using a sensor network

学校读者我要写书评

暂无评论

Mobile robot navigation using a sensor network

IEEE International Conference on robotics and Automation (ICRA)

作者： M.A. Batalin G.S. Sukhatme M. Hattig Robotic Embedded Systems Laboratory Center for Robotics and Embedded Systems Computer Science Department University of Southern California Los Angeles CA USA Open Source Robotics Intel Corporation Hillsboro OR USA

We describe an algorithm for robot navigation using a sensor network embedded in the environment. Sensor nodes act as signposts for the robot to follow, thus obviating the need for a map or localization on the part of the robot. Navigation directions are computed within the network (not on the robot) using value iteration. Using small low-power radios, the robot communicates with nodes in the network locally, and makes navigation decisions based on which node it is near. An algorithm based on processing of radio signal strength data was developed so the robot could successfully decide which node neighborhood it belonged to. Extensive experiments with a robot and a sensor network confirm the validity of the approach.

关键词： Mobile robots Robot sensing systems Orbital robotics Radio navigation Embedded system Sensor phenomena and characterization Computer networks Motion planning State-space methods Mobile communication

The open vision computer: An integrated sensing and compute system for mobile robots

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Quigley, Morgan Mohta, Kartik Shivakumar, Shreyas S. Watterson, Michael Mulgaonkar, Yash Arguedas, Mikael Sun, Ke Liu, Sikang Pfrommer, Bernd Kumar, Vijay Taylor, Camillo J. Open Source Robotics Foundation. Mountain ViewCA United States GRASP Lab University of Pennsylvania PhiladelphiaPA19104 United States

In this paper we describe the open Vision Computer (OVC) which was designed to support high speed, vision guided autonomous drone flight. In particular our aim was to develop a system that would be suitable for relatively small-scale flying platforms where size, weight, power consumption and computational performance were all important considerations. This manuscript describes the primary features of our OVC system and explains how they are used to support fully autonomous indoor and outdoor exploration and navigation operations on our Falcon 250 quadrotor platform. Copyright © 2018, The Authors. All rights reserved.

关键词： Computer vision

Analysis of grasping failures in multi-rigid body simulations

学校读者我要写书评

暂无评论

Analysis of grasping failures in multi-rigid body simulation...

IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)

作者： James R. Taylor Evan M. Drumwright John Hsu George Washington University Washington DC USA Toyota Research Institute Palo Alto CA USA Open Source Robotics Foundation Mountain View CA USA

ISBN: (纸本)9781509046171

Rigid body simulation libraries are sophisticated software systems that include multiple, tricky to implement numerical algorithms: solving initial value problems, root finding, geometric intersection (collision detection) and contact determination, and solving mathematical programming and optimization problems. How and why such systems fail to produce expected behavior is not readily known and not easy to discern; few of the components described above use reference implementations or are modular, for example, which makes attempting to identify points of failure challenging. Additionally, most libraries present an extraordinary number of parameters to be tuned, which complicates assessment. For the field of robotics, such failures are frustrating. One of the most seminal tasks in the domain, grasping of rigid objects, can require considerable parameter tuning. This paper uses a recent development, the support of four open-source physics engines in the GAZEBO simulator, to assess the ability of simulated robots to maintain grasps of rigid objects. We develop a metric that captures grasping performance and run a multitude of experiments to ascertain causes of failure. We have made all of our experimental code and data freely available, which allows others to reproduce our results and the authors of the corresponding physics engines to compete toward maximizing performance on the grasping task.

关键词： Robots Mathematical model Grasping Libraries Numerical models Iterative methods Collision avoidance

LIGHTGRAD: LIGHTWEIGHT DIFFUSION PROBABILISTIC MODEL FOR TEXT-TO-SPEECH

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Chen, Jie Song, Xingchen Peng, Zhendong Zhang, Binbin Pan, Fuping Wu, Zhiyong Shenzhen International Graduate School Tsinghua University Shenzhen China Horizon Robotics Beijing China WeNet Open Source Community China The Chinese University of Hong Kong Hong Kong

Recent advances in neural text-to-speech (TTS) models bring thousands of TTS applications into daily life, where models are deployed in cloud to provide services for customs. Among these models are diffusion probabilistic models (DPMs), which can be stably trained and are more parameter-efficient compared with other generative models. As transmitting data between customs and the cloud introduces high latency and the risk of exposing private data, deploying TTS models on edge devices is preferred. When implementing DPMs onto edge devices, there are two practical problems. First, current DPMs are not lightweight enough for resource-constrained devices. Second, DPMs require many denoising steps in inference, which increases latency. In this work, we present LightGrad, a lightweight DPM for TTS. LightGrad is equipped with a lightweight U-Net diffusion decoder and a training-free fast sampling technique, reducing both model parameters and inference latency. Streaming inference is also implemented in LightGrad to reduce latency further. Compared with Grad-TTS, LightGrad achieves 62.2% reduction in paramters, 65.7% reduction in latency, while preserving comparable speech quality on both Chinese Mandarin and English in 4 denoising steps1 © 2023, CC BY-NC-ND.

关键词： Diffusion

LightGrad: Lightweight Diffusion Probabilistic Model for Text-to-Speech

学校读者我要写书评

暂无评论

LightGrad: Lightweight Diffusion Probabilistic Model for Tex...

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Jie Chen Xingchen Song Zhendong Peng Binbin Zhang Fuping Pan Zhiyong Wu Shenzhen International Graduate School Tsinghua University Shenzhen China Horizon Robotics Beijing China WeNet Open Source Community The Chinese University of Hong Kong Hong Kong SAR China

关键词： Noise reduction Signal processing Probabilistic logic Data models Acoustics Decoding Speech processing

Multimodal Surface Sensing based on Hybrid Flexible Triboelectric and Piezoresistive Sensor

学校读者我要写书评

暂无评论

Multimodal Surface Sensing based on Hybrid Flexible Triboele...

2022 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2022 ACM International Symposium on Wearable Computers, UbiComp/ISWC 2022

作者： Lin, Zenan Lei, Kai Chong Mu, Shilong Song, Ziwu Dai, Yuan Ding, Wenbo Zhang, Xiao-Ping Tsinghua-Berkeley Shenzhen Institute Shenzhen International Graduate School Tsinghua University China Tencent Robotics X Lab China Tsinghua-Berkeley Shenzhen Institute Shenzhen International Graduate School Tsinghua University RISC-V International Open Source Laboratory China Shenzhen International Graduate School Tsinghua University Department of Electrical Computer and Biomedical Engineering Ryerson University China

ISBN: (纸本)9781450394239

Sensing the surface properties through touch, as the most natural perceptual way of humans, has become an important and practical method for human-machine interactions (HMI) and robot manipulations. In this paper, we design a fingertip hybrid flexible tactile sensor for multimodal surface sensing, based on the triboelectric and piezoresistive mechanisms. A real-time tactile sensing system is implemented on a 3D-printed robot finger together with a wireless data acquisition board. A virtual data generation method is proposed to expand the model adaptability under different compression force levels. Moreover, considering the characteristics of data generated by our sensors, a novel deep learning model with a residual structure is developed, named parallel residual convolutional neural network (PR-CNN). Our model outperforms the state-of-the-art models, i.e., Res-CNN, LSTM-FCN and InceptionTime, with over 96% accuracy, on three classification tasks, including textures (13 types), materials (10 types), and combinations of textures and materials (18 types). The proposed system has broad applications in service robots, industrial sorting robots, and HMI. © 2022 Owner/Author.

关键词： Textures

WEKWS: A PRODUCTION FIRST SMALL-FOOTPRINT END-TO-END KEYWORD SPOTTING TOOLKIT

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wang, Jie Xu, Menglong Hou, Jingyong Zhang, Binbin Zhang, Xiao-Lei Xie, Lei Pan, Fuping School of Marine Science and Technology Northwestern Polytechnical University Xi'An China School of Computer Science Northwestern Polytechnical University Xi'An China Horizon Robotics Beijing China WeNet Open Source Community China

Keyword spotting (KWS) enables speech-based user interaction and gradually becomes an indispensable component of smart devices. Recently, end-to-end (E2E) methods have become the most popular approach for on-device KWS tasks. However, there is still a gap between the research and deployment of E2E KWS methods. In this paper, we introduce WeKws, a production-quality, easy-to-build, and convenient-to-be-applied E2E KWS toolkit. WeKws contains the implementations of several state-of-the-art backbone networks, making it achieve highly competitive results on three publicly available datasets. To make WeKws a pure E2E toolkit, we utilize a refined max-pooling loss to make the model learn the ending position of the keyword by itself, which significantly simplifies the training pipeline and makes WeKws very efficient to be applied in real-world scenarios. The toolkit is publicly available at https://***/wenet-e2e/wekws. Copyright © 2022, The Authors. All rights reserved.

关键词： Machine learning

Using parallel stiffness to achieve improved locomotive efficiency with the Sandia STEPPR robot

学校读者我要写书评

暂无评论

Using parallel stiffness to achieve improved locomotive effi...

IEEE International Conference on robotics and Automation (ICRA)

作者： Anirban Mazumdar Steven Spencer Jonathan Salton Clinton Hobart Joshua Love Kevin Dullea Michael Kuehl Timothy Blada Morgan Quigley Jesper Smith Sylvain Bertrand Tingfan Wu Jerry Pratt Stephen Buerger Intelligent Systems Sandia National Laboratories Albuquerque NM USA Open Source Robotics Foundation Mountain View CA USA BioInspired Robotic Lab at the Institute for Human and Machine Cognition (IHMC) Pensacola FL USA

In this paper we introduce STEPPR (Sandia Transmission-Efficient Prototype Promoting Research), a bipedal robot designed to explore efficient bipedal walking. The initial iteration of this robot achieves efficient motions through powerful electromagnetic actuators and highly back-drivable synthetic rope transmissions. We show how the addition of parallel elastic elements at select joints is predicted to provide substantial energetic benefits: reducing cost of transport by 30 to 50 percent. Two joints in particular, hip roll and ankle pitch, reduce dissipated power over three very different gait types: human walking, human-like robot walking, and crouched robot walking. Joint springs based on this analysis are tested and validated experimentally. Finally, this paper concludes with the design of two unique parallel spring mechanisms to be added to the current STEPPR robot in order to provide improved locomotive efficiency.

关键词： Torque Springs Joints Legged locomotion Hip Knee