检索结果-内蒙古大学图书馆

Towards real-time and energy efficient Siamese tracking - a hardware-software approach

学校读者我要写书评

暂无评论

TechRxiv 2022年

Siamese trackers have been among the state-of-the-art solutions in each Visual Object Tracking (VOT) challenge over the past few years. However, with great accuracy comes great computational complexity: to achieve real-time processing, these trackers have to be massively parallelised and are usually run on high-end GPUs. Easy to implement, this approach is energy consuming, and thus cannot be used in many low-power applications. To overcome this, one can use energy-efficient embedded devices, such as heterogeneous platforms joining the ARM processor system with programmable logic (FPGA). In this work, we propose a hardware-software implementation of the well-known fully connected Siamese tracker (SiamFC). We have developed a quantised Siamese network for the FINN accelerator, using algorithm-accelerator co-design, and performed design space exploration to achieve the best efficiency-to-energy ratio (determined by FPS and used resources). For our network, running in the programmable logic part of the Zynq UltraScale+ MPSoC ZCU104, we achieved the processing of almost 50 frames-per-second with tracker accuracy on par with its floating point counterpart, as well as the original SiamFC network. The complete tracking system, implemented in ARM with the network accelerated on FPGA, achieves up to 17 fps. These results bring us towards bridging the gap between the highly accurate but energy-demanding algorithms and energy-efficient solutions ready to be used in low-power, edge systems. © 2022, CC BY.

关键词： Field programmable gate arrays (FPGA)

Energy Efficient Hardware Acceleration of Neural Networks with Power-of-Two Quantisation

学校读者我要写书评

暂无评论

TechRxiv 2022年

关键词： System-on-chip

Playing cards and bidding calls detection for automatic registration of a duplicate bridge game

学校读者我要写书评

暂无评论

TechRxiv 2022年

作者： Wzorek, Piotr Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Krakow Poland

In this work, the implementation of a playing cards and bidding calls detection system for the automatic registration of a duplicate bridge game is presented. For this purpose, two YOLOv4 deep convolutional neural networks were used. During the training process, which was carried out using an automatically generated dataset, we obtained a detector characterised by more than 99.9% accuracy on the test set. The prepared system has been evaluated on videos collected during bridge competitions and is characterised by high accuracy. The application has been implemented and evaluated on a CPU and an embedded GPU. © 2022, CC BY.

关键词： Deep neural networks

Towards real-time and energy efficient Siamese tracking - a hardware-software approach

学校读者我要写书评

暂无评论

arXiv

arXiv 2022年

关键词： Energy efficiency

Traffic Sign Classification Using Deep and Quantum Neural Networks

学校读者我要写书评

暂无评论

arXiv

arXiv 2022年

作者： Kuros, Sylwia Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Kraków Poland

Quantum Neural Networks (QNNs) are an emerging technology that can be used in many applications including computer vision. In this paper, we presented a traffic sign classification system implemented using a hybrid quantum-classical convolutional neural network. Experiments on the German Traffic Sign Recognition Benchmark dataset indicate that currently QNN do not outperform classical DCNN (Deep Convolutuional Neural Networks), yet still provide an accuracy of over 90% and are a definitely promising solution for advanced computer vision. © 2022, CC BY.

关键词： Traffic signs

Traffic Sign Classification Using Deep and Quantum Neural Networks

学校读者我要写书评

暂无评论

TechRxiv 2022年

作者： Kuros, Sylwia Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Kraków Poland

关键词： Traffic signs

Sensorless force estimation on fingertip with gravitational compensation for heavy-duty pneumatic tri-grasper robot

学校读者我要写书评

暂无评论

Sensorless force estimation on fingertip with gravitational ...

Engineering Technology International Conference (ETIC 2022)

作者： A. Irawan M. I. Putra Azahar M. A. Hashimi Robotics Intelligent Systems & Control Engineering (RISC) Research Group Faculty of Electrical & Electronics Engineering Technology Universiti Malaysia Pahang Pahang Malaysia

The paper presents the proposed sensorless force estimator design for pneumatic robot fingertip by using gravitational compensation and pressure changed in pneumatic cylinder piston. The approach is done to replace the commercial force sensor that may be expensive for heavy-duty configuration. The formulation was done by considering the torque of robot's finger joint, finger dimension as well as its actuator and the different pressures in cylinder piston. The gravitational force is calculated from the geometry of the robot's finger as dynamic gain for the force of pneumatic cylinder. The proposed method is validated on a heay-duty pneumatic Tri-grasper Robot with the simple basic movement and blocked randomly by human barehand. The results show that the force output by the estimator is almost identical to the loadcell sensor that attached on the fingertip at about 2% error in average. The sensitivity is a bit low for small and fragile material but enough for heavy-duty application that generally with hard and rough surfaces.

关键词：

Learning Stable and Robust Linear Parameter-Varying State-Space Models

学校读者我要写书评

暂无评论

Learning Stable and Robust Linear Parameter-Varying State-Sp...

IEEE Conference on Decision and control

作者： Chris Verhoek Ruigang Wang Roland Tóth Control Systems Group Eindhoven University of Technology The Netherlands Australian Centre for Robotics School of Aerospace Mechanical and Mechatronic Engineering The University of Sydney Australia Vehicle Industry Research Center Szechenyi Istvan University Hungary

This paper presents two direct parameterizations of stable and robust linear parameter-varying state-space (LPV-SS) models. The model parametrizations guarantee a priori that for all parameter values during training, the allowed models are stable in the contraction sense or have their Lipschitz constant bounded by a user-defined value $\gamma$ . Furthermore, since the parametrizations are direct, the models can be trained using unconstrained optimization. The fact that the trained models are of the LPV-SS class makes them useful for, e.g., further convex analysis or controller design. The effectiveness of the approach is demonstrated on an LPV identification problem.

关键词：

Robust control of Interleaved Boost Converter for Open-Cathode PEM Fuel Cell systems

学校读者我要写书评

暂无评论

Robust Control of Interleaved Boost Converter for Open-Catho...

IEEE International Conference on Automatic control and Intelligent systems (I2CACIS)

作者： Jie Ying Tan R.M.T Raja Ismail Institute of Postgraduate Studies Universiti Malaysia Pahang Pekan Pahang Malaysia Robotics Intelligent Systems & Control Engineering (RISC) Research Group Faculty of Electrical & Electronics Universiti Malaysia Pahang Pekan Pahang Malaysia

This paper implements a thermal control and a super-twisting sliding mode (STSM) incorporated with an interleaved boost converter (IBC) for an open-cathode proton exchange membrane fuel cell (OC-PEMFC). The implementation of the thermal control is to regulate the stack temperature and adjust the air stoichiometric during the fuel cell current variation. The STSM controller is designed to ensure the fuel cell system’s robustness by achieving the reference values set. Therefore, simulation results discussed that the thermal control provides the suitable stack temperature to the system and maintains the fan voltage during temperature variations. A PI controller is designed and used as a comparison to the proposed STSM controller. Hence, the proposed STSM demonstrates its effectiveness in tracking down the reference current values for the fuel cell system.

关键词：