In light of the worldwide energy crisis, this research investigates methods for reducing the power consumption of digital circuits, focusing specifically on multiplexer design using FPGA families. In order to lower po...
详细信息
Convolution Neural Networks (CNNs) are pivotal in image processing. Prominent CNN models include SqueezeNet, MobileNet, Xception, VGG-16, and AlexNet. These models transform 2D data to 1D, emphasizing maximum intensit...
详细信息
A fieldprogrammable gate array (FPGA) is a type of programmable logic device that the consumer can modify after production to carry out a variety of tasks, from fundamental logic gate operations to AI systems and bey...
详细信息
Nothing is perfect. Even carefully developed hardware components can sometimes exhibit unexpected behaviour. This is even more likely in harsh environments, like the one satellites are exposed to. However, on satellit...
详细信息
ISBN:
(纸本)9798350302233
Nothing is perfect. Even carefully developed hardware components can sometimes exhibit unexpected behaviour. This is even more likely in harsh environments, like the one satellites are exposed to. However, on satellites hardware analysis is limited to the data obtained by a constrained set of integrated sensors. Therefore, only expected failure cases can be traced, or only indirect monitoring is possible. In this paper, we present an adaptable way to analyse hardware in the field, to solve this problem. Our approach was implemented in-flight on ESA's OPS-SAT satellite, where such an unexplainable fault was noticed. Fortunately, OPS-SAT contains a reconfigurable field-programmable Gate Array (FPGA) allowing a logic analyser functionality to be implemented. Thereby, any signal present in the FPGA fabric becomes traceable, including all externally connected ones. Here, we describe the implementation, its deployment and the successful execution on the OPS-SAT satellite. Using the captured traces, we performed an in-depth analysis of the erroneous behaviour. Our understanding is, that this is the first time such a technique has been implemented on a flying spacecraft. We hope to encourage in-field adoption, especially in-space FPGA reconfiguration, to drive future innovation.
This paper introduces the two-channel digital image processing technology. The system uses FPGA as its core processing unit. Infrared thermal imager and CCD camera were used for the shooting. The FPGA composed of prog...
详细信息
Reliability and energy efficiency are critical challenges in the design of modern Network-on-Chip (NoC) architectures, where the efficient mapping of computational tasks to processing cores plays a pivotal role in ove...
详细信息
This paper introduces an effective online compensation mechanism for time-to- digital converters (TDCs) implemented on field-programmable Gate Arrays (FPGAs), which are highly sensitive to process, voltage, and temper...
详细信息
ISBN:
(纸本)9798350361056;9798350361049
This paper introduces an effective online compensation mechanism for time-to- digital converters (TDCs) implemented on field-programmable Gate Arrays (FPGAs), which are highly sensitive to process, voltage, and temperature (PVT) variations. The online compensation strategy exploits the synergy between an external high-precision clock signal and the internal chip clock for real-time correction of TDC outputs. Our TDC architecture incorporates a coarse-fine counting scheme, complemented by a start-stop calibration module and a ring oscillator for precise time compensation. The proposed TDC system has been successfully implemented and tested on a Xilinx Artix-7 35T FPGA board. Measurement results demonstrate that our system achieves a minimum time resolution of 18.9 picoseconds (ps) and maintains a root mean square (RMS) resolution within 20 ps, even under temperature variations.
This paper presents the implementation of a 28 nm field-programmable gate array (FPGA)-based time-to-digital converter (TDC), which has both high resolution and low-cost characteristics. Due to the simple TDC structur...
详细信息
In this paper, different hardware deployment schemes are designed to accelerate the overall reasoning process of convolutional neural network according to the characteristics of large amount of calculation in convolut...
详细信息
暂无评论