The significant advancements in research within the domain of Advanced Driver-Assistance systems (ADAS) have been significantly propelled by the integration of Artificial Intelligence (AI) and Deep Learning (DL). The ...
详细信息
ISBN:
(数字)9798350384819
ISBN:
(纸本)9798350384826
The significant advancements in research within the domain of Advanced Driver-Assistance systems (ADAS) have been significantly propelled by the integration of Artificial Intelligence (AI) and Deep Learning (DL). The detection and classification of traffic signals are pivotal for ensuring driver safety and comfort, prompting the development of numerous detection and classification methodologies in recent times. Convolutional Neural Networks (CNNs) and field Programmable Gate Arrays (FPGAs) have individually demonstrated notable efficacy in signal processing, image detection, and image classification tasks. However, the amalgamation of CNNs and FPGAs for traffic signal classification offers enhanced flexibility, accuracy, and resource efficiency. This study proposes a software-hardware co-design approach for the development of a resource-efficient real-time traffic signal classification system tailored for FPGAs utilizing the CNN algorithm. The software implementation has been carried out within the TensorFlow framework, while the hardware implementation is executed via the Xilinx Vivado 2023.1 platform and rigorously evaluated using the Xilinx ZCU102 FPGA board. Experimental findings demonstrate that the traffic signal classification accuracy can attain approximately 99.98% at a frame rate of 84139 frames per second (FPS) on the FPGA platform, with a power consumption of 4.4W. The exceptionally low latency and heightened accuracy exhibited by the proposed system render it highly advantageous for ADAS and other safety-critical applications.
The focus of this paper is on the development of Sband radar using the Lyrtech Small Form Factor (SFF) Software Defined Radio (SDR) for generating the Chirp-Z signal. The Lyrtech SDR platform has a high speed AD/DA co...
详细信息
The focus of this paper is on the development of Sband radar using the Lyrtech Small Form Factor (SFF) Software Defined Radio (SDR) for generating the Chirp-Z signal. The Lyrtech SDR platform has a high speed AD/DA conversion board and a tunable RF board, which enable a flexible radar design. To achieve a functional radar system, the development kids of the SFF SDR have to be fully utilized. This paper introduces the usage of both software and hardware resources of the SDR for converting the SDR into a waveform-programmable radar, which can benefit the study of radar systems using generic SDR platforms. Based on our approach, we develop an S-band radar which is used to transmit a Chirp-Z radar waveform modulated on a 2.2GHz carrier. The structure and the development kits of the SFF platform are introduced, and the development of the S-band radar by using the development kits discussed. Finally, both transmitted and received Chirp-Z signals using the developed radar system are demonstrated.
The upper mid-band (or FR3, spanning 6–24 GHz) is a crucial frequency range for next-generation mobile networks, offering a favorable balance between coverage and spectrum efficiency. From another perspective, the sy...
详细信息
ISBN:
(数字)9798350354058
ISBN:
(纸本)9798350354065
The upper mid-band (or FR3, spanning 6–24 GHz) is a crucial frequency range for next-generation mobile networks, offering a favorable balance between coverage and spectrum efficiency. From another perspective, the systems operating in the near-field in both indoor environment and outdoor environments can support line-of-sight multiple input multiple output (MIMO) communications and be beneficial from the FR3 bands. In this paper, a novel method is proposed to measure the near-field parameters leveraging a recently developed reflection model where the near-field paths can be described by their image points. We show that these image points can be accurately estimated via triangulation from multiple measurements with a small number of antennas in each measurement, thus affording a low-cost procedure for near-field multi-path parameter extraction. A preliminary experimental apparatus is presented comprising 2 transmit and 2 receive antennas mounted on a linear track to measure the 2 × 2 MIMO channel at various displacements. The system uses a recently-developed wideband radio frequency (RF) transceiver board with fast frequency switching, an FPGA for fast baseband processing, and a new parameter extraction method to recover paths and spherical characteristics from the multiple 2 × 2 measurements.
Radio Tomographic Imaging (RTI) seeks to reconstruct useful images of the environment through measurements of received signal strength (RSS) in a network of radios. RSS data is compiled over many transmissions, and af...
详细信息
Radio Tomographic Imaging (RTI) seeks to reconstruct useful images of the environment through measurements of received signal strength (RSS) in a network of radios. RSS data is compiled over many transmissions, and after processing the data an estimated image of the spatial loss field (SLF) of the area can be recovered. This image represents the estimated radio frequency signal attenuation in dB that occurs at each pixel location. An ideal RTI system could form images of floor plans and objects that are located inside buildings or behind other obstructions [1]. One of the challenges in RTI is the presence of multipath signals. Such signals reach the receiver at some delay with respect to the line-of-sight (LOS) signal. Presently, there has been no attempt to deal with such signals other than to reject them, at best. Thus, potentially useful links in a network of radios cannot be utilized, because their LOS signal path does not pass through the area to be imaged. This paper presents a method by which multipath signals may be treated as useful measurements in an RTI system. Doing so informs the existing SLF image, improving image quality and reducing root-mean-squared-error (RMSE), which is the focus of this paper. It is also possible to use such measurements to create an estimate of the area's reflectivity loss field, or RLF. Such an estimated image represents the reflection coefficient (in dB) seen at each reflective pixel location. This image could be useful in identifying interior features that are reflective, such as large metal objects or metal walls. Thus, two separate images are estimated simultaneously, an SLF image of pixel attenuation, and an RLF image of pixel reflectivity values. Both estimated images remain linear functions of the received data. We refer to this new model as the Augmented Multipath Linear generating Model, or AMLGM. After presenting the AMLGM model, we discuss the results of our simulations showing improvements in both subjective and obj
The use of video camera surveillance systems has dramatically increased over recent years. This is due in no small part to the reduction in costs and increased reliability afforded by CCD camera technology. However th...
The use of video camera surveillance systems has dramatically increased over recent years. This is due in no small part to the reduction in costs and increased reliability afforded by CCD camera technology. However the capability of designers and installers has not kept pace with the hardware development. With end users demanding ever reducing costs many systems are designed and installed by electrical engineers who have no training in broadband signal communications. This has led to systems being installed with less than adequate performance or where performance has degraded a year or so down the line. The objective of VIDEQUAL is to provide a fully quantitative measurement of system performance. This is achieved by placing or defining targets in the camera field of view and/or injecting test signals at the camera. Measurements are then made either after the viewing monitor or directly from the output of the video switcher. It is intended that the measuring equipment employs standard PC architecture using a commercial video signal capture board to acquire an image for processing. Dependant on the application this image could be stored for future reference or audit purposes. The measurements are chosen to best reflect the system performance and should be simple to make and be easy to understand. The output from the equipment is in the form of either a screen display or a printed report. Where the equipment is undertaking an automatic comparison an audible/digital alarm would be provided
In this study, virtual global navigation satellite system data is created to be used when global navigation satellite system (GNSS) data of unmanned aerial vehicles (UAV) is jammed. Nowadays, when GNSS signals disappe...
详细信息
ISBN:
(数字)9781728172064
ISBN:
(纸本)9781728172071
In this study, virtual global navigation satellite system data is created to be used when global navigation satellite system (GNSS) data of unmanned aerial vehicles (UAV) is jammed. Nowadays, when GNSS signals disappear, the most common method used as aided navigation data is to obtain location information by using inertial measurement units based on MEMS (micro-electro-mechanical) system. However, the position error generated by inertial units increases in proportion to the square of time. In this study, the position information of the UAV is determined by calculating the slant range, elevation and azimuth angles between the electro optic ground station and the UAV. In this structure, 3 different sensors are used. The laser distance meter is used to measure the distance between the ground station and the UAV, the magnetic field meter is used to measure the angle of the azimuth and the accelerometer is used to measure the elevation angle. The sensor data obtained from the STEVAL-STLKT01V1 development board were processed and the effect of the measurement errors on the position error was analyzed. This study provides methods for aided navigation system developers and sheds light on the tradeoffs by giving effect to the position error of sensor measurement errors. It is thought that the study will help to establish an aided navigation system for UAVs that may work in harsh environmental and atmospheric conditions in the future.
WiFi-based indoor positioning has been extensively studied. A fundamental issue in such solutions is the collection of WiFi fingerprints. However, due to real-world constraints, collecting complete fingerprints at all...
详细信息
ISBN:
(数字)9798350384611
ISBN:
(纸本)9798350348439
WiFi-based indoor positioning has been extensively studied. A fundamental issue in such solutions is the collection of WiFi fingerprints. However, due to real-world constraints, collecting complete fingerprints at all intended locations is sometimes prohibited. This work considers the WiFi fingerprint inpainting problem. This problem differs from typical image/video inpainting problems in several aspects. Unlike RGB images, WiFi field maps come in any shape, and signal data may follow certain distributions. Therefore, it is difficult to forcefully fit them into a fixed-dimensional matrix, as done with processing images in RGB format. As soon as a map is changed, it also becomes difficult to adapt it to the same model due to scale issues. Furthermore, such models are significantly constrained in situations requiring outward inpainting. Fortunately, the spatial relationships of WiFi signals and the rich information provided among channels offer ample opportunities for this generative model to accomplish inpainting. Therefore, we designed this model to not only retain the characteristic of regression models in generating fingerprints of arbitrary shapes but also to accommodate the observational outcomes from densely deployed APs. This work makes two major contributions. Firstly, we delineate the distinctions between this problem and image inpainting, highlighting potential avenues for research. Secondly, we introduce novel generative inpainting models aimed at capturing both inter-AP and intra-AP correlations while preserving latent information. Additionally, we incorporate a specially designed adversarial discriminator to enhance the quality of inpainting outcomes.
暂无评论