Search

This article presents two novel neural network models based on recurrent neural network (RNN) for radio frequency power amplifiers (RF PAs): instant gated recurrent neural network (IGRNN) model and instant gated implict recurrent neural network (IGIRNN) model. In IGRNN model, two state control units are introduced to ensure the linear transmission of hidden state and solve the problem of vanishing gradients of RNN model. In contrast with conventional RNN model, IGRNN can better describe the long-term memory effect of power amplifier, more in line with the physical distortion characteristics of power amplifier. Furthermore the instantaneous gates are used to express the input information implicitly to reduce the redundancy of the input information, and a simpler IGIRNN model is proposed. The complexity analysis indicates that the proposed models have significantly lower complexity than other RNN-based variant structures. A wideband Doherty RF PA excited by 100MHz and 120MHz OFDM signals was employed to evaluate the performance. Extensive experimental results reveal that the proposed IGRNN and IGIRNN models can achieve better linearization performance compared with RNN model and traditional GMP model, and have comparable performance with lower computational complexity compared with the state-of-the-art RNN-based variant models, such as gated recurrent unit (GRU) model.

Data assimilation is a useful approach to obtain the most accurate estimate of the initial and boundary conditions for Numerical Weather Prediction (NWP). This study is to investigate data assimilation of Automatic Weather Station (AWS) data with Weather Research and Forecasting (WRF) Model for heavy rainfall forecasting in Thailand. Two heavy rainfall events occurred in the Northeast of Thailand, due to two tropical storms (i.e., SONCA; 24th_29th July, 2017 and PODUL; 28th_30th August, 2019) were simulated. Three assimilation experiments for each event were designed using two different AWS data sources provided by the Hydro-Informatics Institute (HII) and the Thai Meteorological Department (TMD) together with a combination of the AWS data sources. The results shown that the WRF model assimilated with the AWS observation simulates the extreme rainfall events more accurately than that of non-data assimilation of heavy rainfall in local areas. The WRF model shows better performance when assimilated with a combined AWS dataset with a positive bias score of less than 0.3 and POD larger than 0.6 in both heavy rainfall events.

In the current social development, many new types of intelligent science and technology such as big data are rapidly becoming generalized, and now various companies and enterprises are gradually adopting digital operations, and the application of big data technology to digital marketing has become The shortcut in the development of modern enterprises has become the consensus in the development of enterprises, and it is becoming more and more popular in the development of enterprises. The new digital marketing model based on big data technology takes big data technology as the core foundation of digital marketing. The rapid development of digital marketing models has not only made a huge contribution to the rapid development of social economy, but also produced various problems in the development of new digital marketing. The most common one is that the marketing data and models are not accurate enough. Only by establishing direct communication channels with consumers and accurately positioning consumers' needs for corporate products and services can new enterprises find out the marketing purpose, and then use various big data technologies and smart terminal technologies to accurately push marketing information to public. It not only saves the cost of corporate marketing, but also avoids the unnecessary influence that is unintentionally passed on to consumers who do not need marketing products, thereby maximizing the effect of digital marketing. The previous marketing concepts can no longer meet the needs of today's era of great changes, and new social forms require brand-new marketing theories.

With the development of my country's economy and society, the development of cross-border foreign trade companies has gradually attracted people's attention. In order to make foreign trade companies grow, they must implement e-commerce strategies reasonably according to the rare opportunities and challenges in the Internet era, so as to conform to the status quo of trade development. To this end, this article aims to discuss China's cross-border e-commerce strategy to promote foreign trade transformation and modernization from the perspective of industrial integration, as well as the current challenges of my country's foreign trade enterprise transformation. In the internet age, this paper examines the economics of cross-border e-commerce in the context of the trend of industrial integration in the internet age, uses appropriately realistic assumptions and theories, and analyzes the development of foreign trade patterns in recent years. Years old. Consider it a specific research topic and bring new knowledge in e-commerce foreign trade field. These new ideas will help change and improve the foreign trade system. Experimental results show that China's cross-border e-commerce is positively impacting the transformation and modernization of foreign trade.

Quadrotors are making a difference to numerous civilian and military applications. There are various approaches to realize formation flight control. In this paper, a comparison between sliding mode and second-order consensus approaches for quadrotor leader-follower formation flight was conducted. Sliding mode and second-order consensus algorithms were employed to design the formation controller for steady formation flight. Both leader and the follower quadrotors are controlled by proportional integral derivative controllers. The leader follows the desired trajectory while the followers follow the instruction calculated by the formation controller. Simulations are carried out to analyze the effect of these two approaches. The result indicates that the followers can follow the leader's trajectory and keep the formation fairy well under the control of both approaches. However, the sliding mode approach can achieve faster position convergence effect, while the second-order consistency approach has better disturbance suppression performance in case of disturbance.

Compressive sensing (CS) theory has achieved signifificant success in the field of synthetic aperture radar (SAR) imaging. Recent studies have shown that SAR imaging for sparse scene can also be successfully performed with one-bit quantized data. Many algorithms were proposed for one-bit SAR imaging. These algorithms can recover data only if the sensing matrix is accurate. But when the SAR system contains phase error (i.e., the sensing matrix has error), the SAR image recovered by these classical algorithms will defocus. In this work, We introduce a novel framework where phaseless transformation is adopted to realize autofocus in one-bit SAR imaging. The proposed method has excellent focusing ability and reduces hardware cost appropriately. The simulation results of the point targets and extended targets demonstrate the effectiveness of our proposed method.

In this paper, we present a new method for the linearization of radio frequency (RF) power amplifies (PAs) based on the sample selection method (SSM) with incremental dimension of coefficients. In SSM, insufficient training samples can cause some problems such as the low convergence speed and serious noise interference. In order to solve these problems, we propose to use coefficient vectors with increasing dimension in iterations and then to introduce the dimension reduction (or pruning) technique to ensure the priority selections of important coefficients in training. Experimental results show that the proposed method can achieve better normalized mean square error (NMSE) and adjacent channel power ratio (ACPR) with faster convergence speed compared with SSM.

The use of GNSS to determine the orthometric height requires a geoid model which is used as a transformation linkage between ellipsoidal height and orthometric height. GNSS derived ellipsoidal height is purely mathematical, having no physical meaning. So, in most of the geodetic applications, it is necessary to use physically defined orthometric height. Orthometric height can be obtained accurately from traditional spirit levelling which is labor-intensive, tedious and costly. Conversion of GNSS-derived ellipsoidal height into orthometric height is the popular surveying technique in GNSS user community. This approach is required to integrate a geoid model which provides separation of geoid with reference ellipsoid. The EGM2008, a global geoid model is widely employed for the purpose yielding sub-meter accuracy. This paper evaluated the orthometric height derived from GNSS measurements and Global Geo-potential Model, EGM2008 in Bangladesh. GNSS Survey as well as spirit levelling was carried out connecting fifteen points distributed in the area. GNSS observations were taken using dual frequency geodetic receivers in differential static mode and levelling works were carried out using digital levels. The geoidal undulation for the corresponding points were obtained using the EGM2008 Geoid model. The orthometric height of these points are computed using ellipsoidal height and the geoid undulation. These heights are compared with the orthometric heights obtained from levelling. The difference in these two, was used to compute the root mean square error, mean and standard deviation which are found 17.57 cm, 17.18 cm and 3.76 cm, respectively. The results show that EGM2008 model is appropriate with root mean square errors. It is proven that orthometric height can be obtained at centimeter level accuracy by this method for any Engineering, Surveying or Mapping works where rapid heights are required.

In this article, from the perspective of rice processing loss, this paper proposes a method to evaluate the processing loss of rice produced by enterprises using data analysis. The specific content of this method is to build a fitting simulation model of unit power consumption and rice production rate based on production data of various types of enterprises across the country at different scales, equipment, and technological processes, thereby establishing a rice processing loss evaluation system; Based on the production data of the enterprise to be evaluated, an enterprise data fitting optimization model of unit power consumption and rice production rate was constructed; and the results of evaluation and diagnosis were combined with processing losses of rice processing enterprises. The application process of this evaluation method found the change law of unit power consumption and rice production rate in different dimensions, and gave evaluation conclusions and optimization plans for the evaluation of the production status of the enterprise. The application of this evaluation method to the grain industry can be a rice processing enterprise. Provide practical guidance on energy saving and consumption reduction during processing.

In this paper, a new compact H-plane dual-band waveguide band-pass filter is proposed. The filter is composed of one thick dielectric substrate with copper on both sides. Three complementary split ring resonators (CSRRs) and three asymmetric resonator slots are etched on copper layers. The asymmetric resonator slots on the single copper layer can generate two transmission zeros to form a wide pass-band. The CSRRs aim to introduce another transmission zero in the wide pass-band and strengthen the out-of-band suppression. Because of the short distance, two copper layers interact with each other, strengthening the flatness of pass-band and enhancing the selectivity of the filter. The isolation between two pass-bands has also been greatly improved. The proposed filter has a dual-band band-pass response and operates at 8.69/11.63GHz. Insertion loss of each frequency is at around 0.5 dB and the 3dB bandwidth is 0.48/0.54 GHz, respectively. The prototype is measured with fairly good performance, but the insertion loss is a little higher. Also, the total length of this filter is only 3.245mm, which is much smaller than similar filters.

The compressive sensing framework aims to stably recover sparse signals at reduced sampling rate. Recent research has shown that only the sign information of the measurements can achieve accurate signal recovery. Most work focused on the real-valued signals. In many applications, however, the signals are complex-valued. Also, the existing reconstruction algorithms always involve large matrix-vector multiplications which make them more time and memory consuming than traditional matched filtering-based methods. In this paper, we propose a 1-bit compressive sensing reconstruction algorithm for recovering complex sparse signals using approximated observation matrix which is deduced from the inverse of chirp-scaling method. This algorithm directly operates on complex data and greatly reduce the computation time and memory usage. Both simulated and real data validate the efficiency of the proposed algorithm.

In recent years, one-bit compressed sensing (CS) has been applied to the field of synthetic aperture radar (SAR) imaging and shows great potential. The existing models are, based on application of the sensing matrix acquired by exact observation functions. As a result, the corresponding reconstruction algorithms are much more time consuming than traditional matched filter (MF)-based focusing methods, especially in high resolution and wide swath systems. This paper presents a novel one-bit compressed sensing approach for SAR imaging using weighted L1-Norm Minimization based on approximated observation.It adopts the approximated SAR observation model deduced from the inverse of MF-based methods.It can be applied to high-quality and high-resolution imaging under sub-Nyquist rate sampling, while decreasing the computational cost substantially both in memory and time, resulting in a fast approach.Simulations results show that our method can perform sparse SAR imaging effectively with one-bit quantized data for large scale applications.