Search

Recently, the combination of convolutional neural network (CNN) and long short-term memory (LSTM) exhibits better performance than single network architecture. Most of these studies connect LSTM networks behind CNNs. When operating on hardware, the current design of CNN-LSTM is similar to a pipeline architecture. However, the classic structure lead to a feature loss when data is sent to LSTM since CNN is not good at extracting temporal features. At the same time, as the depth and scale increases, it will bring a huge amount of computation, which makes hardware implementation difficult. Based on that, a parallel CNN-LSTM architecture is proposed, in which two networks extract features from the input data synchronously, being proven to be more effective than classical CNN-LSTM. This paper designs a parallel CNN-LSTM computing device based on FPGA. The device is divided into control unit and operation unit. Control stream and data stream transport between the two units, ensuring the proper running of the device. A highly parallel multi-channel convolution layer and pooling layer are designed to improve the calculation efficiency. A 4-stage pipeline structure is adopted to implement the LSTM part. This paper makes full use of on-chip BRAM to design a look-up table for activation function approximation, reducing the resource consumption by 95% compared with the traditional polynomial approximation. Finally, we verify our device under cooperative spectrum sensing (CSS) and handwritten classification scenarios. Proposed device reaches higher accuracy in two scenarios compared with classic CNN-LSTM structure as well as faster calculating speed, and the overall project power is limited below 2W. The scalability and limitation of this computing device are also discussed.

This paper discusses the regularity of the GP-V’-rings in conjunction with idempotent reflexivity for the first time. We mainly discuss the weak and strong regularity of the GP-V’-rings using generalized weak ideals, weakly right ideals, and quasi-ideals. We show the following: (1) If R is an idempotent reflexive semi-abelian left GP-V’-ring whose every maximal essential left ideal is a generalized weak ideal, a weakly right ideal, or a quasi-ideal, then R is a reduced left weakly regular ring. (2) R is a strongly regular ring if and only if R is an idempotent reflexive semi-commutative left GP-V’-ring whose every maximal essential left ideal is a generalized weak ideal, a weakly right ideal, or a quasi-ideal. (3) If R is a semi-primitive idempotent reflexive ring whose every simple singular left R-module is flat, and every maximal left ideal is a generalized weak ideal, then, for any nonzero element a∈R, there exists a positive integer n such that an≠0, and RaR+lan=R.

Cognitive radio (CR), as an emerging technology to improve the utilization of radio spectrum, the fundamental of CR technology is spectrum sensing, due to the detection performance being affected by various factors, spectrum sensing is challenging to achieve accurately. In recent years, many spectrum sensing algorithms have been proposed, such as energy detection algorithm, matched filter detection algorithm, cyclic stationary detection algorithm, etc. However, these algorithms are model-driven and require certain prior information. If the model assumptions are inaccurate or the prior information is challenging to obtain, the algorithms’ detection performance will be degraded. The development of artificial intelligence technology and deep learning provides a new way to realize spectrum sensing. In this paper, we design a cooperative spectrum sensing model based on the parallel connection of convolutional neural network (CNN) and long-short-term memory (LSTM), which makes full use of the complementary feature extraction capabilities of CNN and LSTM networks. Among them, CNN is used to extract hidden spatial features, and LSTM network is used to extract time features. Both CNN and LSTM can process the original dataset directly avoiding information feature loss when the network is connected serially. Experimental result shows that the detection performance of the proposed algorithm outperforms the conventional cooperative detection algorithm under low SNR condition. For example, when the number of cooperative users is 9 and the transmit power is 10, the detection probability of the proposed algorithm in this paper can reach more than 90%, which is much higher than the detection performance of other spectrum detection algorithms.

Facade and rooftop agriculture has great potential in regulating urban local microclimate, building energy conservation and carbon emission reduction. This study aims to address the limitation that the existing studies lack multi-effect simulations of integrated implementation of facade and rooftop agriculture. It takes into account the variability and complexity of thermal coefficients in planting layers during the simulations. By analyzing in-situ observations, the study examines the regulatory impact of integrated agriculture on urban microclimate. Furthermore, it uses observed data instead of DeST temperature database data to simulate the energy-saving effect and carbon emission reduction. In order to analyse the average economic efficiency of rooftop and facade agriculture in China, the study conducts a meta-analysis of existing cases in China. The results of the study are as follows: (1) The thermal insulation efficiency of the facade and rooftop integrated agriculture varies with weather conditions and diurnal variations. It reaches the strongest (1.68°C) on sunny days and around noon, (2) The integrated implementation of facade and rooftop agriculture on a normal home building in Changsha can save a total of 12,226.30 kW·h (5%) of electricity per year, and thus reducing 2,809.26 kg of carbon emissions, and (3) Based on meta-analysis and LCA calculation, the average net present value of the project in China for 40a is RMB -937712.18. The economic efficiency needs to be optimized. The results of this study reveal the ideal ecological benefits of facade and rooftop agriculture and the economic efficiency that hinders their implementation, providing a theoretical basis for building energy conservation and renovation, and the economic efficiency optimization in the regions with hot summer and cold winter.

Land use transition and its eco-environmental effects are important research topics. Its essence is the process that human activities exert interference to the ecological environment in the process of social and economic development, and the ecosystem resists interference and recovers and adapts to interference. The article starts from the transition of land use dominant morphology and takes ecological resilience as the breakthrough point. Based on four periods of land use data, this article studied the spatio-temporal evolution of land use and ecological resilience and the response of ecological resilience to land use transition in China from 1990 to 2020. The results showed as follows: (1) During the study period, the construction land in China continued to increase, and the forest land, grassland, and farmland showed a fluctuating trend. (2) The spatial distribution pattern of ecological resilience showed the characteristics of “high in the southeast and low in the northwest”. The mean value and total value of ecological resilience in the region decreased first and then increased, taking 2010 as the dividing line. The difference in ecological resilience increased first and then decreased. (3) Ecological land and construction land are the main types of land that affect the changes in ecological resilience. The higher the proportion of ecological lands such as forest land, grassland, and waters, the smaller the variable coefficient of ecological resilience. The higher the proportion of construction land, the greater the difference in ecosystem elasticity among different types of areas.