In this paper, the realizable short-code direct sequence code-division multiple-access (DS-CDMA) system occupying Nyquist chip bandwidth and exhibiting zero direct-current (DC) is developed. By limiting the spreading waveform to exhibit zero DC and continuous bandlimited spectrum, the realizability and DC nulling constraints on spreading codewords are derived to enable such system realization. Based on both constraints, a low-complexity implementation structure in spreading and despreading is also addressed. Several frequently-adopted spreading codes are examined for realization feasibility. It is shown that Nyquist chip bandwidth and zero-DC transmission can be practically achieved for the bandlimited DS-CDMA system using two conventional even-length orthogonal spreading codes at the sacrifice of very little user capacity. A new odd-length orthogonal spreading code is also proposed to enable such system realization.
Under the boom of low-orbit satellite Internet constellation construction, an LDPC decoder design method is proposed for the outgoing physical layer of satellite Internet in order to carry out effective monitoring for typical satellite signals with high bandwidth and low time delay. The scheme uses a partially parallel architecture, where the node update step based on the normalized min-sum algorithm omits the resource consumption and compresses the number of clocks consumed by operations such as multi-level comparison of message values, which helps to improve the decoder throughput rate. Based on Virtex-7 series hardware platform, we realize the high rate decoding of long and short frame LDPC codes with different modulation methods in DVB-S2/S2X standard. At 204.8 MHz operating frequency, the decoder can reach 768 Mbps throughput. The decoder implemented in this paper is capable of meeting the rate and performance requirements of typical satellite systems for targeted signal monitoring.
A safety helmet detection model has been developed based on the original YOLOv5 network model, which addresses the problems of low accuracy and poor robustness for small targets in complex natural environments. Firstly, we add the layer to detect small targets, resulting in the shallow detection scale, which makes the model better able to detect small objects. Secondly, introducing the transformer self-attention module in the backbone network, which can extract global features with rich semantic information, improves the detection performance. Thirdly, The Ghost module demands fewer parameters and less computational complexity for features to be generated, reducing the pressure to learn model parameters. Finally, the loss function EIoU Loss is applied to improve the detection accuracy and accelerate the convergence of the model. Experimental results show that the average accuracy means (MAP), accuracy rates (P), and recall rates (R) of the improved algorithm are increased by 1.2%, 0.4%, and 0.2%, respectively while reducing the model size, which proves that the algorithm is more effective than the original algorithm at detecting safety helmets.
The rise and development of software-defined networking (SDN) has brought a new development direction to the Internet. As a new network architecture, SDN has better flexibility and programmability than traditional network architectures, which brings new opportunities for the development of traffic management. In this paper, the development and trend of SDN are firstly introduced, and then some representative works of SDN in data center traffic management are analyzed. Finally, some research results related to traffic management based on the combination of SDN and artificial intelligence (AI) technology are expounded, and the challenges and development prospects of SDN traffic management are pointed out.
The development of 5G networks has brought higher requirements for services such as switching, routing, and QoS (Quality of Service) of the carrying network and the core network. Packet classification is an essential technology which plays an important role in guaranteeing the high bandwidth and low latency of the network. Many hardware-based algorithms have been presented to meet the high requirements. However, the resource overhead on chip become greater and greater as the network scale expands so fast that it is urgent to improve the storage efficiency. Thus, an effective classification method based on extended bloom filter and cuckoo hash is proposed in this paper. It preserves the feature of the extended bloom filter by locating entries according to the comparison result of multiple hash counter groups, and further introduces the hash index comparison which slows down the increase of the counter groups greatly. In addition, a cuckoo hash with limited number of collisions and a reasonable deletion mechanism are also presented in this paper to ensure higher stability. The experimental results show that the method has the advantages of simple structure, wire-speed lookup capability, high stability, and high storage utilization.
Wireless sensor networks (WSNs) generally consist of thousands of sensor nodes, each one supplied by a battery or harvested energy. To prolong the lifetime of wireless sensor networks, wake-up receivers (WuRxs) are typically employed. WuRxs can selectively activate sensor nodes by decoding a signal called wake-up call (WuC). Therefore, they optimize the power management of WSNs, by allowing communication when requested. In this article, an ultra-low power implementation of a sequential WuRx concept is proposed. The WuRx is composed of AND gates, switches, and monostable circuits. The monostable circuit is implemented through a transistor that generates a pulse according to a simple RC network and a switch. The proposed WuRx decodes the WuC signal by comparing the duty cycle of the received signal with the output signals of the monostable circuits. The ultra-low power implementation is validated at simulation level. The WuRx consumes 32.8nW when decoding a 3-bit wake-up call signal and 153nW when decoding 11 bits.
In recent years, the detection of illegal and harmful messages which plays an significant role in Internet service is highly valued by the government and society. Although artificial intelligence technology is increasingly applied to actual operating systems, it is still a big challenge to be applied to systems that require high real-time performance. This paper provides a real-time detection system solution based on artificial intelligence technology. We first introduce the background of real-time detection of illegal and harmful messages. Second, we propose a complete set of intelligent detection system schemes for real-time detection, and conduct technical exploration and innovation in the media classification process including detection model optimization, traffic monitoring and automatic configuration algorithm. Finally, we carry out corresponding performance verification.
Combined with the actual conditions of VHF communication system in inland waters, the distance, frequency and antenna height of the radio wave coverage effect calculation model proposed in Recommendation ITU-R P.1546-6 are modified to obtain the VHF path loss function curve adapted to inland waters. The results of the real ship test show that the modified radio wave coverage effect prediction model can predict the coverage effect of VHF communication system better in inland waters, and has high promotion and application value.
The image and text matching task plays an essential role in bringing the semantic gulf between language and vision. It still remains challenging since previous methods lack a detailed comprehension of the contextual interplays reflected in diverse visual relationships between objects. In this work, we present one novel dual relation-aware synergistic attention (DRSA) network to produce visual representations that incorporate crucial semantic concepts and salient objects of image scenes. First, we construct each image as two subgraphs and perform the multi-type inter-object interactions utilizing the sentence-guided graph attention mechanism. Specially, two types of visual relations are exploited: Implicit Relations extracting the latent dynamics between objects and Explicit Relations capturing the semantic dependencies and relative geometric positions. Second, a synergistic fusion module is designed to adaptively merge the implicit, explicit, and all-mixed relation features based on the sentence contexts, which works like multi-head attention. Third, adversarial learning is conducted to reinforce the interaction of implicit and explicit relation encoding modules to explore more effective multi-view associations between images and sentences. Experiments show that DRSA outperforms existing state-of-the-art approaches on two widely used datasets (i.e., MSCOCO, Flickr30K), proving the efficacy of our proposed elaborate matching technique.
In order to solve the problem of key abuse in attribute-based encryption, a traceable ciphertext-policy attribute-based encryption scheme is proposed based on the learning with error over ring and the ordered binary decision diagram structure. Firstly, the specific information of the user is embedded in each private key to realize the traceability of the key. Secondly, the access of illegal users and malicious users is avoided by maintaining the white list. In addition, the scheme adopts the access structure of ordered binary decision diagram, and increases the positive and negative values of attributes on the basis of supporting the operation of “AND”, “OR” and “Threshold”. Finally, through analysis, the scheme meets the security of anti-collusion attack and chosen-plaintext attack security, reduces the storage and computing overhead, and is more practical than other schemes.
Wireless revolution has evolved in ever-increasing demands on an already constrained wireless spectrum, driving the quest for increased bandwidths, greater channel capacity, and faster access rates. In-band full-duplexing (IBFD) is a promising technology that aims to answer this mitigation call by bolstering spectral efficiency through simultaneous same frequency band transmission and reception. However, transmitter-to-receiver leakage, or self-interference, remains the most barring frustration to IBFD realization. Therefore, self-interference cancellation (SIC) is necessary to suppress any polluting SI to a level below the receiver’s noise floor. The level of energy reduced by SIC determines the quality of the radio as well as overall system throughput. In this paper, a simulation model of an IBFD transceiver is developed in MATLAB, with inclusion of various sources of distortion. Subsequent results detailing an investigative study done on a fully adaptive tapped-branch analog self-interference canceller are shown. Evaluation of the results reveal marginal effect on the SIC efficacy due to transmission path noise and nonlinear distortion alone. However, expansion of model consideration for conceivable cancellation hardware nonlinearity reveals an indirectly proportional degradation of SIC performance by up to 35dB as distortion levels vary from −80dBm to −10dBm. These results indicate consideration of such nonidealities should be an integral part of cancellation hardware design for the preclusion of any intrinsic cancellation impediments.
The liquid crystal holographic antenna is a high-gain antenna that can obtain good directivity. Moreover, the dielectric constant of the liquid crystal can be changed by adjusting the bias voltage of the liquid crystal material to realize the beam scanning characteristics of the liquid crystal holographic antenna. However, because there are too many antenna elements in a holographic antenna, it is very difficult to apply a bias voltage to each element individually. This paper adopts a new unit structure [1], under the premise of not affecting the performance of the antenna, the metal patch of all the units can be connected together, the entire antenna only needs one bias voltage to realize the beam of the liquid crystal holographic antenna scanning. Based on this antenna unit structure, this paper designs an antenna that works at 27GHz. The main beam can be scanned from Phi=0°, Theta=40° to Phi=25°, Theta=32°, and the gain is 16dBi.
Attention mechanisms have been widely studied and applied in various computer vision tasks because of their ability to establish inter-dependencies between channels and spatial positions. Recent works show it necessary to provide light-weight and effective implementations for higher computation efficiency. To this end, this paper proposes to restrict the attention to a local region for object detection and presents a distance-based attention mechanism to incorporate contextual information from regional areas, instead of the full image, in an efficient and effective way. Furthermore, we implement the proposed approach in an object detection model that combines tricks from several state-of-the-art models. Specifically, the popular one-stage RetinaNet is used as the backbone and the two-step classification and regression, feature enhancement module, max-out operation and multi-scale testing are adopted to obtain a high detection accuracy. Experimental results on WIDER FACE show that the proposed detection model outperforms many state-of-the-art models and the proposed attention mechanism can greatly increase the detection accuracy without introducing extra computational overheads.
This paper presents a detailed analysis of the crosstalk that occurred in multi-channel pipelined-successive approximation register (pipelined-SAR) analog-to-digital converters (ADCs). According to the circuit model simulation results, the capacitive crosstalk dominates while inductive crosstalk is negligible. The different paths of crosstalk are analyzed and their influences are compared. The result shows that most crosstalk will transform into noise in the output, degrading the signal-to-noise ratio (SNR) performance of the ADC. And some specific types of digital outputs can generate large harmonics that affect spurious-free dynamic range (SFDR) performance. Furthermore, an adaptive-filter-based cancellation method is proposed for input coupled crosstalk. The method is validated in the test of a 16-channel 14-bit 50-MS/s ADC, and the measurement results show that crosstalk is effectively eliminated.
The prediction of failure numbers is helpful to develop and optimize the strategies of operation and management for smart electricity meters. In this paper, a failure number prediction method for smart electricity meters based on Weibull distribution and odds ratio was proposed. Then strategies for spare parts inventory and rotation were developed based on the on-site failure data of smart electricity meters. The odds ratio was used to compute the distribution of failure number of meters. The confidence intervals of distribution parameters and failure numbers in the future time interval were given. The research results have reference value for making operation plan and reserve strategy of electricity meters.
Under the background of COVID-19, zero contact offline office and distance teaching mode has become a major trend in the education industry. In order to study the network operation and maintenance (O&M) technology in the campus network environment, the basic mechanism and operation performance of two traditional O&M solutions (separation and centralized) are compared first, based on the remote office engineering practice project in local university. It also points out the defects of the two schemes. In order to solve the defects of the traditional scheme, a new linkage O&M solution is proposed, which cooperates VPN and fortress machine. Practical research shows that the hardware occupancy of the new scheme under high O&M concurrent threads is 15% ~ 20% lower than that of the traditional scheme, which effectively improves the utilization of fortress machine O&M system and reduces the load of VPN, the campus network information security is realized.
Weak signal detection is widely used in many aspects. However, limited by the signal-to-noise ratio, detection under low signal-to-noise ratio has always been a challenge. The performance of traditional energy detection methods has deteriorated significantly under low signal-to-noise ratio or faced with noise uncertainty. The use of cross correlation methods can effectively improve the detection signal-to-noise ratio but threshold setting is closely related to noise power. To deal with noise uncertainty, in this paper, we propose a new test statistics for cross correlation detection, that is peak-to-average ratio. Firstly through mathematical derivation and simulation peak-to-average is proved to be robust against noise uncertainty. Then we design a detection algorithm for cross-correlation detection in the case of multiple signals, which expands the application scenarios of cross correlation detection. Finally, based on the experimental platform built, the performance of the algorithm is verified by taking the received UAV image transmission signal as an detection target.
In dynamic spectrum access (DSA) technology, it is particularly important for the secondary users (SUs) to accurately determine the vacant spectrum to ensure that the activity of the primary users (PUs) is not interrupted. However, in reality, SUs can’t sense all channels or accurately sense the channel state. In this paper, we consider DSA problem where PUs and SUs share the spectrum in an overlay manner under the condition that the priori information of the system is unknown and SUs have limited sensing capability. An algorithm that combines dropout with multi-layer perceptron (MLP) and deep Q-network (DQN) is proposed. SUs obtain channel occupation pattern of PUs through learning. Simulation results show that the algorithm has better performance in protecting PUs from interference, reducing conflicts between SUs, and increasing the average channel capacity compared with three other algorithms: myopic, random access and DQN-RC (reservoir computing).
Civil unmanned aerial vehicle (UAV) is an increasing threat to personal privacy and public safety. Passive UAV detection and localization is a hot topic. This paper presents a platform with six synchronized channels for real-time locating the UAV. In this system, an antenna array is arranged in a short baseline, and its matching RF front-end adopts the synchronous design. The front-end supports a sweep frequency from 230MHz to 6GHz. The passive location functions are implemented with the hardware processing unit and a computer. The rational division of algorithms helps accelerate the algorithm and realize a real-time performance. Finally, an experiment was designed to test it. The results of the experiment proved the ability and accuracy of the platform.
This paper proposed a wide-locking-range Divide-by-3 Injection-Locked Frequency Divider (ILFD3). In this paper, a fourth-order second harmonic enhanced resonator is proposed to strength the second harmonic and improve the mixing efficiency of the feedback signal and the injection signal. A new switched-capacitor tuning structure is also proposed. By setting sub-bands between every two operation bands, a wider locking range can be obtained without using much more switched capacitors. The locking condition based the fourth-order resonator is also analyzed theoretically in this paper. Fabricated in 28nm CMOS technology, the proposed ILFD can achieve locking-range of 22.5–30.4GHz (29.8%), and the total DC current of this circuit is 3.67mA with 0.9V supply.
