Search

The problem of string stability control for the heterogeneous vehicular platoon subject to actuator fault and external disturbances is investigated. Specifically, a quadratic spacing error policy is applied to improve string stability despite any vehicle in the platoon occurring actuator fault. Then, a super twisting-based robust string stability controller is developed to achieve the fault-tolerant ability of the whole vehicular platoon system and ensure the spacing error converges to zero quickly. Moreover, the asymptotic convergence and the string stability in the whole vehicular platoon are also proved, respectively. Finally, the proposed control strategy is carried out on the heterogeneous vehicular platoon with the actuator fault and external disturbances, and effectiveness and control performance can be verified.

The development of information technology has brought changes in industrial technology. Traditional industry is declining and intelligent industry is rising. Compared with the traditional industry, intelligent industry has a stronger demand for computing resources and is more sensitive to delay. To meet these requirements, this paper designs a computing platform integrating edge and cloud resources. The platform is mainly composed of edge computing nodes, cloud system, and AI(Artificial Intelligence) component. Edge computing nodes provide network services and computing offload services, which can quickly respond to industrial tasks. Cloud system provides computing resources and storage services for high resource consumption tasks in industry. AI components make full use of idle resources in the edge network to form clusters to accelerate AI inference. Facing different industrial scenarios, the platform can flexibly allocate computing resources in the network and provide computing offload service for tasks in the industry. Finally, we designed three experimental scenarios to verify the effectiveness of our platform. In the three experimental scenarios, compared with the traditional task execution method, our platform has brought obvious productivity improvement and accelerated the completion speed of the task.

In a group decision making (GDM) problem, the information is fused to obtain a collective result, which helps to choose the best solution/s to the problem. Recently, a new representation model called Extended Comparative Linguistic Expressions with Symbolic Translation (ELICIT), which extends the representation of the comparative linguistic expressions (CLEs) to a continuous domain combining the advantages of the hesitant fuzzy linguistic term sets and 2-tuple linguistic representation model has been proposed to model experts' preferences. Due to the need of fusing experts' preferences in GDM processes, it is convenient to have enough flexible aggregation operators for such processes. However, so far, only two aggregation operators have been introduced to aggregate ELICIT information in GDM problems, the fuzzy arithmetic mean operator and the Bonferroni mean operator. Thus, it seems necessary to define new aggregation operators with different features to model wide range of decision-making scenarios. One widely used operator to aggregate preferences in decision making is the OWA operator. The key issue to apply the OWA operator is the reordering process of the arguments. However, the ELICIT information does not have an inherent order because it is represented by a fuzzy number. Therefore, the aim of this contribution is to define the type-1 ELICIT OWA operator by using crisp and fuzzy weights, particularly interval weights, and define a multi-criteria group decision making model which applies the type-1 ELICIT OWA operator to fuse the information. Additionally, an experimental study is introduced to demonstrate the feasibility of the proposed aggregation operator.

This paper introduces a high-speed visible light communication system based on Direct Current biased Optical-OFDM (DCO-OFDM), which combines the clustering algorithm (K-mediods) in machine learning with visible light communication to achieve post-equalization and signal demodulation. This equalization method can effectively reduce the impact of noise, facilitate the judgement of distorted constellation points, and enhance system performance. The system utilizes a commercial blue LED as emission source, a pre-equalization circuit to expand the modulation bandwidth, and a clustering algorithm as a post-equalizer. Based on DCO-OFDM, we experimentally demonstrated a 1.06 Gbit/s visible light link with 1m free space transmission, and experimental results indicated that the bit error rate (BER) is under the forward error correction threshold of 3.8×10-3.

The speckle noise in the echocardiogram reduces the image quality and brings difficulties to the clinical diagnosis and subsequent image processing. In order to effectively remove noise, this paper proposes a method for suppressing echocardiographic speckle noise based on bilateral filtering and piecewise linear transformation. Firstly, bilateral filtering is applied to filter the echocardiogram, then a sine-type function is applied to enhance the brightness of the image, and finally the piecewise linear transformation is applied to adjust the contrast of the image. Experimental results show that this method can effectively suppress speckle noise in echocardiograms, highlight the various chamber areas in the echocardiogram, improve contrast, and provide a good basis for clinical diagnosis and subsequent image feature extraction and recognition.

This paper proposes a medical MRI image enhancement method based on curvelet transform and fuzzy algorithm. First, the MRI image is subjected to curvelet positive transform to obtain the curvelet coefficients at various scales and directions, and then the Monte-Carlo test method is used to estimate each scale noise variance, and then apply hard threshold shrinkage processing to the curvelet coefficients. Finally, the Pal-King algorithm with modified membership function is used to perform fuzzy enhancement on the image after inverse curvelet transformation to obtain the final result image. We selected a brain MRI image to test the algorithm, the experimental results show that compared with the other two enhancement algorithms, the algorithm in this paper has higher PSNR and CONTRAST, which can effectively suppress noise, enhance the edges and details of the image, and has better visual effects.

A transient response enhancement circuit applied in peak-current mode converter is proposed, which can improve the system transient response. Because the bandwidth of the traditional peak-current mode Boost converter is limited by the right half plane zero, the transient response speed is limited during the load step. A circuit is proposed which can adaptively adjust the output voltage of the Boost converter and optimized the system transient response. The proposed circuit is simulated and verified by 0.18um BCD process. The recovery time was reduced from 67$\mu$s to 49$\mu$s while the load current changed from 1A to 200mA. The recovery time was reduced from 47$\mu$s to 35$\mu$s while the load current changed from 200mA to 1A.

In Multimedia Broadcast/Multicast Service (MBMS), Multicast/broadcast Single Frequency Network (MBSFN) and Single-cell Point-to-multipoint Broadcast are two important ways to organize networks to provide multicast/broadcast service. MBSFN shows its unique advantage on the enhanced signals at the boundary of two cells, however, highly requiring synchronization of symbols. In 5G era, we develop MBSFN in more scenarios, e.g. vehicular network. In this paper, we compare two ways of organizing networks and propose the optimization models using successful transmission probability (STP) to evaluate the performance in MBSFN mode and PTM mode separately. Then, we evaluate the data throughput based on that the users’ distribution follows the Poisson Point Process and get the approximation ratio of data throughput between MBSFN and PTM with the given transmitting power. The simulation results show the comparison between two modes, which well proves our theory.

A wideband microstrip antenna for X-band which is coupled by an H-shaped aperture is designed in this paper. Dual mode resonance is formed by parasitic patch technique and the scanning performance of the antenna element is improved by employing isolation columns. The VSWR bandwidth of the proposed antenna is 40% and its low profile feature facilitates the integration of array systems. Good agreement has been obtained between simulation and experimental results.

Couplers in monitoring network of airborne phased array radar are often arranged with equal spacing. This will cause the resonance of the network and degradation of the coupling performance. This phenomenon is analyzed and studied in this paper. The equivalent circuit model shows that the homo-phase superposition of reflexed signals by each coupling aperture is responsible for this phenomenon. On this basis, two optimization methods are studied with a 40-cell network as an example. The lengths equal to integer multiple of half the dielectric wavelengths of intra-band frequencies should avoid to be adopted as the distance between each coupler. These methods effectively suppress the resonance and make the disturbance of intra-band coupling signal less than 0.5dB. Measuring results show good accordance with the simulation data. Design requirements are satisfied.

Service placement is a significant issue in mobile edge computing (MEC) system. Many works have proposed efficient offline approaches for service placement problems in MEC system. However, because of the randomness and uncertainty of mobile networks, it is impractical for these approaches to be implemented. Facing these uncertainty, we propose an online service placement scheme for MEC system without knowing service demand and network states in advance. In order to maximize the long-term accumulated reward obtained by service placement with limited resource constraint, we analyse this problem by a combinatorial multi-armed bandit (MAB) framework. In addition, because we simultaneously consider the service placement and resource allocation among services, it can be formulated as a multiple choice knapsack problem (MCKP) in each time slot. To solve this long-term reward maximization problem, we first propose a combinatorial upper bound confidence(CUCB)-based online service placement and resource allocation scheme. Then, we analyse the performance of this algorithm theoretically. Finally, simulation results show the efficiency of the algorithm.

In order to solve the problem of interference of ambient light noise in visible light communication (VLC), an adaptive filtering algorithm based on a convolution neural network (CNN) is proposed in this paper. For different channels, CNN is used to learn the relationship between the receiver signal and the original transmitted symbol, and then the received signal is classified using the trained model. Besides, we conducted the VLC experiment to collect data, and the performance of CNN and the least mean square algorithm (LMS) algorithms was compared and analyzed using differential phase shift keying (DPSK) sample data collected through the experiment. The experimental results show that under the same signal-to-noise ratio (SNR), the Bit error rate (BER) of DPSK signal recovery based on CNN is 40% lower than that of the LMS algorithm.

Long-distance cables are often damaged by external forces due to site construction, resulting in protective layer grounding fault. Conventional fault location and detection methods, such as bridge method and traveling wave method, require on-site measurements along the cable, which results in heavy maintenance workload. For this reason, a method of online diagnosis and location of high-voltage cable based on sheath current is proposed. The simulation results show that when a fault occurs, the sheath current will change with the change of fault location and grounding resistance. Therefore, the external insulation faults of cables can be monitored, diagnosed and located on-line based on sheath current.

The pruning of spurious skeletons is an important issue when apply skeletons for object analysis and recognition. Most existing pruning methods use threshold-like parameter to control the effects of pruning. Such approaches usually encounter difficulties in parameter tuning and the quality of pruning. A skeleton pruning method is proposed in this work. This method decomposes a skeleton into a number of skeletal components (SCs), and the terminal SCs are removed one by one according to a saliency measure. SCs may be merged into a new one after each removal. The removal process continues until the desirable number of terminals are achieved. The method can essentially be regarded as the sorting of the terminals, and the truncation of the terminals at the given number. This method is suitable for applications where the desired number of terminals can be determined a priori. Experimental results show that the method can effectively prune the skeletons from the test image bases, and it has the advantage that the terminal number parameter is very intuitive and easy to set for the applications under concern.

In this paper, a new flame detection method using infrared images is proposed. By analyzing the spectrum of the sun, a 940nm narrow-band filter is applied to the camera to obtain an infrared image. Then, a threshold selection method for the flame is applied to obtain the binarized image, and then the similarity of consecutive video frames and the roundness of the contour are used to make the final judgment. Experimental results demonstrated that the proposed method has the advantages of high speed, low cost, and high accuracy.

As a high-risk industry for safety accidents in the construction industry, falling from a height is the most likely one to be occurred. Based on the finite element analysis method, this paper combines the safety net model with the human body model to obtain the human body-net dynamic simulation model. Through simulation, the effects of different factors on the fall protection performance of safety net are analyzed from different aspects. The maximum contact force of the human body model can be used as an evaluation index to reflect the protective performance of the safety net effectively in the research process. The research results show that under the condition of meeting the protection requirements and economy, reducing the height between the safety nets, and choosing a safety net with a larger overall size and a smaller mesh side length can effectively improve the protection performance of the safety net. This study can provide a reference for choosing, arranging and installing the safety net during construction industry.