Haifeng Wang

A combined GPS UWB and MARG locationing algorithm for indoor and outdoor mixed scenario

An indoor or outdoor positioning system is hard to position in cross-region and complex environment. In order to make up for the loss of blind zone in scenario change, we proposed a cooperative positioning system based on GPS/UWB/MARG, which can achieve the seamless positioning between buildings in the hybrid scene. In the hybrid scene, using a weighted fusion algorithm to achieve the cooperative positioning of GPS and UWB, and MARG is used to improve the positioning accuracy of GPS. With data optimization and performance analysis of a subsystem, we process GPS/MARG data and UWB data by weighted fusion method. The optimal positioning information in different positioning environments will be output after independent judgment. The results show that, compared with GPS/MARG positioning system, the average positioning accuracy of the cooperative positioning system is improved by 64% in the hybrid scene. In addition, it expands the application scenario of a single positioning system.

Research on the Technology Used to Inspect the Visual Appearance of Tropical Fruit, based on Machine Vision Color Space

Machine vision (computer vision) technology and image processing has been widely used for the automatic classification of fruit—representing an area for future research. In this chapter, we mainly use the method of machine vision to study color characteristics, size contours, and surface textures of fruit surfaces in the hue, saturation, and value (HSV) color space. The chapter goes on to realize the automatic classification of Hainan tropical fruit presenting typical characteristics.

Research on External Quality Inspection Technology of Tropical Fruits Based on Computer Vision

With the computer vision technology in the image processing has been widely used, which for the automatic classification of fruit provides a research space. This paper mainly uses the method of computer vision, combined with the problem of grade quality detection of agricultural products in agricultural research hotspots. Taking tropical fruit of Hainan as the research object, taking mango as the experimental object, extracting the characteristics of fruit image, explore the differences in the external size and color of different types of fruits, and establish a visual quality inspection technology for tropical fruits based on computer vision.

Application of Clinical Diagnosis and Treatment Data of Coronary Heart Disease Based on Association Rules

According to the characteristics of Chinese medicine diagnosis and treatment of coronary heart disease and the need for mining, this paper will introduce the association rules mining, from the medical treatment of patients with all aspects of the disease and the basis of coronary heart disease diagnosis and treatment of Chinese medicine between the basis of digging out the law of Chinese medicine, the application of association rules algorithm, Get a series of rules of coronary heart disease syndrome, for the diagnosis and prevention of coronary heart disease provides an important basis for decision-making.

Cluster computing data mining based on massive intrusion interference constraints in hybrid networks

To solve the problem that massive intrusion data in hybrid networks greatly interfere network intrusion detection and cause relatively great difficulty to detection due to their frequency discontinuity, a mining algorithm of massive intrusion cluster computing data in hybrid networks based on spectral feature extraction under fixed constraints of time–frequency window is proposed in this paper. The multi-component cross-detection method is used to collect massive intrusion information in hybrid networks and construct a model of massive intrusion signal in hybrid networks. Cascading notch method is used to suppress intrusion interference under constraints of fixed time–frequency window, and extract fundamental quantity and primary function with locality in massive interference information, and obtain a complete energy distribution spectrum on the time–frequency plane. The energy distribution spectrum is used as guidance function to realize cluster calculating data mining with massive intrusion interference constraints. Simulation results show that, in the intrusion detection process of signal-to-noise ratio from −xa015 to −xa05xa0dB, the detection accuracy of the method proposed in this paper is always better than that of others. When the signal-to-noise ratio is −xa09xa0dB, the detection accuracy of this method is over 90%. When the signal-to-noise ratio is −9xa0dB, the detection accuracy of this method can reach 100%. About the detection time, especially after the Number of intrusion date is 2000, the detection time difference between the three methods is increased, and the detection time of other methods is 2–3 times of paper method. The method proposed can accurately locate mass intrusion distribution sources in hybrid networks under the large frequency oscillation of intrusion data in the hybrid networks, realize network intrusion detection and interference suppression and can filter interference information well, so this method improves intrusion detection probability in hybrid networks.

Interrupt protection control of anti-interference nodes in network based on band sampling decision filter modulation

In network real-time communication, interrupt protection control needs to be conducted for interference nodes due to the multipath interference in the link layer. However, there have always been some incorrect operations in selecting optimal nodes, such as large errors and non-optimal masking results. Therefore, an interrupt protection method for anti-interference nodes in network real-time communication based on band sampling decision filter modulation is proposed in this paper. In this method, a multipath transmission link model of network real-time communication is constructed; an impulse response function of network communication is established; copy autocorrelation processing for signals of communication node sampling is conducted to realize the transformation from time domain to frequency domain; band sampling decision filter is used to conduct noise suppression for interference nodes; an interrupt decision is conducted for filtering output signals in network real-time communication according to the least mean square error criterion, and whether to conduct interrupt mask for nodes is decided according to the threshold, so as to realize the interrupt protection control for anti-interference nodes in network real-time communication. The simulation results show that by using this method to conduct interrupt protection for communication nodes, the output bit error rate is in general lower than those got by the traditional methods, and the highest bit error rate does not exceed 0.5. It tends to be flat from −xa04SNR/dB, and completely unaffected by the interference nodes after 4SNR/dB. The method proposed in this paper restrains the interference nodes, and effectively improves the anti-interference capability and the communication quality.