Improved Viterbi Algorithm Research Articles

Non Intrusive Power Load Monitoring Method Based on Machine Learning

In order to improve the accuracy of non-invasive power load monitoring and shorten the monitoring time, a new non-invasive power load monitoring method based on machine learning is proposed. Firstly, the power load data is collected by sensors and normalized. Secondly, based on the normalized results of power load data, the RBF neural network model in machine learning is used for iterative training to extract power load characteristics. Finally, according to the characteristics of power load, the hidden Markov model is used to transform the non-invasive power load monitoring problem into a decoding problem, and the improved Viterbi algorithm is used to solve the hidden Markov model, and the non-invasive power load monitoring is completed according to the obtained state sequence.

An Improved Viterbi Algorithm for Adaptive Instantaneous Angular Speed Estimation and Its Application Into the Machine Fault Diagnosis

It is a very challenging task to diagnose the fault of rotating machinery under variable speed condition. The traditional tacholess order tracking (TLOT) method still has the problems of high computation complexity and low adaptability. Aiming at these problems, a new adaptive instantaneous angular speed (IAS) estimation method based on the improved Viterbi algorithm (VA) and the energy centrobaric correction method is developed. Since the traditional VA cannot process the long time series and find the fundamental order automatically, an improved penalty function for ridge search is proposed. Meanwhile, the search band of IAS is adaptively optimized to accelerate the speed of ridge search. Furthermore, an energy centrobaric correction approach is used for improving the accuracy of instantaneous frequency estimation and the antinoise performance. With the obtained IAS, the nonstationary time-domain vibration signal is resampled as the stationary angle-domain vibration signal by the Vold–Kalman filtering and the Hilbert transform. Finally, the envelope order spectrum is used to diagnose the mechanical faults of a civil aircraft engine. With the datasets of aircraft engine, rolling bearing, and wind turbine, the experimental results show that the improved VA can adaptively estimate IAS faster and achieve higher accuracy of fault detection than the typical IAS estimation methods. Meanwhile, its antinoise ability is verified.

A Method of Optimizing Network Topology Structure Combining Viterbi Algorithm and Bayesian Algorithm

With Internet entering all walks of life, development of internet and usage expansion demand better performance, especially the application of 5G network that adopts NAS networking mode. Some of the network bandwidth cannot fully support the current network demand, which causes network fluctuations and other concerns. In this paper, a method for optimizing the topological structure of the bottom layer of the communication network is proposed that has outage performance close to optimal routing scheme. In specific, path in areas with poor network conditions is first optimized using Viterbi algorithm. Then, network element nodes on the path are optimized using Bayes recommendation algorithm for reasonable flow distribution. Dual planning of improved Viterbi algorithm is used to realize the main and standby path planning, and then, Bayesian recommendation algorithm based on the average value is used to optimize the network elements. Therefore, it is very efficient to realize overall performance optimization.

Target Signal Separation of Missile Group Based on Micro-motion Frequency Correlation

Object micro-motion feature extraction method for isolated targets is relatively mature, but the resolution and micro-motion feature extraction method for group targets still needs further research. Ballistic missiles in the middle of the throw-off process is taken to be the example, when the false target is released. A method of target separation by micro-Doppler frequency is proposed, and the target is modeled. A parameterized expression is given. The narrow band echo is processed by STFT and the time-frequency plane is obtained. After that, the BM3D algorithm is used to filter and reduce noise to smooth the curve. The improved Viterbi algorithm is used to extract the micro-Doppler curve of the group target. Finally, the Fourier transform is performed. According to the frequency correlation, the group target is separated, and the simulation verified this method.

Research on Non-intrusive Load Decomposition Based on FHMM

Non-intrusive load monitoring and decomposition, as one of the important parts of intelligent power utilization system, can deeply analyze users’ internal load components and obtain user’s electricity consumption information from different scales, which is of great significance to users and power companies. In this paper, a non-intrusive load decomposition method based on factorial hidden Markov model using low frequency data is proposed. Kmeans-II algorithm is used to cluster the working state of a single load, the results of which are used to calculate the parameters of the HMM for individual load model. The total load model is represented by a factorial hidden Markov model, which transforms the load decomposition into an optimization problem with maximum probability. The improved Viterbi algorithm based on event detection is proposed to solve this optimization problem, so as to obtain the working state sequence and realize load decomposition. Finally, the correctness and practicability of the method are verified by an example.

A non-intrusive load decomposition method for residents

Aiming at the problem that the large amount of data involved in the existing load decomposition method leads to low decomposition efficiency and high hardware requirements, a non-intrusive load decomposition method based on hidden Markov model (HMM) and improved Viterbi algorithm is proposed. Monitor steady state current of each load and quantify the data obtained and then establish the probability distribution of quantized monitoring values by statistics. The states that a load has can be identified from the probability distribution. The running states of each load at the same moment are arranged in sequence to form a combination state. HMM is created based on the correlation of combination states in the transition and the correlation between the combination state and the quantized monitoring value of the total current. Based on matrix sparsity, a compression algorithm is used to reduce data storage, and a query algorithm and an improved Viterbi algorithm are used to avoid unnecessary calculations about zero-probability terms to further improve the decomposition efficiency. Experimental results show that this method can effectively improve the decomposition efficiency and need lower hardware requirements while accurately obtaining the running state of various loads.

A non-intrusive load decomposition algorithm for residents

In view of the large amount of data involved in the existing decomposition algorithm, which leads to low decomposition efficiency and high hardware requirements, a non-intrusive load decomposition method based on hidden Markov model (HMM) and improved Viterbi algorithm is proposed. First, the steady-state current data of the load is monitored and collected, and then the monitoring value is quantified and the probability distribution of the quantized monitoring value is created. Finally, the state of the load is identified from the probability distribution. According to the correlation of a composite state transition composed of states of multiple loads, the corresponding HMM model is created. Based on the matrix sparsity, the compression algorithm is used to reduce the amount of data stored. By using the query algorithm and the improved Viterbi algorithm, the calculation of zero probability items can be avoided, and the decomposition efficiency can be greatly improved. Results from simulation and real data have been used to verify the performance of the proposed algorithm.

New Miller Codes for Run-Length Control in Visible Light Communications

Designing run-length limited codes for visible light communication systems must account for multiple performance factors, including spectral efficiency, power efficiency, dc balance, and flicker avoidance. This paper reports a new class of enhanced Miller codes, termed eMiller codes, which are capable of achieving highly desirable performances in all of these accounts. An improved Viterbi algorithm (VA), termed $mn$ VA, is developed to help further enhance the performance of eMiller codes by preserving multiple candidate sequences at each decoding stage. This performance-enhancing algorithm introduces little complexity increase compared with the original VA. Analysis on flicker control, power spectral density, and minimum Hamming distance demonstrates the all-around wellness of these new codes. Extensive simulations are carried out to evaluate eMiller codes by themselves and in practical visible light communication (VLC) systems. It is shown that the original VA already allows eMiller codes to deliver a performance noticeably better than conventional Miller and FM0/FM1 codes (and on par with Manchester codes). This result is particularly exciting, as eMiller codes are also more spectrally efficient than Manchester codes. The $mn$ VA further allows eMiller codes to surpass Manchester codes and 4B6B codes in practical RS-coded VLC systems. Simulation results confirm the superb performance of the RS-eMiller schemes.

An improved Viterbi algorithm for IF extraction of multicomponent signals

Viterbi algorithm (VA) applied to time–frequency (TF) representation is a highly performed instantaneous frequency (IF) estimator for discrete-time signals, but it suffers from switch problem (SP) at the intersected points of multicomponents on TF plane. To suppress the SP in VA, an improved VA (IVA) presented in this paper assumes that the IF variation trends between two adjacent IF variation are not large, and then, a novel penalty function is introduced and added to the original VA. To verify the algorithm, the proposed algorithm applied to several multicomponent signals is firstly simulated; then, how parameter in the new penalty function influences the performance is analyzed. Comparison of the proposed algorithm with VA on signals in the background of noise is also made in the next. Simulations indicate that in contrast to the original VA, the proposed IVA can effectively suppress the SP caused by the intersected IFs and thus can achieve more accurate IFs for the multicomponent signals especially those with monotonous IFs.

A New Method for User Dynamic Clustering Based On HSMM in Model of SaaS

This paper deeply studies the phenomenon of hard to satisfy the user's personalized services and only a few researches on users themselves in the model of Software as a Service (SaaS), then proposes a users' behavior feature extraction model based on Hidden Semi-Markov Models (HSMM) to solve the problem of getting users hidden information on SaaS platform first. The model uses the probability distribution of state duration time to control user's browsing behaviors, combines hidden states which describe features with time relativity, and applies improved Viterbi algorithm to get user features sequence. Then cluster users by dynamic K-means algorithm, which doesn't need to give K cluster centers in the process of clustering but adjusts center value automatically through the comparison of clustering quality in every iterative process, finally gets optimal clustering results. Detailed simulation analysis demonstrates that the presented algorithm is of high efficiency of space and time and is more stable.