299 results about "State transition probability" patented technology

The invention relates to mobile tracking of vehicle targets in the communication and transportation field. A self-adapting interactive multiple model mobile target tracking method comprises the following steps of establishing mixed initialized input. including a covariance matrix of a mixed initial condition and a mixed initial state. of each model; establishing constant velocity (CV) and constant acceleration (CA) motion models; updating: calculating covariance matrix and innovation of an error according to a Kalman filter formulation; constructing a likelihood function of a target motion module by utilizing the innovation of a Kalman filter result, and calculating a Markov state transition probability matrix; carrying out estimation output after being fused by utilizing the Markov state transition probability matrix as weight of switch among each motion model. According to the self-adapting interacting multiple model mobile target tracking method disclosed by the invention, the problem that the error is increased or the tracking is failed caused by non-matching of a filter model and a target motion model due to the motion of a target in the traditional interactive multiple model algorithm is solved; the self-adapting interacting multiple model mobile target tracking method has the advantages of low calculation complexity and good tracking effect, and can be applied to target tracking of motor vehicles in the communication and transportation field.

The invention relates to a modeling approach and a modeling system of an acoustic model used in speech recognition. The modeling approach includes the steps of: S1, training an initial model, wherein a modeling unit is a tri-phone state which is clustered by a phoneme decision tree and a state transition probability is provided by the model, S2, obtaining state information of a frame level based on the fact that the initial model aligns the tri-phone state of phonetic features of training data compulsively, S3, pre-training a deep neural network to obtain initial weights of each hidden layer, S4, training the initialized network through error back propagation algorithm based on the obtained frame level state information and updating the weights. According to the modeling approach, a context relevant tri-phone state is used as the modeling unit, the model is established based on the deep neural network, weight of each hidden layer of the network is initialized through restricted Boltzmann algorithm, and the weights can be updated subsequently by means of error back propagation algorithm. Therefore, risk that the network is easy to get into local extremum in pre-training is relieved effectively, and modeling accuracy of the acoustic model is improved greatly.

The invention discloses a pure electric automobile remaining mileage model predication method based on path information. The pure electric automobile remaining mileage model predication method based on path information comprises the following steps of analyzing driver history running data, extracting the path information and generating a state transferring probability matrix satisfying driver behavior characteristics; generating a predicated automobile speed on the basis of road information of a future path and the corresponding state transferring probability matrix; establishing a parameter estimation model to estimate running parameters affecting energy consumption and remaining mileage of an automobile; and establishing an RDR calculation model to predicate a vehicle remaining mileage,wherein an energy consumption predication model is used for calculating a vehicle energy consumption rate by using the predicated automobile speed obtained by an automobile speed predication model andthe running parameters estimated by the parameter estimation model as model input; a remaining energy predication model is used for estimating vehicle battery remaining energy; by integrating the vehicle energy consumption rate and the battery remaining energy, the vehicle remaining mileage can be predicated and can be displayed by using a remaining mileage display model.

The invention discloses a prediction method of driving risk based on a hidden Markov model. The method comprises the steps of (1) classifying driving risk states through a cluster analysis method based on vehicle operating characteristics, (2) estimating the influences of a driver behavior and surrounding traffic environment characteristics on a transition probability between driving risk states through multiple logistical models for different driving risk states, (3) with a risk state as a hidden state, with an actual observed vehicle movement variable as a state output value, with multiple logistic model parameters as parameter initial values of a state transition probability matrix, establishing a hidden Markov chain model that reflects the evolution rule of driving states, and (4) obtaining the vehicle operating characteristics in real time and predicting a future risk state based on the hidden Markov chain model. According to the method, the hidden Markov model which can reflect the above characteristic real-time change and has a variable state transition probability is established, the accuracy and prediction accuracy of a driving risk model are improved, and the real-time requirements of anti-collision warning can be satisfied.

An embodiment of the invention provides a map matching method and device. The map matching method comprises steps as follows: multiple track points are acquired, and at least one candidate road segment corresponding to each track point on an electronic map is acquired; observation probability between a first track point and any one candidate road segment corresponding to the first track point in the multiple track points is calculated, state transition probability between each candidate road segment corresponding to the first track point and each candidate road segment corresponding to a second track point is calculated, and the second track point and the first track point are adjacent track points; a road segment sequence matched with the multiple track points is determined according to the observation probability and the state transition probability. According to the map matching method and the device, the optimal road segment sequence matched with the track points is obtained under the condition that a limited quantity of the track points with low positioning precision are obtained, and the map matching robustness is improved.

A plurality of stochastic models is built that predict the probabilities of state transitions for components in a complex system. The models are trained using output observations from the system at runtime. The overall state and health of the system can be determined at runtime by analyzing the distribution of current component states among the possible states. Subsequent to a low level component failure, the state transition probability stochastic model for the failed component can be analyzed by uncovering the previous states at N time intervals prior to the failure. The resulting state transition path for the component can be analyzed for the causes of the failure. Additionally, component failures resulting from the failure, or worsening state transition, in other components can be diagnosed by uncovering the previous states at the N times prior to the failure for multiple components in the system and then analyzing the state transition paths for correlations to the failed component. Additionally, transitions to worsening states can be predicted using an action matrix. The action matrix is created beforehand using state information and transition probabilities derived from a component's stochastic model. The action matrix is populated probabilities of state transitions at a current state for given actions. At runtime, when an action is requested of a component, the probability of the component transitioning to a worsening state by performing the action can be assessed from the action matrix by using the current state of the component (available from the stochastic model).

The invention discloses a residual error posterior-based abnormal value online detection and confidence degree assessment method. The method comprises the steps of collecting data, establishing time series data, performing linear fitting on the time series data to obtain a linear combination formula of data at a current moment and p pieces of previous data, and predicting a data value of subsequent time; comparing the predicted data value with an actually detected data value to obtain a predicted residual error series; determining a probability density function of the predicted residual error series by adopting a KDE (Kernel Density Estimation) method; performing posterior ratio check on the predicted residual error series, and judging whether the data at the current moment is an abnormal point or not; and by taking the time series data as an input, building an SOM state model, obtaining state series and state transition probability matrixes, defining an abnormal scoring function, and outputting an abnormal score. By comparing the probability that the data is the abnormal point with the probability that the data is a normal point, the abnormal value in the pollutant discharge concentration time series data is identified online, so that the accuracy and reliability of abnormal value judgment are improved.

The invention provides a city rescue intelligent agent dynamic path planning method based on an improved ant colony algorithm, and belongs to the technical field of robot simulation. Aiming at the issue of path planning of a dynamic changing environment in robot rescue simulation, the classic ant colony algorithm is improved, the target dominance is introduced, the approach for calculating the probability of state transfer of ants is modified, and the pheromone updating rules adapt to a situation in which the road condition is unknown and dynamically changes, a rescue intelligent agent is given complex tasks, and the path planning requirements are inconsistent in a rescue environment, so that a rescue robot can search the optimum path according to the task attribute. Compared with the classic ant colony algorithm, the improved ant colony algorithm of the invention is effective and reliable.

The invention provides a method for planning tasks of unmanned aerial vehicles on the basis of Q(lambda) algorithms. The method includes a step of carrying out environment modeling, a step of initializing Markov decision process models, a step of carrying out Q(lambda) algorithm iterative computation and a step of computing the optimal paths according to state value functions. The method particularly includes initializing grid spaces according to the minimum flight path section lengths of the unmanned aerial vehicles, mapping coordinates of the grid spaces to obtain airway points and representing circular and polygonal threat regions; building Markov decision models, to be more specific, representing flight action spaces of the unmanned aerial vehicles, designing state transition probability and constructing reward functions; carrying out iterative computation on the basis of constructed models by the aid of the Q(lambda) algorithms; computing each optimal path of the corresponding unmanned aerial vehicle according to the ultimate convergent state value functions. The unmanned aerial vehicles can safely avoid the threat regions via the optimal paths computed according to the ultimate convergent state value functions. The method has the advantages that the traditional Q learning algorithms and effectiveness tracking are combined with one another, accordingly, the value functionconvergence speeds can be increased, the value function convergence precision can be enhanced, and the unmanned aerial vehicles can be guided to avoid the threat regions and autonomously plan paths.

A method and apparatus are described where account information is used to predict possible outcomes of a HELOC. To predict the possible outcomes, HELOC account state transition probabilities are modeled. The transition probabilities, determined by historic data regression analysis, provide the framework for a Monte Carlo simulation. The simulation is seeded with HELOC account information. A calculation engine takes the account information and simulates an elapse of time using a random number generator and the state transition probabilities. The simulation results in updated account information predicting a possible outcome over the elapsed time interval. The updated account information in turn may be used by the calculation engine to simulate the next elapse of time. This method may be iteratively repeated with the account information propagated forward until the end of the prediction period is reached.

The present invention belongs to the research field of equipment fault prediction, and relates to an equipment fault prediction method in a power communication network based on the Weibull distribution and the hidden semi-Markov model (HSMM). According to the method provided by the present invention, the actual degradation of power equipment is considered, and the hidden Markov model (HMM) with the excellent condition monitoring and degradation diagnosis and recognition capability is introduced; the equipment failure rate is modeled by using the Weibull distribution; the Taylor series expanding-combining like terms method is used to estimate degradation factors; and a state dwell time probability and a state transition probability matrix of the HSMM is calculated, the actual evolution of the equipment fault can be satisfied, the parameter fitting error can be reduced, and better prediction can be performed on the equipment fault state in the communication network.

The invention relates to a wireless locating method which can be used for location with high degree of accuracy in a mixed environment of sight distance and non-line of sight. The method first sets up motion equations and observation equations of wireless location and then expresses state transition probability model of the non-line of sight and the sight distance, which can make use of rectified extended Kalman filter (EKF) to estimate the motion state and the non-line of sight state according to measured values obtained by every base station and then blends the motion state and the non-line of sight state together through the use of a data fusion method to get the estimation of the motion state at the present moment and at last on-line wireless device position solutions can be realized through loop iteration. The method of the invention can effectively solve the non-line of sight influence in wireless location so as to effectively improve the motion state estimation of wireless devices, which has robustness to LOS/NLOS transition probability in different environments. At the same time, the method is suitable for VLSI parallel processing, operand can meet real time requirements, and the method is suitable for different signal measuring methods such as TOA, RSS, etc.

The invention discloses an emergency logistic transfer station site selection method based on an improved ant colony algorithm. The method comprises the steps of: constructing a multi-target optimizing model with an maximum emergency demand coverage degree and minimum emergency cost serving as a target, and combining constraint conditions of the optimizing model to determine the number of the temporary transfer stations; combining the clustering ant colony algorithm and a greedy algorithm to solve the multi-target optimizing model, successively performing clustering analysis on coverage situations of disaster points under each disaster level, determining grouping of the disaster points and making a primary site selection scheme of the transfer stations; evaluating the unselected transfer stations, replacing the primary selected transfer station with the transfer station of a maximum evaluation value, performing transfer station repositioning until the site selection scheme does not change; and outputting the final transfer station site selection scheme. According to the invention, the ant colony algorithm is improved on the application of the state transition probability, the sharing tabu table and the greedy algorithm idea, the convergence rate and the solving quality are increased and a theoretical basis is provided for the optimizing of the emergency logistic transfer station site selection.

The invention provides a multi-unmanned aerial vehicle (UAV) track planning method based on a culture ant colony search mechanism, which includes the following steps: (1) carrying out mesh generationon a standard space according to a grid method; (2) building a multi-UAV track planning model, including the number of UAVs, the start and end points and a threat model; (3) initializing the start point and the end point; (4) initializing an ant colony algorithm, including: initializing an ant colony and calculating a heuristic factor and a guide factor; and (5) assigning all ants to an initial node, and updating taboo knowledge; selecting next node for transfer according to the taboo knowledge and the state transfer probability until there is no optional node or a destination node is selected, updating historical knowledge, and updating pheromones according to the historical knowledge; and outputting a shortest path if the maximum number of iterations is achieved, and continuing the process until U multi-UAV optimal multi-path tracks are obtained. The problem that it is difficult to find the optimal flight tracks of unmanned aerial vehicles due to slow search and heavy computing burden is solved, and multi-UAV track planning is realized.

The invention provides an oil field output prediction method based on a dynamic radial basis function neural network. The method comprises the steps that 1, factors which affect the output are determined according to oil field situations, and historical data are obtained and divided into a training data set and a test data set; 2, unitization processing is conducted on the data sets through a deviation standardization method; 3, an RBF neural network structure is adjusted in a dynamic mode through a sensitivity method, and a temporary RBF neural network prediction model is established; 4, a model error is corrected through a state transition probability matrix, and a stable RBF neural network oil output prediction model is obtained; 5, verification is conducted on the model through the test data sets obtained in the first step to judge whether the model meets expectations or not; 6 oil field output prediction is conducted through the output prediction model which meets the expectations and obtained in the fifth step. According to the oil field output prediction method based on the dynamic radial basis function neural network, the problem that the hidden layer neurons are too many or too small is avoided. and the obtained model has an adaptive adjustment function; second correction is conducted on a prediction error, and the prediction result is more accurate and reasonable.

In order to obtain customer state transition probabilities and short-term rewards conditioned by actions, customer behaviors are modeled with a hidden Markov model (HMM) using composite states each composed of a pair of a customer sate and a marketing action. Parameters of the estimated hidden Markov model (the composite state transition probabilities and a reward distribution for each composite state) are further transformed into the customer state transition probabilities and the distribution of rewards for each customer state conditioned by marketing actions. In order to model purchase properties in more detail, a time interval between purchases (called an inter-purchase time, below) is always included as an element in the customer state vector, thereby allowing the customer state to have information on the probability distribution of the inter-purchase time.

A system and method recognizes speech securely. The system includes a client and a server, The client is configured to provide securely speech in a form of an observation sequence of symbols, and the server is configured to provide securely a multiple trained hidden Markov models (HMMs), each trained HMM including a multiple states, a state transition probability distribution and an initial state distribution, and each state including a subset of the observation symbols and an observation symbol probability distribution. The observation symbol probability distributions are modeled by mixtures of Gaussian distributions. Also included are means for determining securely, for each HMM, a likelihood the observation sequence is produced by the states of the HMM, and means for determining a particular symbol with a maximum likelihood of a particular subset of the symbols corresponding to the speech.

The invention relates to the technical field of mobile communications and provides a method for selecting a safe relay for multiple targets in a mobile collaborative network. In the method provided by the invention, a channel fading threshold as well as and the various states and the state transition probability of candidate relay nodes are determined through periodically transmitting a training sequence and feedback information; when the fading speed of some channel is greater than the threshold, dynamic relay selection is carried out, or else, semi-static relay selection is carried out; and during the process of relay selection, the optimal safe relay is selected to collaborate data transmission according to the total compensation determined by a plurality of target values, such as system safety capacity, and the like. With the adoption of the method provided by the invention, the characteristics of wireless channels in the mobile collaborative network can be accurately and timely described, and the dynamic selection of the optimal safe relay for the multiple targets is realized through lower complexity, so that the method has strong flexibility and expandability; and the transmission efficiency is further increased through carrying out different relay selections according to channel states.

The invention relates to the technical field of mobile communications and provides a method for selecting an optimal distributed type interference source in a mobile collaborative network. In the method, the possible states and the state transition probability of candidate collaborative nodes are obtained through periodically transmitting a training sequence and feedback information; whether a channel source node can be in direct communication with a destination node is judged, if the direct communication is available, the selection of the optimal interference node is carried out, or else, the selection of an optimal relay-interference pair is carried out; and a corresponding priority symbol value of each possible state of the collaborative nodes is calculated, and a suitable node is selected to collaborate data transmission. With the adoption of the method provided by the invention, the optimal interference source can be selected according the specific states of a channel, so as to respond to the interception of illegal nodes in the mobile collaborative network, the power consumption and the transmission error rate of the collaborative nodes are sufficiently considered at the same time that the secure communication is guaranteed, and the quality of communication is ensured.

The embodiment of the invention provides a method for selecting a network in a heterogeneous network, which comprises the following steps that an access network at each decision time is determined according to benefits, state transition probabilities and state probability vectors of each network in various states; and according to the access network at each decision time, the heterogeneous network obtains the largest system benefit, wherein the decision time is the time for sending an access request or a leave request to the network by a user, and the state of the network is the number of various users probably to be accessed to each network. Because the correlativity of state changes of each network in a time quantum is considered, the access network at each decision time is determined according to the benefits, the state transition probabilities and the state probability vectors of each network in various states at each decision time in a plurality of continuous decision times so that the heterogeneous network obtains the largest system benefits during the decision period comprising the plurality of the decision times.

The invention relates to a keyword identification method based on a hidden markov model, and the method comprises the following steps: S1, constructing the hidden markov model, wherein the hidden markov model comprises five elements including a hidden state S, an observable state O, an initial state probability matrix pi, a hidden state transition probability matrix A and an observation state matrix B; S2, after separating a target article into a word + word class format through a word segmentation algorithm, inputting the article into the built hidden markov model, acquiring an observable state sequence O, then, inputting the observable state sequence O into the built hidden markov model, and thereby obtaining a model mu; S3, based on the built hidden markov model mu and the obtained observation state sequence O = {O1, O2, ..., OT}, calculating a maximum possible value of the hidden state through a viterbi algorithm, thereby identifying whether each word is the keyword. With the method, the device and the storage medium provided by the invention, better universality is realized, the keywords can be executed simultaneously for a relatively long article or a relatively short article, and identification accuracy is high.

For voltage values (observed noise sequence) in an electronic power line (communication medium) which are obtained at a predetermined interval, initial values of noise characteristics based on a statistic of the observed noise sequence itself are decided by a moment method (S301 to S307), the noise characteristics (state transition probabilities and state noise power) for maximization of the likelihood of the observed noise sequence are obtained from the initial values by MAP (Maximum A Posteriori) estimation using a Baum-Welch algorithm (S309 to S312), a state sequence is estimated from the obtained noise characteristics, and an impulsive noise at each time point is detected.