CN103714382A - Multi-index comprehensive evaluation method for reliability of urban rail train security detection sensor network - Google Patents
Multi-index comprehensive evaluation method for reliability of urban rail train security detection sensor network Download PDFInfo
- Publication number
- CN103714382A CN103714382A CN201310752571.6A CN201310752571A CN103714382A CN 103714382 A CN103714382 A CN 103714382A CN 201310752571 A CN201310752571 A CN 201310752571A CN 103714382 A CN103714382 A CN 103714382A
- Authority
- CN
- China
- Prior art keywords
- layer
- value
- output
- neural network
- hidden layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention belongs to the technical field of modern traffic safety, and discloses a multi-index comprehensive evaluation method for the reliability of an urban rail train security detection sensor network. A series of indexes influencing the reliability of the urban rail train security detection sensor network is summarized, and the four indexes of time delay, the occupation ratio, the packet loss probability and the error rate are selected as evaluation indexes. The measured values of the evaluation indexes are normalized through a fuzzy mathematic theory, subordinative function values are obtained as samples, a BP neural network is built, the samples are trained, a weight value and a threshold value are modified through a steepest descent method, the samples are trained repeatedly in the same way until the error of an actual output value and the error of a calculated output value are within an acceptable range, the training is finished, and the obtained BP neural network with expert knowledge can comprehensively evaluate the reliability of the urban rail train security detection sensor network. The brand-new method is provided for evaluating the performance of the urban rail train security detection sensor network, and theoretical and practical supports are provided for optimizing the performance of the urban rail train security detection sensor network.
Description
Technical field
The invention belongs to and belong to Modern Traffic safety technique field.Be particularly related to a kind of many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure.
Background technology
Urban railway transit train, as one of main transport agent of the large capacity public transport in city, ensures that its safe operation has extreme importance and key to whole urban track traffic.The construction of municipal rail train safety detection Sensor Network is the important prerequisite of carrying out city railway train monitoring in transit and safe early warning, it is the necessary basis of formulating the driving safe measures such as municipal rail train Security Strategies, municipal rail train system failure detection, municipal rail train Accident Causes Analysis, guaranteeing that it is the prerequisite of building municipal rail train safety detection Sensor Network that municipal rail train safety detection Sensor Network has height reliability, is the necessary condition that ensures safe train operation.
To the unfailing performance assessment of municipal rail train safety detection Sensor Network, be based on certain single index mostly at present, as time delay, packet loss, the bit error rate etc., or many indexs evaluation of introducing subjective factor, as expert's scoring.Yet single index or expert marking all can not reflect the reliability of safety detection Sensor Network comprehensively, objectively, the reliability that therefore must carry out objectively comprehensive assessment to many index could be more comprehensive, objective, reliably, accurately assess municipal rail train safety detection Sensor Network.
Artificial neural network (Artificial Neural Networks, ANNs) be the artificial intelligence technology of 20th century develop rapidly eighties, it is the information processing system understanding of cerebral tissue structure and active mechanism being proposed based on the mankind, being can parallel processing information, carry out non-linear conversion and self-organization adjustment, has the complicated network system of adaptive learning ability and Error Tolerance.Neural network will be inputted sample and actual output as training sample, by calculating, sample be carried out to the training of abundant number of times, make neural network successfully set up the mapping relations of inputting with outlet chamber, and the neural network model after training just can solve similar problem.Neural network is developed so far, and has been each architectonical, and the most representative have BP neural network, adaptive resonance network, Hopfield network, a self-organized mapping network etc.BP neural network is to take the feed-forward type network that error back propagation study is algorithm, use gradient method of steepest descent to search for, square error minimum according to calculating output valve and real output value is principle, progressively weights and the threshold value of each layer of recursive resolve BP neural network.If calculate, the difference of output valve and real output value is undesirable just adjusts weights and threshold value recalculates, until calculate the difference of output valve and real output value, meets the requirements, and finishes to calculate, and obtains network model.
Summary of the invention
The object of this invention is to provide a kind of many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure, it is characterized in that, the method selects to affect the index of municipal rail train safety detection Sensor Network reliability, choose time delay, occupation rate, packet loss and bit error rate four indices and input sample as network, use BP neural network in conjunction with fuzzy mathematics theory, municipal rail train safety detection Sensor Network reliability to be assessed, comprise the steps:
(1) process sample, the four indices such as time delay, occupation rate, packet loss and the bit error rate are measured, use the subordinate function in fuzzy mathematics theory to be normalized measured value, by formula, calculate corresponding membership function value, make it have relative comparability, and be quantized to one can more interval [1,1] in, as the sample input value of BP neural network input layer, sample output valve is corresponding Sensor Network reliability score value;
(2) build BP neural network, determine network structure, set up BP neural network model.BP neural network comprises an input layer, a plurality of hidden layer and an output layer; Adopt is the most also that most widely used three layers of BP neural network with an input layer, a hidden layer and an output layer are as the BP neural network structure building;
(3) training sample, inputs each index membership function value, calculates the output valve of BP neural network hidden layer by formula, and the output valve of hidden layer is the input value of output layer, calculates the output valve of output layer by formula;
(4) adjust weights, according to real output value with calculate the poor of output valve, adjust weights, return to scheme the 3rd step, recalculate hidden layer output valve and output layer output valve, so adjust weights, calculate, adjust again weights, calculate again, until BP neural network real output value and calculating output valve error are less than the minimal value of setting, reach network convergence state, finish to calculate, obtain final weights;
(5) computed reliability, the BP neural network now obtaining has expert's experimental knowledge, with this model, just can calculate according to indices measured value the reliability of municipal rail train safety detection Sensor Network.
Described step (2) builds BP neural network, because every evaluation index varies in size on the impact of municipal rail train safety detection Sensor Network reliability, for every evaluation index is compared mutually in system, there is relative comparability, must use fuzzy mathematics theory to do normalized to every evaluation index, every evaluation index after processing could be as the training sample of BP neural network, time delay, utilization factor, packet loss, the bit error rate is quantitative target, different is time delay, packet loss, the bit error rate is reverse index, the smaller the better, and utilization factor belongs to forward index, be the bigger the better.
Beneficial effect of the present invention: the present invention crosses and measures the multinomial performance index that affect Sensor Network reliability, according to municipal rail train safety detection Sensor Network, the four indices data such as test time delay, occupation rate, packet loss and the bit error rate, in conjunction with fuzzy mathematics theory, use BP neural network, the unfailing performance of municipal rail train safety monitoring Sensor Network is carried out to comprehensive, objective, reliable, assessment accurately, and to this Sensor Network reliability marking, make it more directly perceived, be convenient to people and understand.Use fuzzy mathematics theory by each index measurement value normalization, calculate its membership function value, make it have comparability, by building BP neural network, sample is carried out to enough training, obtain having the BP neural network of expertise, the unfailing performance of assessment municipal rail train safety detection Sensor Network, for the net assessment of municipal rail train safety detection Sensor Network performance provides reference and foundation, for ensureing that safe train operation provides technical support.
Accompanying drawing explanation
Fig. 1 is concrete operations process flow diagram of the present invention.
Fig. 2 is three layers of BP neural network topology structure figure of the present invention.
Embodiment
The invention provides a kind of many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure, below in conjunction with accompanying drawing, preferred embodiment is elaborated;
Embodiment 1:
Figure 1 shows that many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure process flow diagram, contain following steps:
(1) process sample.
Because every evaluation index varies in size on the impact of municipal rail train safety detection Sensor Network reliability, for every evaluation index can be compared mutually in system, there is relative comparability, must use fuzzy mathematics theory to do normalized to every evaluation index, the every evaluation index after processing just can be used as the training sample of BP neural network.Time delay, utilization factor, packet loss, the bit error rate are quantitative target, and time delay that different is, packet loss, the bit error rate are reverse indexs, the smaller the better, and utilization factor belongs to forward index, is the bigger the better.The present invention is utilizing BP neural network to do before comprehensive assessment train safety detection Sensor Network reliability, use fuzzy mathematics theory, first the actual measured value of every evaluation index is done to normalized, according to membership function, measured value is normalized in [1,1] this dimensionless interval.
The normalization formula of forward index:
The normalization formula of reverse index:
Transfer index measurement value to membership function value on interval [1,1]:
Wherein,
be i measured value of k evaluation index,
be the mean value of k all measured values of evaluation index,
for intermediate conversion variable,
be the membership function value of i measured value correspondence in interval [1,1] of k evaluation index, the i.e. input value of BP neural network input layer.
From above-mentioned concrete transfer process, can find out, the measured value of forward index is larger, the membership function value corresponding with it is larger, when measured value reach a certain height, membership function value approaches " saturated ", and the application of formula (3) can prevent that evaluation index measured value is excessive and cause comprehensive assessment inaccurate.
(2) build BP neural network, determine network structure, set up network model.
BP neural network structure comprises input layer, output layer and one or more hidden layer.Generally, a hidden layer just can make neural network have certain precision and ability to express; The present invention adopts three layers of BP neural network with an input layer, a hidden layer and an output layer as the BP neural network structure building.
BP neural network input layer unit number by train safety detection Sensor Network reliability assessment index number, determined.The present invention adopts time delay, utilization factor, packet loss, bit error rate four indices to assess.If input layer vector X=is (x
1, x
2..., x
n)
t,
k=(1,2 ..., p), input layer unit number P=4 in the present invention.X wherein
ibe i sample,
i the measured value corresponding membership function value in interval [1,1] that is k evaluation index is input layer input value, and n is sample number.
One of output layer unit, is the reliability score value of train safety detection Sensor Network, and the present invention adopts centesimal system scoring.Output vector is made as Y=(y
1, y
2..., y
n)
t,
s=(1,2 ..., m), output layer unit number m=1 in the present invention.Wherein
it is the output valve of s unit of output layer in the i time training.
Hidden layer is to exist in order to calculate, and does not possess concrete meaning, and hidden layer unit number q can calculate by formula.
In formula, a is the constant between [1,10].If hidden layer output vector is output layer input vector is U=(u
1, u
2..., u
n)
t,
t=(1,2 ..., q).Wherein
it is the output valve of t unit of hidden layer in the i time training.
If input layer is v to hidden layer weight vector, hidden layer is w to output layer weight vector.BP neural network topology structure as shown in Figure 2.
(3) training sample.
BP neural network belongs to feed-forward type network, and training process comprises forward-propagating and two stages of error back propagation.First, input layer input sample, processes through hidden layer, passes to output layer, and this is the forward-propagating stage.If actual output with calculate output valve and be not inconsistent, by error with certain form backpropagation, be transmitted to hidden layer and pass to again input layer, error is shared equally to each layer of unit, obtain each layer of elemental error signal, revise each layer of weights, this is the error back propagation stage.The continuous adjustment process of weights is the training process of network.
From input layer to hidden layer:
Initial weight ν and initial threshold θ are set, are all made as random minimum value.
The membership function value of input evaluation index, as the input sample of BP neural network, is calculated the output valve of hidden layer and output layer.
In formula (5), ν
ktfor k neuron of BP neural network input layer and hidden layer t the neuronic weights that are connected,
be i measured value of k evaluation index corresponding membership function value in interval [1,1], i.e. input layer input value, θ
tt the neuronic threshold value for hidden layer.
Hidden layer neuron activation function adopts sigmoid function:
f(x)=1/[1+exp(-x)] (6)
From hidden layer to output layer:
The output valve of hidden layer neuron now
be the neuronic input value of output layer, output layer neuronal activation value is
W in formula (7)
tsfor t neuron of BP neural network hidden layer and s neuron of output layer connected be connected weights, r
ss the neuronic threshold value for output layer.
Output layer neuron activation functions is the same with hidden layer neuron activation function, adopts sigmoid function:
f(x)=1/[1+exp(-x)] (9)
(4) adjust weights.
Utilize method of steepest descent, according to formula, adjust output layer weights ω
ts:
ω
ts(i+1)=ω
ts(i)+Δω
ts
Wherein, η is scale-up factor, i.e. learning rate is made as between [0,1] one compared with decimal.ω
ts(i+1) be the i+1 time training in hidden layer to output layer weights.
Utilize method of steepest descent, according to formula, adjust output layer weights ν
kt:
ν
kt(i+1)=ν
kt(i)Δν
kt
Wherein, ν
kt(i+1) be the i+1 time training in input layer to hidden layer weights.
Return to step (3), train next time, until BP neural network real output value and calculating output valve error are less than the minimal value of setting, reach network convergence state.
(5) computed reliability
The BP neural network model now obtaining has expert's experimental knowledge, according to the index measurement value providing, just can assess arbitrary safety detection Sensor Network reliability, according to the reliability score value of output, just can understand the reliability standard of Sensor Network.
Claims (5)
1. many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure, is characterized in that, contains following steps:
(1) process sample, the four indices such as time delay, occupation rate, packet loss and the bit error rate are measured, use the subordinate function in fuzzy mathematics theory to be normalized measured value, by formula, calculate corresponding membership function value, make it have relative comparability, and be quantized to one can more interval [1,1] in, as the sample input value of BP neural network input layer, sample output valve is corresponding Sensor Network reliability score value;
(2) build BP neural network, determine network structure, set up BP neural network model.BP neural network comprises an input layer, a plurality of hidden layer and an output layer; Adopt is the most also that most widely used three layers of BP neural network with an input layer, a hidden layer and an output layer are as the BP neural network structure building;
(3) training sample, inputs each index membership function value, calculates the output valve of BP neural network hidden layer by formula, and the output valve of hidden layer is the input value of output layer, calculates the output valve of output layer by formula;
(4) adjust weights, according to real output value with calculate the poor of output valve, adjust weights, return to scheme the 3rd step, recalculate hidden layer output valve and output layer output valve, so adjust weights, calculate, adjust again weights, calculate again, until BP neural network real output value and calculating output valve error are less than the minimal value of setting, reach network convergence state, finish to calculate, obtain final weights;
(5) computed reliability, the BP neural network now obtaining has expert's experimental knowledge, with this model, just can calculate according to indices measured value the reliability of municipal rail train safety detection Sensor Network.
2. a kind of many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure according to claim 1, it is characterized in that, described step (2) builds BP neural network, because every evaluation index varies in size on the impact of municipal rail train safety detection Sensor Network reliability, for every evaluation index is compared mutually in system, there is relative comparability, must use fuzzy mathematics theory to do normalized to every evaluation index, every evaluation index after processing could be as the training sample of BP neural network, time delay, utilization factor, packet loss, the bit error rate is quantitative target, different is time delay, packet loss, the bit error rate is reverse index, the smaller the better, and utilization factor belongs to forward index, be the bigger the better.
3. a kind of many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure according to claim 2, is characterized in that, described utilization fuzzy mathematics theory is done normalized to every evaluation index, sets up evaluation index normalization model,
The normalization formula of forward index:
The normalization formula of reverse index:
Transfer index measurement value to membership function value on interval [1,1]:
Wherein,
be i measured value of k evaluation index,
be the mean value of k all measured values of evaluation index,
for intermediate conversion variable,
be that i measured value of k evaluation index is in interval
the membership function value of interior correspondence, the i.e. input value of BP neural network input layer.
4. a kind of many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure according to claim 1, is characterized in that, described step (2) builds BP neural network model
If input layer vector X=is (x
1, x
2..., x
n)
t,
k=(1,2 ... p), x wherein
ibe i sample,
i the measured value corresponding membership function value in interval [1,1] that is k evaluation index is input layer input value, and n is sample number;
Output vector is made as Y=(y
1, y
2..., y
n)
t,
s=(1,2 ..., m), wherein
it is the output valve of s unit of output layer in the i time training; This output layer unit is the reliability score value of train safety detection Sensor Network;
Hidden layer is to exist in order to calculate, and does not possess concrete meaning, and hidden layer unit number q can calculate by formula,
In formula, a is the constant between [1,10], and establishing hidden layer output vector is that output layer input vector is U=(u
1, u
2..., u
n)
t,
t=(1,2 ..., q).Wherein
it is the output valve of t unit of hidden layer in the i time training;
If input layer is v to hidden layer weight vector, hidden layer is w to output layer weight vector.
5. a kind of many index comprehensives of municipal rail train safety detection Sensor Network reliability appraisal procedure according to claim 1, is characterized in that, described step (3) training sample, sets up sample training model,
From input layer to hidden layer:
Initial weight υ and initial threshold θ are set, are all made as random minimum value;
The membership function value of input evaluation index, as the input sample of BP neural network, is calculated the output valve of hidden layer and output layer;
In formula (5), ν
ktfor k neuron of BP neural network input layer and hidden layer t the neuronic weights that are connected,
be i measured value of k evaluation index corresponding membership function value in interval [1,1], i.e. input layer input value, θ
tfor t neuronic threshold value of hidden layer,
Hidden layer neuron activation function adopts sigmoid function:
f(x)=1/[1+exp(-x)] (6)
From hidden layer to output layer:
The output valve of hidden layer neuron now
be the neuronic input value of output layer, output layer neuronal activation value is
W in formula (7)
tsfor t neuron of BP neural network hidden layer and s neuron of output layer connected be connected weights, r
ss the neuronic threshold value for output layer.
Output layer neuron activation functions is the same with hidden layer neuron activation function, adopts sigmoid function:
S neuron output value of output layer is
Described step (4) is adjusted weights, comprising: establishing known real output value is
real output value with the difference of calculating output valve is
Utilize method of steepest descent, according to formula, adjust output layer weights ω
ts:
ω
ts(i+1)=ω
ts(i)+Δω
ts
Wherein, η is scale-up factor, i.e. learning rate is made as between [0,1] one compared with decimal; ω
ts(i+1) be the i+1 time training in hidden layer to output layer weights;
Utilize method of steepest descent, according to formula, adjust output layer weights ν
kt:
ν
kt(i+1)=ν
kt(i)+Δν
kt
Wherein, ν
kt(i+1) be the i+1 time training in input layer to hidden layer weights; Return to step (3), train next time, until BP neural network real output value and calculating output valve error are less than the minimal value of setting, reach network convergence state, finish to calculate, obtain final weights.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310752571.6A CN103714382B (en) | 2013-12-31 | 2013-12-31 | A kind of multi-index comprehensive evaluation method for reliability of urban rail train security detection sensor network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310752571.6A CN103714382B (en) | 2013-12-31 | 2013-12-31 | A kind of multi-index comprehensive evaluation method for reliability of urban rail train security detection sensor network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103714382A true CN103714382A (en) | 2014-04-09 |
CN103714382B CN103714382B (en) | 2016-08-17 |
Family
ID=50407337
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310752571.6A Active CN103714382B (en) | 2013-12-31 | 2013-12-31 | A kind of multi-index comprehensive evaluation method for reliability of urban rail train security detection sensor network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103714382B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104537238A (en) * | 2014-12-27 | 2015-04-22 | 广州至德电力科技有限公司 | Networked relay protection reliability evaluating system |
CN105046377A (en) * | 2015-09-06 | 2015-11-11 | 河海大学 | Method for screening optimum indexes of reservoir flood control dispatching scheme based on BP neural network |
CN106845776A (en) * | 2016-12-21 | 2017-06-13 | 吴中区穹窿山倪源交通器材经营部 | A kind of Rail Transit System runs safety evaluation method |
CN106886487A (en) * | 2015-12-15 | 2017-06-23 | 北京京航计算通讯研究所 | Method for evaluating FPGA software reliabilities |
CN108964969A (en) * | 2018-05-07 | 2018-12-07 | 中国铁路总公司 | The high-speed railway signal system method for predicting of hybrid neural networks and AR model |
CN109240254A (en) * | 2017-07-10 | 2019-01-18 | 卢照敢 | Vehicle condition assessment and fault early warning method based on vehicle condition data flow |
CN113219282A (en) * | 2021-04-30 | 2021-08-06 | 华能国际电力股份有限公司上海石洞口第二电厂 | Novel microcomputer protection constant value checking method |
CN113326962A (en) * | 2020-02-28 | 2021-08-31 | 南京理工大学 | Subway train ATO speed curve prediction method based on BP neural network |
CN113923139A (en) * | 2021-12-15 | 2022-01-11 | 北京城市轨道交通咨询有限公司 | Method and device for evaluating reliability of train control data communication system |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103235843A (en) * | 2013-04-03 | 2013-08-07 | 北京交通大学 | Train operation optimal control simulation method and system for urban railway system |
-
2013
- 2013-12-31 CN CN201310752571.6A patent/CN103714382B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103235843A (en) * | 2013-04-03 | 2013-08-07 | 北京交通大学 | Train operation optimal control simulation method and system for urban railway system |
Non-Patent Citations (3)
Title |
---|
WENLONG P ETAL.: "Access point research in rail train safety monitoring sensor network", 《THIRD INTERNATIONAL CONFERENCE ON DIGITAL MANUFACTURING AND AUTOMATION(ICDMA)》, 31 July 2012 (2012-07-31) * |
王艳辉 等: "铁路货运车辆运行状态安全评估方法及应用研究", 《铁道学报》, vol. 33, no. 4, 30 April 2011 (2011-04-30) * |
贾天丽: "轨道列车走行部滚动轴承故障诊断研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑 》, 15 September 2011 (2011-09-15) * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104537238A (en) * | 2014-12-27 | 2015-04-22 | 广州至德电力科技有限公司 | Networked relay protection reliability evaluating system |
CN104537238B (en) * | 2014-12-27 | 2017-10-10 | 广州至德电力科技有限公司 | A kind of networking reliability of relay protection assessment system |
CN105046377A (en) * | 2015-09-06 | 2015-11-11 | 河海大学 | Method for screening optimum indexes of reservoir flood control dispatching scheme based on BP neural network |
CN106886487A (en) * | 2015-12-15 | 2017-06-23 | 北京京航计算通讯研究所 | Method for evaluating FPGA software reliabilities |
CN106845776A (en) * | 2016-12-21 | 2017-06-13 | 吴中区穹窿山倪源交通器材经营部 | A kind of Rail Transit System runs safety evaluation method |
CN106845776B (en) * | 2016-12-21 | 2021-03-12 | 中兴(温州)轨道通讯技术有限公司 | Rail transit system operation safety assessment method |
CN109240254A (en) * | 2017-07-10 | 2019-01-18 | 卢照敢 | Vehicle condition assessment and fault early warning method based on vehicle condition data flow |
CN108964969A (en) * | 2018-05-07 | 2018-12-07 | 中国铁路总公司 | The high-speed railway signal system method for predicting of hybrid neural networks and AR model |
CN108964969B (en) * | 2018-05-07 | 2021-12-07 | 中国铁路总公司 | High-speed railway signal system flow prediction method based on hybrid neural network and AR model |
CN113326962A (en) * | 2020-02-28 | 2021-08-31 | 南京理工大学 | Subway train ATO speed curve prediction method based on BP neural network |
CN113219282A (en) * | 2021-04-30 | 2021-08-06 | 华能国际电力股份有限公司上海石洞口第二电厂 | Novel microcomputer protection constant value checking method |
CN113219282B (en) * | 2021-04-30 | 2022-12-27 | 华能国际电力股份有限公司上海石洞口第二电厂 | Novel microcomputer protection constant value checking method |
CN113923139A (en) * | 2021-12-15 | 2022-01-11 | 北京城市轨道交通咨询有限公司 | Method and device for evaluating reliability of train control data communication system |
CN113923139B (en) * | 2021-12-15 | 2022-03-01 | 北京城市轨道交通咨询有限公司 | Method and device for evaluating reliability of train control data communication system |
Also Published As
Publication number | Publication date |
---|---|
CN103714382B (en) | 2016-08-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103714382A (en) | Multi-index comprehensive evaluation method for reliability of urban rail train security detection sensor network | |
CN103226741B (en) | Public supply mains tube explosion prediction method | |
CN103606006B (en) | Sludge volume index (SVI) soft measuring method based on self-organized T-S fuzzy nerve network | |
CN109117987A (en) | Personalized street accidents risks based on deep learning predict recommended method | |
CN110401978B (en) | Indoor positioning method based on neural network and particle filter multi-source fusion | |
CN106355540A (en) | Small- and medium-sized reservoir dam safety evaluating method based on GRA-BP (grey relational analysis and back propagation) neural network | |
CN108171641A (en) | A kind of rail traffic emergency preplan appraisal procedure | |
Ning et al. | GA-BP air quality evaluation method based on fuzzy theory. | |
Hamoda | Modeling of construction noise for environmental impact assessment | |
CN105427062A (en) | Reservoir dam collapse risk analysis method based on cloud theory | |
Mansourkhaki et al. | Comparative application of radial basis function and multilayer perceptron neural networks to predict traffic noise pollution in Tehran roads | |
CN106896724A (en) | For the tracking system and tracking of sun tracker | |
CN108538301B (en) | Intelligent digital musical instrument based on neural network audio technology | |
CN116307775A (en) | Evaluation method and system | |
Ma et al. | Traffic flow forecasting research based on Bayesian normalized Elman neural network | |
Xie et al. | A method of flood forecasting of chaotic radial basis function neural network | |
CN113191689B (en) | Land suitability evaluation method for coupling principal component analysis and BP neural network | |
CN112183847B (en) | QRNN improved Stacking algorithm-based train running wind speed probability prediction method | |
CN107505931A (en) | A kind of aircraft engine maintenance level decisions method based on depth belief network | |
CN115034648B (en) | Bridge engineering risk assessment method based on BP neural network under condition of few samples | |
Wang et al. | A model of coal mine water disaster emergency logistics risk assessment based on AHP and BP neural network | |
Yekta et al. | Suspended sediment estimation of Ekbatan reservoir sub basin using adaptive neuro-fuzzy inference systems (ANFIS), artificial neural networks (ANN), and sediment rating curves (SRC) | |
Akhter | Application of ANN for the Hydrological Modeling | |
CN105654395A (en) | Comprehensive water quality evaluating method based on compensation type Hopfield neural network | |
Fang | Traffic flow prediction analysis method based on the improved genetic algorithm. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |