CN105608833A - Early-warning system and method for danger of express delivery vehicle - Google Patents

Abstract

The invention relates to an early-warning system and method for danger of an express delivery vehicle. A system monitoring terminal node collects environment parameter monitoring data in a compartment and position monitoring data of an express delivery vehicle and sends the data to a coordinator; the coordinator gathers the monitoring data and sends the data to a gateway module; the gateway module carries out processing on the monitoring data and then carries out data frame format conversion and protocol conversion, and then uploads the processed data to a cloud server; the cloud server receives the data information transmitted by the gateway module and stores the data information into a database, and the gateway module sends the data information to an alarm unit simultaneously; and the alarm unit determines whether the data information is abnormal; if so, the alarm unit sends an alarm signal out and transmits the alarm information to the gateway module and the cloud server in real time; and if not, safety information is transmitted to the gateway module in real time and is displayed on a display screen. According to the invention, early-warning for danger that may occur during transportation of the express delivery vehicle is realized; the safety transportation of the express delivery vehicle is guaranteed; and the security and survivability of the system are high.

Description

A kind of express delivery car danger early warning system and method

Technical field

The invention belongs to logistics monitoring and warning technical field, relate to a kind of express delivery car danger early warning system and method.

Background technology

The proposition of plan along with " internet+", internet thinking has incorporated majority's work and lifeIn, electric business is as the important economic industry pillar of China, and shopping at network has become one of most people lifeWhen part, the express delivery that connects enterprise and client just becomes the pass of promoting economic development and improving public life qualityKey factor. The express delivery amount of China in 2015 has broken through 20,000,000,000, will present the coming years at a high speedDeveloping state, in 2015 that just pass by " two 11 " shopping joint, the logistics order of day cat platform justReach 4.67 hundred million, China has become genuine the first in the world express delivery big country. But, be accompanied by magnanimityWhen express delivery occurs, safety problem also becomes abnormal pressing problem urgently to be resolved hurrily.

Because packaging is improper, traffic condition is not suitable for, receive while posting part the lack of standardization and courier of Visually Inspected and mishandle, may there are a lot of safety problems in the reasons such as mail, for example battery of mobile phone is met high in express delivery car transports processTemperature is blasted, chemicals is bright because revealing appears in packages in damaged condition, fuel tank leakage of oil causes assembling chance in compartmentFire is blasted etc. Occur that at express delivery car the scene of blast, spontaneous combustion also often appears at news the inside, seriousThreaten courier's life security and road safety, can cause reputation impact and economic loss to Express firm simultaneously.Especially the shopping such as " two 11 ", shop-establishment celebration, festivals or holidays concentrative time interval, express delivery amount is extremely large, and security risk is askedInscribe very outstanding. On November 12nd, 2015, in Jiangxi just there is the accident that express delivery car burns in prosperous paulownia, causesThe express delivery that is worth more than 100 ten thousand yuan is all burnt. Therefore design a kind of express delivery car danger early warning system, ensureTransporting safely of express delivery just becomes the task of top priority.

Now the safety guarantee of express delivery is mainly leaned on to the observation of driver to express delivery compartment, by virtue of experience come to determineWhether express delivery car compartment is in a safe condition, and the express delivery car combustion incident of for example prosperous paulownia in Jiangxi is exactly at first by drivingThe person of sailing finds that in rearview mirror trunk has flare, and general fire is in uncontrollable in this caseState, loses also very large. At present, some existing Express Logistics watch-dogs are only to temperature in compartmentHumidity and vehicle location are monitored, and issuable flammable explosive gas in driving process etc. are not carried outMonitoring, cannot accomplish the dangerous real-time early warning to occurring in transportation, to traffic safety also with regard to nothingMethod is accomplished effective guarantee. And in the numerous monitoring of environmental of number of sensors, effectively do not countAccording to fusion treatment, cannot avoid the burst data mistake that may occur, while transmitting a large amount of SDI,Data transmission efficiency is very low; In addition, real-time monitored data does not realize to be uploaded and cloud storage, entangles if produceRight-safeguarding evidence accurately confusingly cannot be provided.

To sum up, the shortcoming that Express Logistics monitoring technology exists is at present as follows:

(1) when express delivery car is storing and when peril of transportation material, may produce owing to happening suddenly some fortuitous eventRaw flammable explosive gas, for this situation, does not also realize real-time monitoring at present, thereby cannot be rowCar is made safely guarantee.

(2) driver cannot obtain real-time compartment monitored data, and can not make anticipation to unusual condition.

(3) also do not realize at present the Data Fusion to Real-time Monitoring Data, cannot be to occurringBurst data mistake is corrected, and the efficiency of transmission of data is very low in the time that data volume is larger.

(4) monitored data does not realize network timely and uploads and cloud storage, on the one hand when occurring abnormal conditionsShi Wufa realizes the quick and precisely processing of the departments such as fire-fighting; On the other hand, if cannot carry while producing customers' conflictSupply right-safeguarding evidence accurately.

Summary of the invention

First technical problem that the present invention will solve is to provide a kind of express delivery car danger early warning system, this systemAdopt environmental aspect and express delivery truck position letter in wireless sensor network technology Real-Time Monitoring express delivery car compartmentBreath, in the time there is abnormal data in express delivery car compartment, can and alarm.

In order to address the above problem, express delivery car danger early warning system comprises that the monitoring being arranged in express delivery car compartment is wholeEnd node, ZigBee telegon, ZigBee-3G/4G gateway module, Cloud Server, warning device; MonitoringAmbient parameter monitored data in the sensor group collection compartment of terminal node and the position monitoring number of express delivery carAccording to, and send to the ZigBee telegon being arranged in driver's cabin by wireless transmission method; ZigBee associationAdjust device that the monitored data of monitoring terminal node collection is gathered and sent to by Serial Port Line connected modeZigBee-3G/4G gateway module; After ZigBee-3G/4G gateway module is processed monitored data, more rightIt carries out data frame format conversion and the protocol conversion of Zigbee and 3G/4G, then by 3G/4G networkData message after conversion is uploaded to Cloud Server; Cloud Server receives by ZigBee-3G/4G gateway moduleThe data message sending is also stored in database; Meanwhile, ZigBee-3G/4G gateway module is by numberIt is believed that breath sends to warning device by Serial Port Line connected mode; Be arranged in the warning device judgement in driver's cabinWhether data message occurs extremely, if send alarm signal, is given warning message real-time Transmission simultaneouslyZigBee-3G/4G gateway module is also presented on display screen; The warning message that warning device sends in real time simultaneouslyBe transferred to Cloud Server; If data message normally by security information real-time Transmission to ZigBee-3G/4G gatewayModule is also presented on display screen.

It is right that described ZigBee-3G/4G gateway module utilizes the Kalman filtering of Pearson came relevant treatment to improve one's methodsMonitored data is carried out fusion treatment, then the data frame format that it is carried out to Zigbee and 3G/4G is changed and agreementConversion.

The present invention adopts environmental aspect and the express delivery in wireless sensor network technology Real-Time Monitoring express delivery car compartmentTruck position information, mainly pay close attention to express delivery car in transportation owing to occurring that abnormal conditions cause occurring dangerousThe concentration of gas change and compartment in the situation of change of smog, temperature, to abnormal in the compartment occurringState carries out phonic warning to driver in time, and by Data fusion technique by monitored data and early warning informationWith 3G/4G packet reach Cloud Server in form, store and supply and check.

Adopt the present invention, in the time there is abnormal data in express delivery car compartment, can carry out alarm to driver,And upload on network, take in time corresponding emergency treatment for network monitoring personnel; Can be by monitoringEnvironmental data information exchange cross 3G/4G signal and send to Cloud Server, generate monitoring journal, and with chartForm stores.

The present invention detects ambient parameter and positional information in express delivery car compartment, for possibility in compartmentThe dangerous substance occurring, when these materials run into emergency case in transportation, produces while occurring revealingHazardous gas and when burning the situation such as smog, high temperature that produces, carry out real-time monitoring, abnormal when occurringShi Jinhang audio alert, the situation such as the burning of express delivery article, blast that prevention may occur, ensures road safetyWith property safety; Data are sent to Cloud Server by 3G/4G network and store, for issuableDispute provides right-safeguarding evidence accurately, and provides the foundation of rescue promptly and accurately for fire department.

The sensor group of described monitoring terminal node comprises humiture integral sensor, smokescope sensor,Hazardous gas concentration sensor and express delivery truck position sensor; Humiture integral sensor, smokescope passIn sensor, hazardous gas concentration sensor and express delivery truck position sensor difference Real-time Collection express delivery car compartmentHumiture, smokescope, hazardous gas concentration and express delivery truck position information, and send it to ZigBeeTelegon.

Described hazardous gas concentration sensor comprises carbonomonoxide concentration monitoring sensor, concentration of alcohol monitoringWith sensor, methane concentration monitoring sensor; These sensor Real-time Collections by smokescope, an oxygenChange carbon gas concentration, ethanol gas concentration, concentration of methane gas are transferred to ZigBee telegon.

Second technical problem that the present invention will solve is to provide a kind of express delivery car danger early warning method, under this comprisesState step:

One, gather express delivery car ambient parameter monitored data and it is carried out to data processing;

Two, the data message obtaining after data processing is carried out the data frame format conversion of Zigbee and 3G/4GAnd protocol conversion, then judge that according to the data message after conversion whether express delivery car compartment is in a safe condition,If monitored data is presented on display screen in real time, otherwise output alarm information, simultaneously by warning messageBe presented on display screen and this warning message and car position information are uploaded to Cloud Server.

In described step 1, utilize the Kalman filtering of Pearson came relevant treatment to improve one's methods environmental data to enterRow fusion treatment, concrete steps are as follows:

1) establish for the each parameter index that will measure and adopt M sensor; By the difference of obtaining source dataSource data is divided into M group by source, is respectively SD₁、SD₂、……SD_i……、SD_M; By every group of source numberTime is that order is arranged according to this, and every N sampling instant source data is a sequence;

2) for SD_iGroup data, i ≠ 1 and i ≠ M, calculate SD_i-1Group data and SD_i+1The mark of group dataAccurate poor; If i=1, calculates SD_i+1Group and SD_i+2Group data standard is poor; If i=M, calculates SD_M-1Group and SD_M-2The standard deviation of group data; If standard deviation is not 0, go to step 3);

3) obtain SD according to formula (1-1), (1-2), (1-3), (1-4)_i-1Group data and SD_iGroup dataPearson correlation coefficient ρ_i-1,i，SD_iGroup data and SD_i+1The Pearson correlation coefficient ρ of group data_i,i+1：

${\mathrm{\ρ}}_{i-1,i}=\frac{E\left({\mathrm{SD}}_{i-1}{\mathrm{SD}}_i\right)-E\left({\mathrm{SD}}_{i-1}\right)E\left({\mathrm{SD}}_i\right)}{\sqrt{E\left({{\mathrm{SD}}_{i-1}}^2\right)-E^2\left({\mathrm{SD}}_{i-1}\right)}\sqrt{E\left({{\mathrm{SD}}_i}^2\right)-E^2\left({\mathrm{SD}}_i\right)}}---\left(1-1\right)$

$E\left({\mathrm{SD}}_i{\mathrm{SD}}_{i+1}\right)=\frac{\underset{n=1}{\overset{N}{\Σ}}\left({\mathrm{SD}}_{in}{\mathrm{SD}}_{(i+1)n}\right)}{N}---\left(1-2\right)$

${\mathrm{\ρ}}_{i,i+1}=\frac{E\left({\mathrm{SD}}_i{\mathrm{SD}}_{i+1}\right)-E\left({\mathrm{SD}}_i\right)E\left({\mathrm{SD}}_{i+1}\right)}{\sqrt{E\left({{\mathrm{SD}}_i}^2\right)-E^2\left({\mathrm{SD}}_i\right)}\sqrt{E\left({{\mathrm{SD}}_{i+1}}^2\right)-E^2\left({\mathrm{SD}}_{i+1}\right)}}---\left(1-3\right)$

$E\left({\mathrm{SD}}_i{\mathrm{SD}}_{i+1}\right)=\frac{\underset{n=1}{\overset{N}{\Σ}}\left({\mathrm{SD}}_{in}{\mathrm{SD}}_{(i+1)n}\right)}{N}---\left(1-4\right)$

Wherein ρ_i-1,iSD_i-1Group data and SD_iThe Pearson correlation coefficient of group data, ρ_i,i+1SD_iGroup numberAccording to SD_i+1The Pearson correlation coefficient of group data, E represents the mathematic expectaion of data group; N represents SD_iNumberAccording to sequence number according to time sequence in group, n=1,2 ..., N; SD_inRepresent SD_iIn group data, sequence number isThe source data of n;

4) judge SD_iWhether the Pearson correlation coefficient of group data and adjacent two group data is all less than 0.5,If conclude that abnormal interference appears in data, revises these group data, otherwise do not revise; If SD_iGroupData exception, when it is revised, first calculates the data SD in these group data k moment_{k_i}All with this momentThe difference DV of data mean value_{k_i}, then by difference DV_{k_i}Absolute value | DV_{k_i}| with the mark of these all data of momentAccurate poor S_kCompare, if | DV_{k_i}|>S_kAnd | DV_{k_i}|<2S_k, will extremely organize data correction for all numbers of this momentAccording to median and mean of mean; If | DV_{k_i}|>2S_k, will extremely organize data modification is all numbers of this momentAccording to median; If | DV_{k_i}|<S_k, do not revise; Then carry out data merging, calculate in this moment and repairThe average of all data just, and set it as the merging data in this moment; Merged through this stepAfter sample data Z (k), k=1,2 ..., N;

5) set up a discrete control procedure system model, this system is carried out with the linear random differential equationDescribe, shown in (2):

X(k)＝A·X(k-1)+W(2)

In formula (2), X (k) is expressed as the system state variables corresponding in the k moment, and X (k-1) is expressed as at k-1The system state variables that moment is corresponding, X (0)=Z (1); A is systematic parameter matrix, and W represents system noise;

6) utilize formula (3), (4), (5), (6), (7) to carry out the optimization observability estimate of system, obtainData message after fusion;

X(k|k-1)＝A×X(k-1|k-1)+W(3)

P(k|k-1)＝A×P(k-1|k-1)×A'+Q(4)

Kg(k)＝P(k|k-1)×H'/(H×P(k|k-1)×H')(5)

X(k|k)＝X(k|k-1)+Kg(K)×(Z(k)-H×X(k|k-1))(6)

P(k|k)＝(I-Kg(k)×H)×P(k|k-1)(7)

X in formula (k|k-1) is expressed as by previous moment state variable by the result of estimating to draw,X (k-1|k-1) is expressed as the observability estimate value that previous moment state variable has, and P (k|k-1) is X (k|k-1)Covariance value, P (k-1|k-1) is X (k-1|k-1) covariance value, matrix A ' represent the transposed matrix of A; Q tableShow the covariance value of system noise W; Kg (k) is kalman gain; H is complete 1 measurement parameter matrix,H ' is the transposed matrix of H; X (k|k) is the optimization observability estimate of k moment system state variables, P (k|k)For the covariance of X (k|k), coefficient I is complete 1 coefficient matrix;

Concrete estimation procedure is as follows:

When k=1, utilize respectively formula (3), (4) to calculate X (k|k-1), P (k|k-1); When X (0|0) is k=1The sample data Z (k) carving; P (0|0) gets the value between 0 to 1, here P (0|0)=0.5; Then profitCalculate kalman gain Kg (k) by formula (5), recycling formula (6) is calculated k moment system state variablesOptimization observability estimate X (k|k);

Utilize formula (7) to calculate the covariance P (k|k) of X (k|k);

The optimization observability estimate X (k|k) that utilization calculates and the covariance P (k|k) of X (k|k) are to formula (3), (4)In X (k-1|k-1), P (k-1|k-1) upgrade, next moment is calculated in recycling formula (5), (6), (7)Optimization observability estimate X (k|k) and the covariance P (k|k) of X (k|k); Carry out according to this recursive operation, until askGo out the observability estimate value X (k|k) in all k moment, i.e. data message after fusion treatment.

In described step 2, judge the method for the whether in a safe condition and output alarm information in express delivery car compartmentAs follows:

To the temperature receiving, smokescope, carbon monoxide gas concentration, ethanol gas concentration and methane gasBulk concentration data compare judgement; If smokescope data are greater than the smokescope threshold value of setting, and oneCarbonoxide concentration is greater than the carbonomonoxide concentration threshold value of setting, and output represents that the warning message of the condition of a fire appears in compartment;If smokescope data are greater than the smokescope threshold value of setting, and carbonomonoxide concentration is lower than an oxygen of settingChange carbon concentration threshold, output represents the warning message that compartment dust is too much;

In the smog data that receive lower than the smokescope threshold value of setting, if alcohol gas is denseDegrees of data is greater than the ethanol gas concentration threshold value of setting, and temperature is higher than the temperature threshold of setting, and exports representativeThere is the warning message of risk of explosion in compartment; If ethanol gas concentration is greater than the ethanol gas concentration threshold value of setting,And temperature is lower than the temperature threshold of setting, and output represents that compartment has fuel gas to reveal warning message, simultaneously defeatedGo out ambient parameter data and car position information;

Smokescope threshold value in smokescope data lower than setting, and ethanol gas concentration data are lower than settingThe situation of ethanol gas concentration threshold value under, dense if concentration of methane gas data are greater than the methane gas of settingDegree threshold value, and temperature data is higher than the temperature threshold of setting, and exports and represents that there is the alarm signal of risk of explosion in compartmentBreath; If concentration of methane gas data are greater than the concentration of methane gas threshold value of setting, and temperature data is lower than establishingFixed temperature threshold, output represents the warning message that compartment has fuel gas to reveal, simultaneously output environment parameterData and car position information;

Smokescope threshold value in smokescope data lower than setting, and ethanol gas concentration data are lower than settingEthanol gas concentration threshold value, simultaneously concentration of methane gas data are lower than the concentration of methane gas threshold value feelings of settingUnder condition, output environment supplemental characteristic and car position information.

Owing to existing a large amount of dissimilar sensors at one time physical index to be carried out to data in the present invention, therefore in system, there are a considerable amount of multi-source heterogeneous data in data acquisition. The number of nodes of sensor networkMore and be subject to external interference, occur that catastrophic failure is unavoidable; And the physical parameter of measuring possesses necessarilyComplexity and polytropy, therefore, there are careless mistake data or the larger data of error in actual measured value sometimes.By eliminating the above-mentioned systematic error of being discussed, the present invention carries out Kalman filtering algorithm in conjunction with Pearson came related operationImprovement, first utilize Pearson came related operation to revise the abnormal data in measurement data, then makeCarry out the fusion of monitored data with Kalman filtering algorithm.

The data anastomosing algorithm that the present invention adopts is that the Kalman filtering based on Pearson came relevant treatment is improved algorithm.Kalman filtering algorithm is that a kind of computer that is applicable to is processed and the recursion algorithm for estimating that carries out calculating in real time. ShouldThe recursion estimation principle of algorithm is: build input number according to this best estimate criterion of least mean-square error simultaneouslyAccording to the model of noise variable, and enter as basis taking the estimated value of previous moment and the observation of current timeOne step estimates the state variable value of current time, is the estimated value of current time state variable. NamelyThe essence of saying Kalman filtering is: the logical thinking with " estimating-actual measurement-corrigendum " carries out recursion.

Beneficial effect:

(1) realized the danger early warning that may occur in the transport of express delivery car, ensured that express delivery car transports safely.

(2) driver can obtain the service of monitored data real time inspection and audio alert service.

(3) adopt the Kalman filtering of Pearson came relevant treatment to improve algorithm, monitored data is carried out to data and meltClose processing, improved monitored data accuracy and reliability, reduced the data volume in transmitting procedure, improveEfficiency of transmission, reduced transfer of data cost.

(4) current network signal strength signal intensity is compared, by monitored data by the stronger 3G of signal or 4GNetwork is sent to Cloud Server storage, has expanded communication range, has improved security and the survivability of system.

(5) can be according to the analysis to monitored data, judge in compartment and whether occur extremely, if occur abnormalCarry out audio alert, thereby make driver can take measures in time the unusual condition occurring to process,Effectively to control the dangerous circumstances.

Brief description of the drawings

Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.

Fig. 1 express delivery car of the present invention danger early warning system global structure figure.

The structure chart of Fig. 2 monitoring terminal node.

The structure chart of Fig. 3 telegon.

Fig. 4 gateway structure chart.

Fig. 5 warning device structure chart.

Fig. 6 is main program flow chart of the present invention.

Fig. 7 is the functional flow diagram of warning device of the present invention.

Fig. 8 is the flow chart of data anastomosing algorithm of the present invention.

Data anastomosing algorithm result comparison diagram when Fig. 9 is noiseless.

Figure 10 has the time of interference data anastomosing algorithm result comparison diagram.

Figure 11 is the flow chart of data storage of the present invention.

Detailed description of the invention

As shown in Figure 1, express delivery car danger early warning system of the present invention comprises the monitoring being arranged in express delivery car compartmentTerminal node, ZigBee telegon, ZigBee-3G/4G gateway module, Cloud Server, warning device; PrisonSurvey terminal node and be arranged in express delivery car compartment, by sensor group, the ambient parameter in compartment and position are believedBreath etc. carries out periodic periodic monitor, and by wireless transmission method, the monitored data signal collecting is sent outGive the ZigBee being arranged in driver's cabin telegon; ZigBee telegon gathers monitoring terminal nodeMonitored data gathers and sends to ZigBee-3G/4G gateway module by Serial Port Line connected mode; ByZigBee-3G/4G gateway module carries out data fusion operation to the data message of monitoring, and to the number after mergingIt is believed that breath carries out data frame format conversion and the protocol conversion of Zigbee and 3G/4G, then by 3G/4G netThe monitored data information of processing is uploaded to Cloud Server by network; Cloud Server receives by ZigBee-3G/4G netThe monitored data that pass module sends is also stored in database; Meanwhile, ZigBee-3G/4G gateway mouldPiece sends to warning device by the monitored data after merging by Serial Port Line connected mode; Be arranged in driver's cabinWarning device monitored data information is presented on display screen in real time, and fast according to monitored data information judgementPass car compartment and whether occur extremely, if occur extremely, pass through voice and stand by lamp etc. to driver's alarm.Simultaneously by warning message real-time Transmission to ZigBee-3G/4G gateway module and be presented on display screen; Report to the police and fillPut the warning message while real-time Transmission of sending to Cloud Server; If data message is normally real by security informationIn time, is transferred to ZigBee-3G/4G gateway module and is presented on display screen.

1. monitoring terminal node

Monitoring terminal node structure is basic identical, mainly comprises sensor group and radio frequency transceiving module, asShown in Fig. 2. Sensor group comprises express delivery car vehicle environment parameter acquisition sensor group and express delivery truck position informationPick-up transducers; The monitored data of each sensor collection is sent to ZigBee by radio frequency transceiving moduleTelegon.

According to China's " fire hazard classification ", (GB/T4968-2008) fire is special according to the type of combustible and burningProperty can be divided into 6 classes. Category-A fire refers to solid matter fire, generally has the character of organic substance, organic matterMatter, once occur burning, can produce the toxic gas that carbon monoxide is Main Ingredients and Appearance, by temperature, smogAnd carbonomonoxide concentration can detect easily. Category-B fire refers to the solid that liquid maybe can meltMaterial fire, this type of material generally has volatility, and just can hold by the composition that detects gas in airChange places and monitored, when this type of substance combustion, generally can produce carbon monoxide, and volatile liquid refers generally toEthanol. C class fire refers to gas fire, and gas is not generally in the traffic coverage of express delivery car, but due to transportEmergency in process may produce a certain amount of hazardous gas and cause producing fire, now initiation fireGas refer generally to carbon monoxide and methane. D class belongs to hazardous metals fire, belongs to the model of hazardous materials transportationEnclose, general express delivery car can not transport this type of material, in the time there is dangerous situation in this type of material, by carbon monoxideDetect and just can judge the burning or the blast that whether occur this type of material. E class belongs to charged fire, express deliveryCar does not belong to this category. F class is the fire that the culinary art thing in cooking pot causes, also not in this research range.Actual conditions according to express delivery car in the time transporting article, the contingent fire of express delivery car is generally category-A, BWhether class and C class etc., therefore just can be judged and send out by the concentration that detects carbon monoxide, ethanol and methaneGive birth to fire.

During the sensor of monitoring terminal node is selected, the present invention has adopted 6 class sensors, to temperature, relative7 item numbers such as humidity, smog, carbonomonoxide concentration, concentration of alcohol, methane concentration, vehicle location are according to carrying outGather comparatively accurately. That express delivery car vehicle environment temperature sensor and humidity sensor adopt is DHT11Humiture integral sensor; What Smoke Sensor adopted is gas sensor MQ-2 sensor; Dangerous gasBody sensor be carbonomonoxide concentration monitoring CO-SFA-1000 sensor, concentration of alcohol monitor use beWhat C2H5OH-CA1000 sensor, methane concentration were monitored use is CH4-D5W sensor; Express delivery parking stallWhat put information gathering employing is SiRFstarIIIGPS alignment sensor. Monitoring terminal node adopts CC2530Radio transceiver chip, the monitored data that each sensor gathers by radio frequency transceiving module with wireless biographyDefeated mode is transferred to ZigBee telegon.

Fig. 2 is the structural representation of monitoring terminal node, sensor group to the ambient parameter in express delivery car compartment andVehicle position information accurately gathers, and the external data that A/D modular converter is responsible for sensor to transmit is carried outA/D conversion becomes analog signal after data-signal and sends in MCU processing module and process, and it is alsoA data channel. MCU processing module is the core of single-chip microcomputer, and its function is monitored data to carry outPreliminary screening analysis and processing, be packaged into Frame by data according to Zigbee protocol, transmits by interfaceGive ZigBee communication module. ZigBee communication module mainly completes the function of communicating by letter in ZigBee mode.ZigBee communication is a kind of low-power consumption, low rate, and high reliability, cheapness, networking mode easily, and anti-Answer speed fast, can ensure the real-time monitoring to express delivery car compartment, its coverage can reach several when without gainTen meters, meet the required distance of vehicle monitoring completely.

2. telegon

The CC2530 radio frequency transceiving module that telegon part adopts TI company to produce. Telegon converges jointThe nucleus module of point is Zigbee communication module, and this module utilizes antenna and monitoring terminal node to carry out between dataIntercommunication mutually, comprise from monitoring terminal node and receive the ambient parameter data of uploading, send acquisition instructions arriveMonitoring terminal node, ensures the normal data communication of carrying out with Zigbee-3G/4G gateway module. All the other also haveCentral processing unit, auxiliary storage and the processing to data of memory module, concrete structure designs as Fig. 3, in addition,The function of telegon also comprises networking and data receiver.

3.ZigBee-3G/4G gateway module

As shown in Figure 4, ZigBee-3G/4G gateway module comprises STM32F103RCT6 type single-chip microcomputer, ChinaFor MC509CDMA2000 communication module and the ME909s-821 of Huawei communication module.STM32F103RCT6 type single-chip microcomputer is responsible for data receiver and processing, and the MC509s of Huawei communication module is responsible for3G communication, Huawei's ME909s communication module is responsible for 4G communication.

The monitored data informational needs receiving from telegon utilizes gateway to carry out data fusion, to improve monitoring numberTransmit cost according to accuracy and reduction data; The information mutual communication of wireless sensor network and 3G/4G network needsUtilize gateway, make information format in gateway, convert the frame format of 3G/4G to from the frame format of ZigBee.STM32F103RCT6 type single-chip microcomputer in ZigBee-3G/4G gateway module is by the prison receiving from telegonSurvey data message and carry out data fusion, and the data message after merging is sent to report by serial ports connected modeAlarm device. Afterwards, the STM32F103RCT6 type single-chip microcomputer in ZigBee-3G/4G gateway module detects and works asFront network signal intensity, if 3G signal is stronger, carries out protocol conversion and frame by the data message after mergingFormat conversion, is sent in MC509 communication module, and MC509 communication module will be monitored by 3G networkData message sends to Cloud Server and stores; If 4G signal is stronger, by the data message after mergingCarry out protocol conversion and frame format conversion, be sent in ME909s communication module MC909s communication moduleBy 4G network, monitored data information being sent to Cloud Server stores.

4. warning device

Warning device is by LED display, microprocessor and jingle bell installation composition, as shown in Figure 5. Report to the police and fillIt is dense that microprocessor in putting is responsible for receiving environment temperature, relative air humidity, smog that gateway module sendsDegree, carbonomonoxide concentration, concentration of alcohol, methane concentration data, be presented at the visible LED of driver and showOn screen, and itself and the safe range of setting are compared to judgement. If monitored data information is in safe rangeWithin, warning device is not pointed out, if monitored data information exceeds safe range, by jingle bell device toDriver's voice broadcast information or warning information.

5. Cloud Server

Cloud Server is mainly realized data storage function. Cloud Server detects and interrupts shape after operating environment is initializedThe state of state register, checks whether COM1 receives data receiver request, if do not detect,Return and continue inquiry, if find request, information source is verified, and returned and reply instruction, then joinPut the memory location of data in server, extract afterwards data and write in the memory of server.

Fig. 6 is main program flow chart of the present invention. Telegon is set up network; Monitoring terminal node search for networks,If can't find network, monitoring terminal node will continue search, until discovering network; After discovering network, prisonThe various kinds of sensors of surveying on terminal node is believed the ambient parameter in monitored area, express delivery car compartment, vehicle locationBreath etc. carries out Real-time Collection exactly; Monitoring terminal node sends to the data message collecting to be arranged in and drivesSail the ZigBee telegon of chamber; Telegon sends to ZigBee-3G/4G gateway by serial ports by data messageModule; The data message receiving is carried out data fusion behaviour by single-chip microcomputer in ZigBee-3G/4G gateway moduleDo, fuse information is sent to warning device by Serial Port Line connected mode; Warning device receives after fusionData message, judge that whether data abnormal, if data are normal, data message is presented on display screen;If data occur abnormal, on display screen, represent data exception, and pass through language according to the scope of data exceptionSound carries out alarm to driver; ZigBee-3G/4G gateway module judges network signal intensity at that time, selects letterNumber stronger 3G or 4G network, carries out frame format conversion and protocol conversion to fuse information, and data are sentGive Cloud Server; Cloud Server receives the data of gateway, and the information receiving is stored in database.

As shown in Figure 7, warning device to the temperature receiving, smokescope, carbon monoxide gas concentration,Ethanol gas concentration and concentration of methane gas data compare judgement; If express delivery car vehicle environment safe temperatureScope is below 70 DEG C, and the safe range of smokescope is below 10ppm, carbon monoxide gas concentrationSafe range is below 12.5%, and the safe range of ethanol gas concentration is below 3.3%, concentration of methane gasSafe range be below 5%; If the smokescope data that receive show express delivery car vehicle environment smogConcentration is greater than 10ppm, and the carbonomonoxide concentration data that receive show that in express delivery car compartment, carbon monoxide is denseDegree is greater than 12.5%, shows to have occurred the condition of a fire in express delivery car compartment, and burning has produced a large amount of carbon monoxide, reportJingle bell device in alarm device is play the voice of " condition of a fire appears in compartment, please notes the fire extinguishing of stopping "; If connectThe smokescope data received show that express delivery car vehicle environment smokescope is greater than 10ppm, and receive oneCarbonoxide concentration data shows that in express delivery car compartment, carbonomonoxide concentration, lower than 12.5%, shows express delivery car compartmentMiddle smokescope exceeds standard, but do not have on fire, dust excessive interference working sensor, the sound in warning deviceBell device is play the voice of " compartment dust is too much, please notes cleaning ".

In the smog data that receive lower than 10ppm, if the ethanol gas concentration number receivingExpress delivery car vehicle environment ethanol gas concentration is greater than 3.3% according to the show, and the temperature data receiving shows express deliveryCar vehicle environment temperature is higher than 70 DEG C, and showing has alcohol gas to reveal in express delivery car compartment, and has blast windDanger, the jingle bell device in warning device is play the voice of " there is risk of explosion in compartment, please notes strick precaution "; AsThe ethanol gas concentration data that fruit receives show that express delivery car vehicle environment ethanol gas concentration is greater than 3.3%, andThe temperature data receiving shows that express delivery car vehicle environment temperature is lower than 70 DEG C, and showing has second in express delivery car compartmentAlcohol gas leakage, " compartment has fuel gas to reveal to the jingle bell device broadcasting in warning device, please notes ventilationCheck " voice, information relevant to the ambient parameter data and the car position etc. that record is uploaded to cloud simultaneouslyServer.

In the smog data that receive, lower than 10ppm, and the ethanol gas concentration data that receive are lower than 3.3%Situation under, if the concentration of methane gas data that receive show express delivery car vehicle environment concentration of methane gasBe greater than 5%, and the temperature data receiving shows that express delivery car vehicle environment temperature, higher than 70 DEG C, shows express deliveryIn car compartment, have methane gas to reveal, and have risk of explosion, the jingle bell device in warning device is play " carThere is risk of explosion in railway carriage or compartment, please notes strick precaution " voice; If the concentration of methane gas data that receive show fastPass car vehicle environment concentration of methane gas and be greater than 5%, and the temperature data receiving shows express delivery car compartment ringBorder temperature is lower than 70 DEG C, and showing has methane gas to reveal in express delivery car compartment, the jingle bell device in warning devicePlay the voice of " compartment has fuel gas to reveal, and please noting ventilates checks ", simultaneously by the environment ginseng recordingThe information that logarithmic data and car position etc. are relevant uploads to Cloud Server.

In the smog data that receive, lower than 10ppm, and ethanol gas concentration data are lower than 3.3%, connect simultaneouslyThe concentration of methane gas data of receiving, lower than in 5% situation, illustrate that in express delivery car compartment, ambient parameter is normal,Do not occur security risk, warning device is not reported to the police; Complete receive data relatively after, display screen all canThe data message receiving is shown, check at any time the ambient parameter data simultaneously recording for driverAnd the relevant information such as car position uploads to Cloud Server.

In the middle of the present invention moves, measure compared with the parameter index of multiple types owing to the present invention relates to, and eachThe principle that supplemental characteristic merges is close, below only with environment temperature in the compartment in measurement parameter index thisOne index is example, and the data anastomosing algorithm set forth is before carried out emulation testing and its result is analyzed.In the time that the measurement of environment temperature adopts 5 sensors to measure, choose the each monitoring of 10 continuous moment eventuallyThe temperature value data sample that end node collects, as shown in table 1.

Table 1 monitoring terminal node image data sample

Temperature (DEG C)	Sensor 1	Sensor 2	Sensor 3	Sensor 4	Sensor 5
						t1	20.0	19.8	20.1	20.2	19.9
t2	20.0	19.8	20.1	20.2	20.0
						t3	20.1	20.0	20.2	20.2	20.0
t4	20.0	19.9	20.1	20.3	19.9
						t5	20.0	19.8	20.1	20.2	19.9
t6	20.0	19.8	20.1	20.2	19.9
						t7	19.9	19.7	20.0	20.2	19.9
t8	20.0	19.8	20.1	20.0	19.8
						t9	19.9	19.8	20.0	20.1	19.9
t10	20.0	19.8	20.1	20.2	19.9

As shown in Figure 8, Data Fusion method is the Kalman filtering based on the processing of Pearson came related operationImprove algorithm, first utilize Pearson came related operation to revise the abnormal data in measurement data, thenUse Kalman filtering algorithm to carry out data fusion, concrete steps are as follows:

1) be directed to the each parameter index that will measure (taking temperature as example), by detect source data with the timeFor order is arranged, and by the separate sources that obtains source data, data are divided into groups, press and pass by data in table 1Sensor is divided into 5 groups, is respectively T₁、T₂、T₃、T₄、T₅, every group of source data arranged taking the time as order,Every 5 sampling instant temperature datas are as a sequence; In this sequence, the sequence number of temperature data isn＝1,2，……，5。

2) for arbitrary T in sequence_iGroup data, i ≠ 1 and i ≠ 5, calculate its last group of (T_i-1Group) dataThereafter one group of data (T_i+1Group) standard deviation of data, if standard deviation is not 0, can carry out ensuingProcess; If i=1, calculates T_i+1Group and T_i+2Group data standard is poor (is T₂And T₃The standard deviation of group data),If i=5, calculates T_i-1Group and T_i-2The standard deviation of group data (is T₄And T₃The standard deviation of group data). MarkThe degree of fluctuation of the accurate poor data that represent data group, standard deviation is not that 0 explanation data fluctuations is larger, having canCan there are abnormal conditions, need to further judge. I=1,2,3,4,5, represent data group sequence number;

3) if standard deviation is not 0, by T_iGroup data and another two groups of data (T_i-1Group data and T_i+1Group numberAccording to) carry out Pearson came related operation, obtain respectively Pearson correlation coefficient, concrete calculate suc as formula (1-1),(1-2), shown in (1-3), (1-4):

${\mathrm{\&rho;}}_{i-1,i}=\frac{E\left(T_{i-1}{T}_i\right)-E\left(T_{i-1}\right)E\left(T_i\right)}{\sqrt{E\left({T_{i-1}}^2\right)-E^2\left(T_{i-1}\right)}\sqrt{E\left({T_i}^2\right)-E^2\left(T_i\right)}}---\left(1-1\right)$

$E\left(T_i{T}_{i+1}\right)=\frac{\underset{n=1}{\overset{N}{\&Sigma;}}\left(T_n{T}_{(i+1)n}\right)}{N}---\left(1-2\right)$

${\mathrm{\&rho;}}_{i,i+1}=\frac{E\left(T_i{T}_{i+1}\right)-E\left(T_i\right)E\left(T_{i+1}\right)}{\sqrt{E\left({T_i}^2\right)-E^2\left(T_i\right)}\sqrt{E\left({T_{i+1}}^2\right)-E^2\left(T_{i+1}\right)}}---\left(1-3\right)$

$E\left(T_i{T}_{i+1}\right)=\frac{\underset{n=1}{\overset{N}{\&Sigma;}}\left(T_{in}{T}_{(i+1)n}\right)}{N}---\left(1-4\right)$

ρ in formula (1)_i-1,iT_i-1Group data and T_iThe Pearson correlation coefficient of group data, ρ_i,i+1T_iGroupData and T_i+1The Pearson correlation coefficient of group data, is used for representing two linear correlation journeys between data groupDegree, is correlated with between data group under normal circumstances, and Pearson's coefficient is greater than 0.5; E represents the number of data groupTerm is hoped (for example E (T_i) mathematic expectaion be data group T_iAverage; E (T_iT_i+1) expression T_iGroup data withT_i+1The average of group data product); Wherein, n represents T_iSequence number according to time sequence in data group, n=1 here,2,3,4,5; N represents T_iThe number of data in data group, N=5 here, T_inRepresent T_iIn group data groupSequence number is the temperature data of n;

4) judge T_iWhether the Pearson correlation coefficient of group data and adjacent two group data is all less than 0.5, ifTo conclude that abnormal interference appears in data, need to carry out data correction, otherwise needn't revise; To extremelyT_iWhen group data correction, for arbitrary moment k, first calculate T_iOrganize the data T in this moment_{k_i}All with this momentThe difference DV of the average of data (from different groups)_{k_i}(T), then by the absolute value of difference | DV_{k_i}(T) | with the standard deviation S comparison of these all data of moment, if | DV_{k_i}(T) | > S and | DV_{k_i}(T)│ < 2S, by abnormal T_iGroup data correction is all data medians of this moment and mean of mean; If |DV_{k_i}(T) | > 2S, by abnormal T_iGroup data modification is the median of these all data of moment; If |DV_{k_i}(T) | < S, does not revise; Then carry out data merging, calculate in revised institute of this momentThere is the average of data (from different groups), and set it as the merging data in this moment; Through this stepObtain the sample data Z (k) after merging, k=1,2,3,4,5;

5) set up a discrete control procedure system model, this system is described with linear random differentialShown in (2):

X(k)＝A·X(k-1)+B·U(k-1)+W(2)

In formula (2), X (k) is the system state variables corresponding in the k moment, and X (k-1) is expressed as in the time of k-1Carve corresponding system state variables, X (0)=Z (1); What coefficient A and coefficient B represented is systematic parameter,The system model using is during for Multi-model System, and the value of A and B is coefficient matrix, in the present invention, adoptsThe data physical quantity of collection has identical rule with the data physical quantity of previous moment, and therefore can obtain A is unitMatrix, the coefficient A in equation is 1; U (k-1) is expressed as the system control amount corresponding in the k-1 moment, thisIn invention, system is not controlled, U (k-1) is 0; W is system noise;

6) the optimization observation that utilizes the covariance value of the system state variables of built erection system to carry out system is estimatedMeter. Computational process is: the state that estimates system current time according to the state variable of system previous momentVariable X (k|k-1); Utilize the covariance of previous moment state variable covariance value to state variable X (k|k-1)P (k-1|k-1) estimates; According to state variable estimated value, covariance estimated value and revised sample dataZ (k), can obtain X (k|k) and kalman gain Kg (k), obtains the state optimization observability estimate value of moment kX (k|k); Specific formula for calculation is as follows:

X(k|k-1)＝A×X(k-1|k-1)+B×U(k-1)+W(3)

P(k|k-1)＝A×P(k-1|k-1)×A'+Q(4)

Kg(k)＝P(k|k-1)×H'/(H×P(k|k-1)×H')(5)

X(k|k)＝X(k|k-1)+Kg(K)×(Z(k)-H×X(k|k-1))(6)

P(k|k)＝(I-Kg(k)×H)×P(k|k-1)(7)

X in formula (k|k-1) is expressed as by previous moment state variable by the result of estimating to draw,X (k-1|k-1) is expressed as the best observability estimate value that previous moment state variable has, and P (k|k-1) isX (k|k-1) covariance value, P (k-1|k-1) is X (k-1|k-1) covariance value, matrix A ' represent the transposition square of ABattle array, the value of A' is 1; Q represents the covariance value of system noise W; Kg (k) is kalman gain; CoefficientH is expressed as the parameter relevant to measurement of place Study system, due to the sensor adopting in the present inventionBe all that actual measured value is directly corresponding with the physical quantity that needs to measure, therefore H is complete 1 matrix, isNumber H values are that 1, H ' is that the value of the transposed matrix H ' of H is 1; X (k|k) is k moment system state variablesOptimization observability estimate, P (k|k) is the covariance of X (k|k), coefficient I is complete 1 coefficient matrix; ItsRemaining parameter meaning is identical with formula (2);

Concrete estimation procedure is as follows:

When k=1, utilize respectively formula (3), (4) to calculate X (k|k-1), P (k|k-1); The present invention is by dataThe sample data Z (k) in group k=1 moment is assigned to X (0|0); When carrying out data fusion along with algorithm, shapeState variable will be tending towards convergence, therefore for covariance value, generally get value between 0 to 1 all can, thisIn to get P (0|0) be 0.5; Then utilize formula (5) to calculate kalman gain Kg (k), recycling formula (6)Calculate the optimization observability estimate X (k|k) of k moment system state variables;

Utilize formula (7) to calculate the covariance P (k|k) of X (k|k);

The optimization observability estimate X (k|k) that utilization calculates and the covariance P (k|k) of X (k|k) are to formula (3), (4)In X (k-1|k-1), P (k-1|k-1) upgrade, next moment is calculated in recycling formula (5), (6), (7)Optimization observability estimate X (k|k) and the covariance P (k|k) of X (k|k); Carry out according to this recursive operation, until askGo out the observability estimate value X (k|k) in all k moment, i.e. data message after fusion treatment.

Algorithm substitutes actual detected value (data message obtaining after merging) with observability estimate value, makes dataAmount reduces, and improves accuracy rate, and algorithm finishes.

When stable operation of the present invention, 5 groups of data that the monitoring terminal node shown in table 1 gathers are passed through based on skinThe Kalman filtering of your inferior relevant treatment is improved algorithm and is carried out data fusion, and by the actual detected value before mergingCompare with the fluctuation situation of the observability estimate value after fusion, as shown in Figure 9. As can be seen from Figure 9,In glitch-free situation, data anastomosing algorithm of the present invention can be preferably by data on the same group not on a time periodMerge, reduced data volume, fusion results can be found out trend and feature and the source number of source data simultaneouslyMore steady according to comparing, embody the high-transmission efficiency of the data anastomosing algorithm of the present invention's design.

Because the express delivery car danger early warning system and method for the present invention's design has opening, in some special feelingsUnder condition, also there will be some inevitable external interference or system itself sensor and causing for a certain reasonSystem interference. In this case, the source data of systematic survey may occur more careless mistake data orInterfering data. For detecting the anti-interference of data anastomosing algorithm of the present invention, by the monitoring terminal shown in table 15 groups of data that node gathers are added random disturbances, then carry out Data Fusion, and by the interpolation before mergingThe fluctuation situation of the observability estimate value after the actual detected value disturbing and fusion compares, as shown in figure 10.As can be seen from Figure 10, in noisy situation, data anastomosing algorithm of the present invention can be preferably by notData on the same group merge on a time period, have reduced data volume, random disturbances can be carried out to certain disappearingRemove, can find out that the trend of source data and feature make fusion results mild, have embodied design of the present invention simultaneouslyHigh-transmission efficiency, high accuracy and the anti-interference of data anastomosing algorithm.

As shown in figure 11, Cloud Server is mainly realized data storage function. Cloud Server is initial to operating environmentAfter changing, detect the state of interrupt status register, check whether COM1 receives data receiver request, ifDo not detect, return and continue inquiry, if find request, information source is verified, and returned and answerAnswer instruction, the memory location of configuration data in server then, extracts afterwards data and writes serverIn memory.

Claims (7)

1. an express delivery car danger early warning system, is characterized in that comprising the monitoring being arranged in express delivery car compartmentTerminal node, ZigBee telegon, ZigBee-3G/4G gateway module, Cloud Server, warning device; PrisonAmbient parameter monitored data in the sensor group collection compartment of survey terminal node and the position monitoring number of express delivery carAccording to, and send to the ZigBee telegon being arranged in driver's cabin by wireless transmission method; ZigBee associationAdjust device that the monitored data of monitoring terminal node collection is gathered and sent to by Serial Port Line connected modeZigBee-3G/4G gateway module; After ZigBee-3G/4G gateway module is processed monitored data, more rightIt carries out data frame format conversion and the protocol conversion of Zigbee and 3G/4G, then by 3G/4G networkData message after conversion is uploaded to Cloud Server; Cloud Server receives by ZigBee-3G/4G gateway moduleThe data message sending is also stored in database; Meanwhile, ZigBee-3G/4G gateway module is by numberIt is believed that breath sends to warning device by Serial Port Line connected mode; Be arranged in the warning device judgement in driver's cabinWhether data message occurs extremely, if send alarm signal, is given warning message real-time Transmission simultaneouslyZigBee-3G/4G gateway module is also presented on display screen; The warning message that warning device sends in real time simultaneouslyBe transferred to Cloud Server; If data message normally by security information real-time Transmission to ZigBee-3G/4G gatewayModule is also presented on display screen.

2. express delivery car danger early warning system according to claim 1, described in it is characterized in thatZigBee-3G/4G gateway module utilizes the Kalman filtering of Pearson came relevant treatment to improve one's methods to monitoring numberAccording to carrying out fusion treatment, then it is carried out to data frame format conversion and the protocol conversion of Zigbee and 3G/4G.

3. express delivery car danger early warning system according to claim 1, is characterized in that described monitoring terminalThe sensor group of node comprises humiture integral sensor, smokescope sensor, and hazardous gas concentration passesSensor and express delivery truck position sensor; Humiture integral sensor, smokescope sensor, hazardous gasHumiture, smog in concentration sensor and express delivery truck position sensor difference Real-time Collection express delivery car compartment are denseDegree, hazardous gas concentration and express delivery truck position information, and send it to ZigBee telegon.

4. express delivery car danger early warning system according to claim 3, is characterized in that described hazardous gasConcentration sensor comprises that carbonomonoxide concentration monitoring sensor, concentration of alcohol monitoring sensor, methane are denseDegree monitoring sensor; These sensor Real-time Collections by smokescope, carbon monoxide gas concentration, secondAlcohol gas concentration, concentration of methane gas are transferred to ZigBee telegon.

5. an express delivery car danger early warning method, is characterized in that comprising the steps:

One, gather express delivery car ambient parameter monitored data and it is carried out to data processing;

Two, the data message obtaining after data processing is carried out the data frame format conversion of Zigbee and 3G/4GAnd protocol conversion, then judge that according to the data message after conversion whether express delivery car compartment is in a safe condition,If monitored data is presented on display screen in real time, otherwise output alarm information, simultaneously by warning messageBe presented on display screen and this warning message and car position information are uploaded to Cloud Server.

6. express delivery car danger early warning method according to claim 5, is characterized in that in described step 1,The step of utilizing the Kalman filtering of Pearson came relevant treatment to improve one's methods environmental data is carried out to fusion treatment asUnder:

1) establish for the each parameter index that will measure and adopt M sensor; By the difference of obtaining source dataSource data is divided into M group by source, is respectively SD₁、SD₂、……SD_i……、SD_M; By every group of source numberTime is that order is arranged according to this, and every N sampling instant source data is a sequence;

2) for SD_iGroup data, i ≠ 1 and i ≠ M, calculate SD_i-1Group data and SD_i+1The mark of group dataAccurate poor; If i=1, calculates SD_i+1Group and SD_i+2Group data standard is poor; If i=M, calculates SD_M-1Group and SD_M-2The standard deviation of group data; If standard deviation is not 0, go to step 3);

3) obtain SD according to formula (1-1), (1-2), (1-3), (1-4)_i-1Group data and SD_iGroup dataPearson correlation coefficient ρ_i-1,i，SD_iGroup data and SD_i+1The Pearson correlation coefficient ρ of group data_i,i+1：

${\mathrm{\&rho;}}_{i-1,i}=\frac{E\left({\mathrm{SD}}_{i-1}{\mathrm{SD}}_i\right)-E\left({\mathrm{SD}}_{i-1}\right)E\left({\mathrm{SD}}_i\right)}{\sqrt{E\left({{\mathrm{SD}}_{i-1}}^2\right)-E^2\left({\mathrm{SD}}_{i-1}\right)}\sqrt{E\left({{\mathrm{SD}}_i}^2\right)-E^2\left({\mathrm{SD}}_i\right)}}---(1-1)$

$E\left({\mathrm{SD}}_i{\mathrm{SD}}_{i+1}\right)=\frac{\underset{n=1}{\overset{N}{\&Sigma;}}\left({\mathrm{SD}}_{in}{\mathrm{SD}}_{(i+1)n}\right)}{N}---(1-2)$

${\mathrm{\&rho;}}_{i,i+1}=\frac{E\left({\mathrm{SD}}_i{\mathrm{SD}}_{i+1}\right)-E\left({\mathrm{SD}}_i\right)E\left({\mathrm{SD}}_{i+1}\right)}{\sqrt{E\left({{\mathrm{SD}}_i}^2\right)-E^2\left({\mathrm{SD}}_i\right)}\sqrt{E\left({{\mathrm{SD}}_{i+1}}^2\right)-E^2\left({\mathrm{SD}}_{i+1}\right)}}---(1-3)$

$E\left({\mathrm{SD}}_i{\mathrm{SD}}_{i+1}\right)=\frac{\underset{n=1}{\overset{N}{\&Sigma;}}\left({\mathrm{SD}}_{in}{\mathrm{SD}}_{(i+1)n}\right)}{N}---(1-4)$

Wherein ρ_i-1,iSD_i-1Group data and SD_iThe Pearson correlation coefficient of group data, ρ_i,i+1SD_iGroup numberAccording to SD_i+1The Pearson correlation coefficient of group data, E represents the mathematic expectaion of data group; N represents SD_iNumberAccording to sequence number according to time sequence in group, n=1,2 ..., N; SD_inRepresent SD_iIn group data, sequence number isThe source data of n;

4) judge SD_iWhether the Pearson correlation coefficient of group data and adjacent two group data is all less than 0.5,If conclude that abnormal interference appears in data, revises these group data, otherwise do not revise; If SD_iGroupData exception, when it is revised, first calculates the data SD in these group data k moment_{k_i}All with this momentThe difference DV of data mean value_{k_i}, then by difference DV_{k_i}Absolute value | DV_{k_i}| with the mark of these all data of momentAccurate poor S_kCompare, if | DV_{k_i}|>S_kAnd | DV_{k_i}|<2S_k, will extremely organize data correction for all numbers of this momentAccording to median and mean of mean; If | DV_{k_i}|>2S_k, will extremely organize data modification is all numbers of this momentAccording to median; If | DV_{k_i}|<S_k, do not revise; Then carry out data merging, calculate in this moment and repairThe average of all data just, and set it as the merging data in this moment; Merged through this stepAfter sample data Z (k), k=1,2 ..., N;

5) set up a discrete control procedure system model, this system is carried out with the linear random differential equationDescribe, shown in (2):

X(k)＝A·X(k-1)+W(2)

In formula (2), X (k) is expressed as the system state variables corresponding in the k moment, and X (k-1) is expressed as at k-1The system state variables that moment is corresponding, X (0)=Z (1); A is systematic parameter matrix, and W represents system noise;

6) utilize formula (3), (4), (5), (6), (7) to carry out the optimization observability estimate of system, obtainData message after fusion;

X(k|k-1)＝A×X(k-1|k-1)+W(3)

P(k|k-1)＝A×P(k-1|k-1)×A'+Q(4)

Kg(k)＝P(k|k-1)×H'/(H×P(k|k-1)×H')(5)

X(k|k)＝X(k|k-1)+Kg(K)×(Z(k)-H×X(k|k-1))(6)

P(k|k)＝(I-Kg(k)×H)×P(k|k-1)(7)

X in formula (k|k-1) is expressed as by previous moment state variable by the result of estimating to draw,X (k-1|k-1) is expressed as the observability estimate value that previous moment state variable has, and P (k|k-1) is X (k|k-1)Covariance value, P (k-1|k-1) is X (k-1|k-1) covariance value, matrix A ' represent the transposed matrix of A; Q tableShow the covariance value of system noise W; Kg (k) is kalman gain; H is complete 1 measurement parameter matrix,H ' is the transposed matrix of H; X (k|k) is the optimization observability estimate of k moment system state variables, P (k|k)For the covariance of X (k|k), coefficient I is complete 1 coefficient matrix;

Estimation procedure is as follows:

When k=1, utilize respectively formula (3), (4) to calculate X (k|k-1), P (k|k-1); When X (0|0) is k=1The sample data Z (k) carving; P (0|0) gets the value between 0 to 1, here P (0|0)=0.5; Then profitCalculate kalman gain Kg (k) by formula (5), recycling formula (6) is calculated k moment system state variablesOptimization observability estimate X (k|k);

Utilize formula (7) to calculate the covariance P (k|k) of X (k|k);

The optimization observability estimate X (k|k) that utilization calculates and the covariance P (k|k) of X (k|k) are to formula (3), (4)In X (k-1|k-1), P (k-1|k-1) upgrade, next moment is calculated in recycling formula (5), (6), (7)Optimization observability estimate X (k|k) and the covariance P (k|k) of X (k|k); Carry out according to this recursive operation, until askGo out the observability estimate value X (k|k) in all k moment, i.e. data message after fusion treatment.

7. express delivery car danger early warning method according to claim 5, is characterized in that in described step 2,Judge that whether in a safe condition express delivery car compartment method and output alarm information be as follows:

To the temperature receiving, smokescope, carbon monoxide gas concentration, ethanol gas concentration and methane gasBulk concentration data compare judgement; If smokescope data are greater than the smokescope threshold value of setting, and oneCarbonoxide concentration is greater than the carbonomonoxide concentration threshold value of setting, and output represents that the warning message of the condition of a fire appears in compartment;If smokescope data are greater than the smokescope threshold value of setting, and carbonomonoxide concentration is lower than an oxygen of settingChange carbon concentration threshold, output represents the warning message that compartment dust is too much;

In the smog data that receive lower than the smokescope threshold value of setting, if alcohol gas is denseDegrees of data is greater than the ethanol gas concentration threshold value of setting, and temperature is higher than the temperature threshold of setting, and exports representativeThere is the warning message of risk of explosion in compartment; If ethanol gas concentration is greater than the ethanol gas concentration threshold value of setting,And temperature is lower than the temperature threshold of setting, and output represents that compartment has fuel gas to reveal warning message, simultaneously defeatedGo out ambient parameter data and car position information;

Smokescope threshold value in smokescope data lower than setting, and ethanol gas concentration data are lower than settingThe situation of ethanol gas concentration threshold value under, dense if concentration of methane gas data are greater than the methane gas of settingDegree threshold value, and temperature data is higher than the temperature threshold of setting, and exports and represents that there is the alarm signal of risk of explosion in compartmentBreath; If concentration of methane gas data are greater than the concentration of methane gas threshold value of setting, and temperature data is lower than establishingFixed temperature threshold, output represents the warning message that compartment has fuel gas to reveal, simultaneously output environment parameterData and car position information;

Smokescope threshold value in smokescope data lower than setting, and ethanol gas concentration data are lower than settingEthanol gas concentration threshold value, simultaneously concentration of methane gas data are lower than the concentration of methane gas threshold value feelings of settingUnder condition, output environment supplemental characteristic and car position information.