CN106993319A - A kind of algorithm for realizing the switching of GNSS and WIFI systems seamless vertical - Google Patents

Abstract

The present invention be to provide it is a kind of based on fuzzy neural network algorithm come realize GNSS and WIFI systems seamless vertical switching algorithm, to reach that two different alignment systems are mutually merged, realize the purpose of seamless switching indoor and outdoor location navigation.The present invention comprises the following steps：The acquisition of main location information, based on the handover trigger for receiving signal, the handoff algorithms judgement based on Fuzzy Neural-network Control.The beneficial effects of the invention are as follows：The accuracy that WIFI is positioned when overcoming outdoor declines, and GNSS signal may occur to interrupt the deficiency for providing user location of mistake information when indoor, GPS of America, Russia Glonass, European Union Galileo and the big-dipper satellite of China can be made indoors, in the environment such as outdoor and underground, earth's surface, major step is strided forward in terms of satellite fix, navigation application technological orientation.Meanwhile, all have been widely used in fields such as China's military use, International Terrorism, Tunnel testings.

Description

A kind of algorithm for realizing the switching of GNSS and WIFI systems seamless vertical

Technical field

The invention belongs to satellite fix, navigation application technical field, and in particular to one kind is based on fuzzy neural network algorithm To realize the algorithm of GNSS and WIFI systems seamless vertical switching.

Background technology

At present, with the continuous progress of science and technology, people are increasing to the demand of positional information, whether military The fields such as purposes, aircraft navigation, naval vessels ship, earthquake relief work, International Terrorism or Tunnel testing are required for obtaining in positioning Fast and accurately information.And technological progress provides the chance based on Local Service (LBS), LBS for outdoor and room area By the acquisition to customer position information, useful information resource in surrounding enviroment is timely and effectively pushed.

GNSS (Global Navigation Satellite System) full name is GLONASS, is one Stage construction, multi-mode, the complicated combined system of multisystem.It refers to all satellite navigation systems, including the whole world, region With it is enhanced, such as the U.S. GPS satellite navigation system, Russia Glonass satellite navigation systems, Europe Galileo defend Star navigation system, the Beidou satellite navigation system of China, and a series of related assisting navigation strengthening systems, such as U.S.'s wide area Strengthening system (WAAS), EGNOS geostationary Navigation Overlay System (EGNOS) and the Japanese MSAS Multi-functional transporting satellites in Europe Strengthening system (MSAS) etc., in addition to currently build and following other that to be built satellite navigation system.

WiFi technology is a kind of short range wireless transmission technology, can support the nothing of linking Internet very well in the range of hundred meters Line electric signal, belongs to IEEE 802.11x standards.WiFi technology is to use SS technologies (Spread Spectrum, spread spectrum), is had It is applied to always before very strong antijamming capability, SS technologies among the communication of the military of U.S. army, for one to Anti-amyloid-β antibody The class communication technology.The most frequently used at present has four kinds of spread spectrum technologies：TH-SS (spread spectrum during jump), DS-SS (Direct Sequence Spread Spectrum), FH- SS (frequency modulation spread spectrum) and C-SS (continuous wave frequency).WIFI is that can not realize positioning in itself, but by measuring access point AP hairs The signal intensity gone out positions come the position to user terminal.At present, the algorithm based on WIFI location technologies is broadly divided into two classes, A kind of is by the way that the signal intensity (RSSI) of measurement is converted into distance measure.AP position coordinates is, it is known that fixed using triangle Position method calculates the position of user.It is another to be positioned by location fingerprint.Its essential idea is in ginseng in actual environment Its received signal strength is measured at examination point and signal strength data storehouse is set up, when being positioned, what intended receivers were received Signal intensity is compared with database, finds the reference point closest to the target and the estimation as target in database Position.In WIFI systems, node AP is similar to the base station in wireless communication networks, is by radio frequency (Radio Frequency) signal carries out the transmission and communication of signal in WLAN.And signal has very strong penetration power, can With through barriers such as wall, furniture, hence in so that WIFI turns into one of preferred hot spot technology of indoor technical research.

Abroad, it is by Guttorm R.Opshaug and Per to a certain extent to realize seamless positioning technology The paper that professor Enge delivers, it is proposed that realize seamless positioning from mixing GPS and UWB, carried by adding redundancy High position precision, also University of Melbourne, University of Calgary etc. propose many seamless positioning algorithms, but are due to location technology The problems of itself and application scenarios scope, can not form the marketization, simply for the experiment simulation stage.At home, The research of theory and actual application is being carried out to indoor and outdoor seamless positioning now with many domestic colleges and universities and scientific research institution.Tsing-Hua University The adaptive weight fusion estimated algorithm based on particle filter and Kalman filtering is proposed in the research to seamless positioning algorithm, is merged Isomery but 3 kinds of location technologies of complementation, average localization error reduce 52% and 63%, can carry out seamless positioning respectively.South Prosperous university proposes the mode that seamless positioning system is positioned with two modules, IMU/GPS integrated navigation systems and IMU/ ZigBee integrated navigation systems, are found out by testing to obtaining data progress analysis, in outdoor carry out GPS location and indoor auxiliary In the hybrid system of ZigBee positioning, the site error resolved is smaller.In the case of the two no location technologies only The resultant error maximum come is oriented using inertial navigation at 16 meters or so, whole error mean is substantially met at 2 meters or so Positioning requirements.Beijing University of Post ＆ Telecommunication uses the technology LTE in forth generation GSM, and WIFI wireless network isomerous environments Under switch over algorithm research, the handoff algorithms of heterogeneous wireless network are researched and analysed, it is theoretical according to fuzzy logic control, The new vertical handoff algorithms based on fuzzy logic are proposed, switching times are reduced, call office obstruction and packet loss is reduced.Although In this way, algorithm research also all the having some limitations property switched both at home and abroad to GNSS and WIFI systems seamless vertical.

For outdoor GLONASS (GNSS), it has the disadvantage：, indoors can not be fine due to its limitation Receive satellite-signal and can not be positioned, especially in the bustlinier shopping centre in city or built-up region, Because GNSS receiver can not timely receive the enough GNSS position location satellites of quantity, then some letters can occur in region to be measured Number critical region, place that such as high building stands, interior of building, metro area etc..When user is among this kind of region, also having can It can be disturbed by various signals, such as receive electromagnetic wave or WLAN signal, can all cause the influence to outdoor positioning.

Realized using GLONASS to outdoor good location, the current U.S., Russia, European Union and China All start to develop GPS in succession, in the near future, the accuracy for outdoor positioning will be more preferable；It is indoor WIFI location technologies are developed rapidly due to WLAN, and WIFI coverings, and cost have all been laid for large- and-medium size cities Low, transmission data are fast, setting accuracy is high, the advantage such as convenient turns into indoor positioning one preferred technique.It is respective excellent using both Point is used among seamless navigation, and the two different alignment systems are blended, and realizes that indoor and outdoor seamless positioning switches, just turns into The invention solves the problems that the key point of indoor and outdoor location navigation problem.

The content of the invention

The purpose of the present invention is to overcome deficiency of the prior art to be realized there is provided one kind based on fuzzy neural network algorithm The algorithm of GNSS and WIFI systems seamless vertical switching, to reach that two different alignment systems are mutually merged, is realized without seaming and cutting Change the purpose of indoor and outdoor location navigation.

The present invention is achieved through the following technical solutions：

A kind of algorithm for realizing the switching of GNSS and WIFI systems seamless vertical, comprises the following steps：Main location information Obtain, based on the handover trigger for receiving signal, the handoff algorithms judgement based on Fuzzy Neural-network Control.

Further, the acquisition of described main location information, when user is in outdoor environment, is entered using GPS Row is positioned, and the satellite number searched is greater than equal to 4, can just be believed to satisfy location condition.When user is in interior When in environment, mobile terminal can not receive gps signal, then positioned using WIFI.It will receive three more than AP's Signal intensity is more than threshold value as the condition of positioning.As a rule, signal intensity is relatively good in belonging to for more than -75dBm Signal.Because indoor environment is complicated, the factor such as noise, multipath effect is had than larger for measurement distance and actual distance Gap.It is also possible to according to the error of positioning come the accuracy of judgment models and the threshold value reasonability set.Using fingerprint During positioning method, when mobile terminal is in finger-print region, so means that and meet positioning requirements, can be positioned.Work as user When being in indoor and outdoor overlapping region, accordingly switched using four kinds of situations of the analysis fingerprint location, in overlapping The user in region makees out position positioning, realizes the purpose of indoor and outdoor seamless positioning.

Further, it is described based on the handover trigger for receiving signal, moved in mobile terminal from outdoor environment to indoor environment When dynamic, when can receive more than three AP signals all more than threshold value, it is believed that meet triangle polyester fibre condition, opened The handover trigger of beginning.Equally, when mobile terminal is moved from indoor environment to outdoor environment, the number of satellite searched meets positioning base This condition, then take place triggering.But, when interior is using fingerprint location technology, used with fingerprint location nearest neighbor algorithm The distance at family, and the Euclidean distance of obtained signal intensity is exactly a kind of form of expression of received signal strength threshold value. At outdoor, the number of satellite that GPS may search for is then as the form of expression for receiving signal strength threshold value.Secondly, it is Prevent the appearance of " ping-pong " phenomenon, can add sluggish level, residence time and motion of mobile terminals in systems The parameters such as trend.

Further, the handoff algorithms judgement based on Fuzzy Neural-network Control, fuzzy neural network is introduced Into Vertical Handover judging process, using its powerful fuzzy reasoning and self-learning ability intelligence make switching judging.And cutting It is a very important job that corresponding threshold value is set during changing, and is be related to switchover policy and algorithm quality one Individual important indicator.For alignment system mutually switches, signal intensity RSS is a most basic reference quantity, just for A form of expression between different location algorithm and different alignment systems.For GPS location, satellite is for mobile whole Distant for end, its intensity level for receiving signal is relatively stablized, it is possible to according to the number of satellite form received It is the resolving for entering row distance according to signal intensity RSS for central triangle polyester fibre algorithm is positioned for WIFI to represent, Then position is obtained using triangular methods, it is more than three signal intensities received using AP in the flat of -75dBm or so Equal signal intensity is used as reference quantity.For central indoor fingerprinting localization algorithm is positioned for WIFI, it is set according to movement Positioned for the signal intensity Euclidean distance obtained in finger-print region, i.e., signal intensity is calculated to obtained minimum Europe A few Reed distances regard signal intensity RSS to judge another judge situation.It is fixed for the use habit of user User is most concerned with the accuracy of positioning precision in the system of position, and this is also to weigh the most important index of alignment system quality.And Among described location algorithm, positioning precision is also to exist as an important parameter.Gps signal stability is compared with WIFI Signal is high, may can be shaken during outer varying environment switching indoors, the signal stabilization degree that mobile terminal is received also is made For the parameter of measurement.

The flow of algorithm is divided into four parts：Switch decision triggering, pre- judgement, fuzzy neural network switching control, switching Judgement.

(1) switch decision is triggered

What switch decision triggering was used is namely based on the handover trigger for receiving signal and user movement trend.Using receiving The Euclidean distance and number of satellite of signal intensity come judge whether to switching, while setting up threshold value to determine whether to Switch over.

The first step：Number of satellite and signal intensity according to receiving judge whether to need to trigger Vertical Handover algorithm.

Second step：When triggering switching, parameter is input among pre- judgement, if clear and definite switch decision can be carried out, Judged result is then directly exported, later step is no longer carried out；If judgement can not be switched over, at the parameter that receives Reason, carries out the processing of fuzzy neural network switching control below.

3rd step：Switch decision exports judged result according to fuzzy neural network (FNN) switching control and selects what is used Alignment system is positioned.

(2) pre- judgement

Addition is pre- before fuzzy neural network switching control judges it is that, in order to alleviate the complexity of system processing, reduction is System consumption memory consumption.The present invention mainly chooses number of satellite and the signal intensity that receives as judging parameter in advance.

GPS location is carried out if the number of satellite n received is more than default threshold value k, if the signal intensity received WIFI positioning is then carried out more than predetermined threshold value Q, if number of satellite or received signal strength are less than default threshold value, Obtained parameter is input among follow-up fuzzy neural network and switches over control process.

(3) fuzzy neural network switching control

First layer is input layer, and input parameter is number of satellite n and received signal strength Q.

The second layer is membership function layer, and the mainly input value to upper strata carries out Fuzzy processing.Obfuscation is exactly pair Data carry out fuzzy variable processing, such as signal intensity is a parameter of Vertical Handover, but the cost of equipment and battery etc. The parameter of factor can also influence Vertical Handover as an input parameter.Fuzzy set is by different membership function groups Into different with digital collection, among fuzzy set, each value only has all when meeting degree of membership, and resulting value could be into For a set.By the way that obtained parameter value is mapped among membership function, corresponding membership function can be obtained Value.And for the method for expressing of membership function, general data and the numerical value of being defined using the straight line or curve of degree of membership.

Using relatively common trapezoidal and triangle combination situation.Membership function is divided into the subset of five grades：A、 B, C, D, E, X_min and X_max represent the minimum value and maximum of the data measurement of selection respectively, and X1, X2, X3 correspondence are defined Fuzzy set span, have corresponding membership function λ among each subset_i, obtained membership function expression formula For：

λ_i=[λ_A, λ_B, λ_C, λ_D, λ_E]_i

Wherein i represented to need the parameter of obfuscation among algorithm, and the membership function of each parameter is expressed as follows：

(3.1) average field-strength received：In the reception membership function of average field-strength, x-axis represent to receive three with Upper signal intensity is -75dBm RSS average values, and y-axis represents the corresponding membership function of different field intensity values, with the intersection point of x-axis from Origin is defined as to the right [- 65, -55, -45, -35, -25], it is assumed that the mobile-terminated average signal strength received is -45dBm, then Corresponding membership function vector is [0,0,1,0,0].

(3.2) number of satellite is received：The membership function of number of satellite, x-axis represented to receive number of satellite, and y-axis is represented The membership function of number of satellite, the satellite situation measured according to software, by the different corresponding degrees of membership of number of satellite Function is defined as to the right [4,5,6,7,8] from origin, it is assumed that the mobile-terminated number of satellite that receives is 5, then corresponding degree of membership Function vector is [0,1,0,0,0].

(3.3) minimum signal strength Euclidean distance：Mobile terminal finger-print region the minimum euclidean of reference point away from From in membership function, x-axis represents the Euclidean distance of minimum signal strength, and y-axis represents its membership function, and corresponding Minimum signal strength Euclidean distance is defined as to the right [0,2,4,6,8] from origin, it is assumed that mobile terminal is in certain time period meter The Euclidean distance calculated is 5, then corresponding membership function vector is [0,0,0.5,0.5,0].

(3.4) position error：Among the membership function of position error, x-axis represents placement error value, and y-axis represents positioning The membership function of error, and GPS is in position fixing process at present, its position error is at 10 meters or so, and WIFI alignment systems are in room Position error is at 2 meters or so when inside using, the membership function corresponding to position error is defined as to the right [0,4,8,12, 16], different position errors can correspond to different membership function values.

(3.5) stability of the mobile-terminated collection of letters number：Gps satellite relative to other positioning from the ground farther out, therefore its to Relatively stablize for family.Indoors among environment, the factor such as noise, multipath effect, wall is positioned for WIFI systems and surveyed Span is had than larger gap from actual distance.

Third layer is fuzzy rule layer.

By above by after parameters membership function, then a fuzzy operation is performed, come in equivalent to by input Multiple parameters membership function value product.The output valve is exactly corresponding node fitness value.With mathematical formulae 1 represent as Under：

α_ijIt is i-th of parameter value for receiving j systems, α_max、α_minI parameter values are in the application respectively in j systems The maximum threshold and minimum threshold of expression.

4th layer is output layer.The layer is each regular node fitness value α_ijAdd a connection weight β_ij, it is therefore an objective to The single value for being converted into suitably carrying out system switching by vector value is judged that basic thought is existed according to parameters Shared significance level, is weighted in system.Cost function is set up, selection functional value highest system is switched over, it is assumed that The functional value of current system is high, just without switching.It is expressed as follows with mathematical formulae 2：

Score_jFor the positioning target function value of j systems, α_ijRegular node fitness value in j systems is represented, passes through the 3rd Layer output is obtained.β_ijFor connection weight, and it is 1 to meet set weights sum.

Verified by many experiments, finally obtaining weight vectors defined formula is：

β_j=[β₁, β₂, β₃]_j=[0.8,0.1,0.1]

β₁Equivalent to number of satellite or the power of the minimum euclid distance of average signal strength RSS or signal intensity Weight values, β₂The weighted value of positioning precision, β₃Received signal strength the evaluation value of stability.

(4) switch decision

Switch and be primarily generated at GPS and WIFI signal overlapping region, starting stage, mobile terminal is opened and receives ginseng Number, two alignment systems work simultaneously.By the calculating of neuro-fuzzy algorithm and object function, initial location system is judged.If When mobile device initial alignment is gps system, next step judges whether to meet WIFI location condition, and one is introduced in the algorithm Residence time timer, starts timing during handover trigger, if in the time restriction of setting, trigger condition is met, then is performed Switching, otherwise returns and re-starts judgement.Equally, if mobile device initial alignment is WIFI systems, next step judges to receive To number of satellite whether meet GPS location condition, introduce residence time timer, trigger condition met in time restriction then Switching, otherwise returns and re-starts judgement.

According to the number of satellite or the situation of change of signal intensity received, the motion of mobile terminal can be judged substantially The parameters received, are then input among the handoff algorithms of fuzzy neural network by trend, by the mould mentioned Criterion and simple weighted are gelatinized, the functional value of system is obtained.If the functional value of current alignment system is more than to be switched positioning system System, then without hand-off process, if the functional value of current alignment system is less than to be switched alignment system, then switch over.

For AP deployment, the optimization deployment principle of the AP based on signal Euclidean distance is totally divided into two steps：First Step, among given WIFI indoor environments, it is considered to after actual size of the space and layout, the transmission characteristic for determining signal, then root AP numbers are determined according to its characteristic.Second step, is established after number, and distribution planning is carried out to these AP position.Ring indoors Among border, when specifying AP numbers, first need predetermined threshold value S, the size of threshold value can directly affect AP number how much.To AP's Number is on the increase, the signal between test point in being disposed by calculating the AP under varying number and neighbouring reference point Euclidean distance average value, takes threshold value S of its maximum again with setting to compare, and draws the minimum AP number bigger than threshold value S, Here it is AP deployment numbers optimal among this environment.It is not only several to signal Europe in reference point when specifying AP optimum positions Reed distance average is maximum, also to allow all reference point locations error within limits.Two steps Mathematically it is expressed as follows：

So that the average signal Euclidean distance of each reference point and standard deviation f₂Maximum is reached, AP optimum bits are obtained Put, so that it is guaranteed that the positioning performance of each reference point can be more or less the same among room area, obtain the high positioning accurate of comparison Degree.

The beneficial effects of the invention are as follows：

1st, for the increasingly extensive satellite fix of GNSS and WIFI applications and airmanship, it is proposed that one kind is based on mould Paste neural network algorithm to realize the algorithm of GNSS and WIFI systems seamless vertical switching, overcome when user is moved to outdoor WIFI positioning accuracy decline, and when user indoors when GNSS signal may interrupt；When GNSS and WIFI are used in combination When can produce substantial amounts of update the data and may caused by result user can be provided mistake location information deficiency.

2nd, present invention employs a kind of optimized algorithm for being different from traditional algorithm, implement described one kind and realize GNSS and WIFI After the algorithm of system seamless vertical switching, can make the conventional four large satellites positioning of GNSS, the GPS of America of navigation application technology, Russian Glonass, European Union Galileo and the big-dipper satellite of China indoors, in the environment such as outdoor and underground, earth's surface, can Seamless coverage in different localizing environments is realized to merge different location technologies, the smooth of location technology is ensured in varied situations Switching and the generation for reducing " ping-pong ", it may also be said to be the positioning of global GNSS satellite, navigation system further expand and Extension.The U.S., Russia, European Union and China can be promoted to stride forward major step in terms of satellite fix, navigation application technological orientation.Together When in national defense construction such as China's military use, aircraft navigation, naval vessels ship, earthquake relief work, International Terrorism, Tunnel testing and civilian Field all has been widely used.

The present invention is further described with reference to the accompanying drawings and examples.

Brief description of the drawings

Fig. 1 is switchover policy figure of the invention；

Fig. 2 is algorithm flow chart of the invention；

Fig. 3 is pre- judgement operation schematic diagram of the invention；

Fig. 4 is structure of fuzzy neural network figure of the invention；

Fig. 5 is degree of membership-domain figure of the invention；

Fig. 6 is signal intensity membership function figure of the invention；

Fig. 7 is number of satellite membership function figure of the invention；

Fig. 8 is minimum signal strength Euclidean distance map of the invention；

Fig. 9 is position error membership function figure of the invention；

Embodiment

In order that those skilled in the art more fully understand the technical scheme in the present invention, below in conjunction with the accompanying drawings to this hair Bright embodiment is described in detail.

A kind of algorithm for realizing the switching of GNSS and WIFI systems seamless vertical, comprises the following steps：Main location information Obtain, based on the handover trigger for receiving signal, the handoff algorithms judgement based on Fuzzy Neural-network Control.Fig. 1 is of the invention Switchover policy figure.

Further, the acquisition of described main location information, when user is in outdoor environment, is entered using GPS Row is positioned, and the satellite number searched is greater than equal to 4, can just be believed to satisfy location condition.When user is in interior When in environment, mobile terminal can not receive gps signal, then positioned using WIFI.It will receive three more than AP's Signal intensity is more than threshold value as the condition of positioning.As a rule, signal intensity is relatively good in belonging to for more than -75dBm Signal.Because indoor environment is complicated, the factor such as noise, multipath effect is had than larger for measurement distance and actual distance Gap.It is also possible to according to the error of positioning come the accuracy of judgment models and the threshold value reasonability set.Using fingerprint During positioning method, when mobile terminal is in finger-print region, so means that and meet positioning requirements, can be positioned.Work as user When being in indoor and outdoor overlapping region, accordingly switched using four kinds of situations of the analysis fingerprint location, in overlapping The user in region makees out position positioning, realizes the purpose of indoor and outdoor seamless positioning.

Further, it is described based on the handover trigger for receiving signal, moved in mobile terminal from outdoor environment to indoor environment When dynamic, when can receive more than three AP signals all more than threshold value, it is believed that meet triangle polyester fibre condition, opened The handover trigger of beginning.Equally, when mobile terminal is moved from indoor environment to outdoor environment, the number of satellite searched meets positioning base This condition, then take place triggering.But, when interior is using fingerprint location technology, used with fingerprint location nearest neighbor algorithm The distance at family, and the Euclidean distance of obtained signal intensity is exactly a kind of form of expression of received signal strength threshold value. At outdoor, the number of satellite that GPS may search for is then as the form of expression for receiving signal strength threshold value.Secondly, it is Prevent the appearance of " ping-pong " phenomenon, can add sluggish level, residence time and motion of mobile terminals in systems The parameters such as trend.

Further, the handoff algorithms judgement based on Fuzzy Neural-network Control, fuzzy neural network is introduced Into Vertical Handover judging process, using its powerful fuzzy reasoning and self-learning ability intelligence make switching judging.And cutting It is a very important job that corresponding threshold value is set during changing, and is be related to switchover policy and algorithm quality one Individual important indicator.For alignment system mutually switches, signal intensity RSS is a most basic reference quantity, just for A form of expression between different location algorithm and different alignment systems.For GPS location, satellite is for mobile whole Distant for end, its intensity level for receiving signal is relatively stablized, it is possible to according to the number of satellite form received It is the resolving for entering row distance according to signal intensity RSS for central triangle polyester fibre algorithm is positioned for WIFI to represent, Then position is obtained using triangular methods, it is more than three signal intensities received using AP in the flat of -75dBm or so Equal signal intensity is used as reference quantity.For central indoor fingerprinting localization algorithm is positioned for WIFI, it is set according to movement Positioned for the signal intensity Euclidean distance obtained in finger-print region, i.e., signal intensity is calculated to obtained minimum Europe A few Reed distances regard signal intensity RSS to judge another judge situation.It is fixed for the use habit of user User is most concerned with the accuracy of positioning precision in the system of position, and this is also to weigh the most important index of alignment system quality.And Among described location algorithm, positioning precision is also to exist as an important parameter.Gps signal stability is compared with WIFI Signal is high, may can be shaken during outer varying environment switching indoors, the signal stabilization degree that mobile terminal is received also is made For the parameter of measurement.

Fig. 2 is algorithm flow chart of the invention.The flow of algorithm is divided into four parts：Switch decision triggering, pre- judgement, mould Paste neutral net switching control, switch decision.

(1) switch decision is triggered

What switch decision triggering was used is namely based on the handover trigger for receiving signal and user movement trend.Using receiving The Euclidean distance and number of satellite of signal intensity come judge whether to switching, while setting up threshold value to determine whether to Switch over.

The first step：Number of satellite and signal intensity according to receiving judge whether to need to trigger Vertical Handover algorithm.

Second step：When triggering switching, parameter is input among pre- judgement, if clear and definite switch decision can be carried out, Judged result is then directly exported, later step is no longer carried out；If judgement can not be switched over, at the parameter that receives Reason, carries out the processing of fuzzy neural network switching control below.

3rd step：Switch decision exports judged result according to fuzzy neural network (FNN) switching control and selects what is used Alignment system is positioned.

(2) pre- judgement

Fig. 3 is pre- judgement operation schematic diagram of the invention.Pre- judgement is added before fuzzy neural network switching control is In order to alleviate the complexity of system processing, reduce system consumption memory consumption.The present invention mainly chooses number of satellite and reception The signal intensity arrived as judging parameter in advance.

GPS location is carried out if the number of satellite n received is more than default threshold value k, if the signal intensity received WIFI positioning is then carried out more than predetermined threshold value Q, if number of satellite or received signal strength are less than default threshold value, Obtained parameter is input among follow-up fuzzy neural network and switches over control process.

(3) fuzzy neural network switching control；Fig. 4 is structure of fuzzy neural network figure of the invention.

First layer is input layer, and input parameter is number of satellite n and received signal strength Q.

The second layer is membership function layer, and the mainly input value to upper strata carries out Fuzzy processing.Obfuscation is exactly pair Data carry out fuzzy variable processing, such as signal intensity is a parameter of Vertical Handover, but the cost of equipment and battery etc. The parameter of factor can also influence Vertical Handover as an input parameter.Fuzzy set is by different membership function groups Into different with digital collection, among fuzzy set, each value only has all when meeting degree of membership, and resulting value could be into For a set.By the way that obtained parameter value is mapped among membership function, corresponding membership function can be obtained Value.And for the method for expressing of membership function, general data and the numerical value of being defined using the straight line or curve of degree of membership. Fig. 5 is degree of membership-domain figure of the invention.

Using relatively common trapezoidal and triangle combination situation.Membership function is divided into the subset of five grades：A、 B, C, D, E, X_min and X_max represent the minimum value and maximum of the data measurement of selection respectively, and X1, X2, X3 correspondence are defined Fuzzy set span, have corresponding membership function λ among each subset_i, obtained membership function expression formula For：

λ_i=[λ_A, λ_B, λ_C, λ_D, λ_E]_i

Wherein i represented to need the parameter of obfuscation among algorithm, and the membership function of each parameter is expressed as follows：

(3.1) average field-strength received：Fig. 6 is signal intensity membership function figure of the invention, in connecing for average field-strength Receive in membership function, x-axis represents to receive the RSS average values that more than three signal intensities are -75dBm, and y-axis represents different fields The corresponding membership function of intensity values, is defined as to the right [- 65, -55, -45, -35, -25] with the intersection point of x-axis from origin, it is assumed that move The average signal strength that moved end is received is -45dBm, then corresponding membership function vector is [0,0,1,0,0].

(3.2) number of satellite is received：Fig. 7 is the number of satellite membership function figure of the present invention, the degree of membership of number of satellite Function, x-axis represents to receive number of satellite, and y-axis represents the membership function of number of satellite, the satellite measured according to software Situation, is defined as to the right [4,5,6,7,8] from origin, it is assumed that mobile-terminated by the different corresponding membership functions of number of satellite Number of satellite is received for 5, then corresponding membership function vector is [0,1,0,0,0].

(3.3) minimum signal strength Euclidean distance：Fig. 8 is minimum signal strength Euclidean distance of the invention Figure, mobile terminal is in the reference point minimum euclid distance membership function of finger-print region, and x-axis represents minimum signal strength Euclidean distance, y-axis represents its membership function, fixed to the right from origin with corresponding minimum signal strength Euclidean distance Justice is [0,2,4,6,8], it is assumed that mobile terminal is 5 in the Euclidean distance that certain time period is calculated, then corresponding degree of membership Function vector is [0,0,0.5,0.5,0].

(3.4) position error：Fig. 9 is the position error membership function figure of the present invention, the membership function of position error Central, x-axis represents placement error value, and y-axis represents the membership function of position error, and GPS is in position fixing process at present, and it is determined Position error is at 10 meters or so, and position error is at 2 meters or so when WIFI alignment systems are used indoors, by corresponding to position error Membership function is defined as to the right [0,4,8,12,16], and different position errors can correspond to different membership function values.

(3.5) stability of the mobile-terminated collection of letters number：Gps satellite relative to other positioning from the ground farther out, therefore its to Relatively stablize for family.Indoors among environment, the factor such as noise, multipath effect, wall is positioned for WIFI systems and surveyed Span is had than larger gap from actual distance.

Third layer is fuzzy rule layer.

By above by after parameters membership function, then a fuzzy operation is performed, come in equivalent to by input Multiple parameters membership function value product.The output valve is exactly corresponding node fitness value.With mathematical formulae 1 represent as Under：

α_ijIt is i-th of parameter value for receiving j systems, α_max、α_minI parameter values are in the application respectively in j systems The maximum threshold and minimum threshold of expression.

4th layer is output layer.The layer is each regular node fitness value α_ijAdd a connection weight β_ij, it is therefore an objective to The single value for being converted into suitably carrying out system switching by vector value is judged that basic thought is existed according to parameters Shared significance level, is weighted in system.Cost function is set up, selection functional value highest system is switched over, it is assumed that The functional value of current system is high, just without switching.It is expressed as follows with mathematical formulae 2：

Score_jFor the positioning target function value of j systems, α_ijRegular node fitness value in j systems is represented, passes through the 3rd Layer output is obtained.β_ijFor connection weight, and it is 1 to meet set weights sum.

Verified by many experiments, finally obtaining weight vectors defined formula is：

β_j=[β₁, β₂, β₃]_j=[0.8,0.1,0.1]

β₁Equivalent to number of satellite or the power of the minimum euclid distance of average signal strength RSS or signal intensity Weight values, β₂The weighted value of positioning precision, β₃Received signal strength the evaluation value of stability.

(4) switch decision

Switch and be primarily generated at GPS and WIFI signal overlapping region, starting stage, mobile terminal is opened and receives ginseng Number, two alignment systems work simultaneously.By the calculating of neuro-fuzzy algorithm and object function, initial location system is judged.If When mobile device initial alignment is gps system, next step judges whether to meet WIFI location condition, and one is introduced in the algorithm Residence time timer, starts timing during handover trigger, if in the time restriction of setting, trigger condition is met, then is performed Switching, otherwise returns and re-starts judgement.Equally, if mobile device initial alignment is WIFI systems, next step judges to receive To number of satellite whether meet GPS location conditions, introduce residence time timer, trigger condition met in time restriction Then switch, otherwise return and re-start judgement.

According to the number of satellite or the situation of change of signal intensity received, the motion of mobile terminal can be judged substantially The parameters received, are then input among the handoff algorithms of fuzzy neural network by trend, by the mould mentioned Criterion and simple weighted are gelatinized, the functional value of system is obtained.If the functional value of current alignment system is more than to be switched positioning system System, then without hand-off process, if the functional value of current alignment system is less than to be switched alignment system, then switch over.

For AP deployment, the optimization deployment principle of the AP based on signal Euclidean distance is totally divided into two steps：First Step, among given WIFI indoor environments, it is considered to after actual size of the space and layout, the transmission characteristic for determining signal, then root AP numbers are determined according to its characteristic.Second step, is established after number, and distribution planning is carried out to these AP position.Ring indoors Among border, when specifying AP numbers, first need predetermined threshold value S, the size of threshold value can directly affect AP number how much.To AP's Number is on the increase, the signal between test point in being disposed by calculating the AP under varying number and neighbouring reference point Euclidean distance average value, takes threshold value S of its maximum again with setting to compare, and draws the minimum AP number bigger than threshold value S, Here it is AP deployment numbers optimal among this environment.It is not only several to signal Europe in reference point when specifying AP optimum positions Reed distance average is maximum, also to allow all reference point locations error within limits.Two steps Mathematically it is expressed as follows：

So that the average signal Euclidean distance of each reference point and standard deviation f₂Maximum is reached, AP optimum bits are obtained Put, so that it is guaranteed that the positioning performance of each reference point can be more or less the same among room area, obtain the high positioning accurate of comparison Degree.

The beneficial effects of the invention are as follows：

1st, for the increasingly extensive satellite fix of GNSS and WIFI applications and airmanship, it is proposed that one kind is based on mould Paste neural network algorithm to realize the algorithm of GNSS and WIFI systems seamless vertical switching, overcome when user is moved to outdoor WIFI positioning accuracy decline, and when user indoors when GNSS signal may interrupt；When GNSS and WIFI are used in combination When can produce substantial amounts of update the data and may caused by result user can be provided mistake location information deficiency.

2nd, present invention employs a kind of optimized algorithm for being different from traditional algorithm, implement described one kind and realize GNSS and WIFI After the algorithm of system seamless vertical switching, can make the conventional four large satellites positioning of GNSS, the GPS of America of navigation application technology, Russian Glonass, European Union Galileo and the big-dipper satellite of China indoors, in the environment such as outdoor and underground, earth's surface, can Seamless coverage in different localizing environments is realized to merge different location technologies, the smooth of location technology is ensured in varied situations Switching and the generation for reducing " ping-pong ", it may also be said to be the positioning of global GNSS satellite, navigation system further expand and Extension.The U.S., Russia, European Union and China can be promoted to stride forward major step in terms of satellite fix, navigation application technological orientation, together When in national defense construction such as China's military use, aircraft navigation, naval vessels ship, earthquake relief work, International Terrorism, Tunnel testing and civilian Field all has been widely used.

It is exemplified as above be only to the present invention for example, do not constitute the limitation to protection scope of the present invention, it is all It is to be belonged to the same or analogous design of the present invention within protection scope of the present invention.Obviously, described embodiment is only It is the embodiment of a part of the invention, rather than whole embodiments.

Claims (4)

1. a kind of algorithm for realizing the switching of GNSS and WIFI systems seamless vertical, it is characterised in that comprise the following steps：It is main fixed The acquisition of position information, based on the handover trigger for receiving signal, the handoff algorithms judgement based on Fuzzy Neural-network Control.

2. the acquisition of main location information according to claim 1, it is characterised in that：When user is in outdoor environment When, positioned using GPS, and the satellite number searched is greater than equal to 4, can just be believed to satisfy location condition. When user is in indoor environment, mobile terminal can not receive gps signal, then positioned using WIFI.It will receive Signal intensity to more than three AP is more than threshold value as the condition of positioning.As a rule, signal intensity is more than -75dBm Belong to relatively good signal.Because indoor environment is complicated, the factor such as noise, multipath effect, for measurement distance and it is true away from Defection has than larger gap.It is also possible to according to the error of positioning come the accuracy of judgment models and the threshold value set Reasonability.During using fingerprint location mode, when mobile terminal is in finger-print region, so means that and meet positioning requirements, can be with Positioned.It is corresponding to carry out using four kinds of situations of the analysis fingerprint location when user is in indoor and outdoor overlapping region Switching, makees out position positioning to the user in overlapping region, realizes the purpose of indoor and outdoor seamless positioning.

3. it is according to claim 1 based on the handover trigger for receiving signal, it is characterised in that：In mobile terminal from outdoor environment When being moved to indoor environment, when can receive more than three AP signals all more than threshold value, it is believed that meet triangle and determine Position condition, the handover trigger started.Equally, when mobile terminal is moved from indoor environment to outdoor environment, the satellite searched Quantity meets positioning primary condition, then triggering takes place.But, it is near with fingerprint location when interior is using fingerprint location technology Adjacent algorithm obtains the distance of user, and the Euclidean distance of obtained signal intensity is exactly received signal strength threshold value A kind of form of expression.At outdoor, the number of satellite that GPS may search for is then as the performance for receiving signal strength threshold value Form.Secondly, in order to prevent " ping-pong " phenomenon occur, can add in systems sluggish level, residence time and The parameters such as motion of mobile terminals trend.

4. the handoff algorithms judgement according to claim 1 based on Fuzzy Neural-network Control, it is characterised in that：Will be fuzzy Neutral net is incorporated into Vertical Handover judging process, using its powerful fuzzy reasoning and self-learning ability intelligence make switching Judge.And set in handoff procedure corresponding threshold value be it is very important one work, be related to switchover policy and One important indicator of algorithm quality.For alignment system mutually switches, signal intensity RSS is a most basic reference Amount, just for a form of expression between different location algorithms and different alignment systems.For GPS location, defend Star is distant for mobile terminal, and its intensity level for receiving signal is relatively stablized, it is possible to according to what is received Number of satellite form represents, is carried out according to signal intensity RSS for the triangle polyester fibre algorithm among WIFI is positioned The resolving of distance, then obtains position using triangular methods, it be more than three signal intensities being received using AP- 75dBm or so average signal strength is used as reference quantity.For central indoor fingerprinting localization algorithm is positioned for WIFI, it It is to be positioned according to mobile device in the signal intensity Euclidean distance that finger-print region is obtained, i.e., calculates signal intensity Obtained minimum Euclideam distance regard signal intensity RSS to judge another judge situation.Further according to making for user For custom, user is most concerned with the accuracy of positioning precision in alignment system, and this is also to weigh alignment system quality most Important index.And among described location algorithm, positioning precision is also to exist as an important parameter.GPS believes Number stability is high compared with WIFI signal, may can shake when outer varying environment switches indoors, mobile terminal is received Signal stabilization degree also serves as the parameter weighed.

The flow of algorithm is divided into four parts：Switch decision triggering, pre- judgement, fuzzy neural network switching control, switch decision.

(1) switch decision is triggered

What switch decision triggering was used is namely based on the handover trigger for receiving signal and user movement trend.Using receiving signal The Euclidean distance and number of satellite of intensity come judge whether to switching, while set up threshold value determine whether to carry out Switching.

The first step：Number of satellite and signal intensity according to receiving judge whether to need to trigger Vertical Handover algorithm.

Second step：When triggering switching, parameter is input among pre- judgement, if clear and definite switch decision can be carried out, directly Output judged result is connect, later step is no longer carried out；If judgement can not be switched over, the parameter received is handled, Carry out the processing of fuzzy neural network switching control below.

3rd step：Switch decision selects the positioning used according to fuzzy neural network (FNN) switching control output judged result System is positioned.

(2) pre- judgement

It is that, in order to alleviate the complexity of system processing, reduction system disappears that pre- judgement is added before fuzzy neural network switching control Consume memory consumption.The present invention mainly chooses number of satellite and the signal intensity that receives as judging parameter in advance.

GPS location is carried out if the number of satellite n received is more than default threshold value k, if the signal intensity received is more than Predetermined threshold value Q then carries out WIFI positioning, if number of satellite or received signal strength are less than default threshold value, will To parameter be input among follow-up fuzzy neural network and switch over control process.

(3) fuzzy neural network switching control

First layer is input layer, and input parameter is number of satellite n and received signal strength Q.

The second layer is membership function layer, and the mainly input value to upper strata carries out Fuzzy processing.Obfuscation is exactly to data Fuzzy variable processing is carried out, such as signal intensity is a parameter of Vertical Handover, but the factor such as the cost of equipment and battery Parameter can also influence Vertical Handover as an input parameter.Fuzzy set is made up of different membership functions, and Digital collection is different, among fuzzy set, and each value only has when all meeting degree of membership, and resulting value could turn into one Set.By the way that obtained parameter value is mapped among membership function, corresponding membership function value can be obtained.And it is right In the method for expressing of membership function, general data and the numerical value of being defined using the straight line or curve of degree of membership.

Using relatively common trapezoidal and triangle combination situation.Membership function is divided into the subset of five grades：A、B、C、 D, E, X_min and X_max represent the minimum value and maximum of the data measurement of selection, the mould of X1, X2, X3 correspondence definition respectively There is corresponding membership function λ among paste set span, each subset_i, obtained membership function expression formula is：

λ_i=[λ_A, λ_B, λ_C, λ_D, λ_E]_i

Wherein i represented to need the parameter of obfuscation among algorithm, and the membership function of each parameter is expressed as follows：

(3.1) average field-strength received：In the reception membership function of average field-strength, x-axis represents to receive more than three letters Number intensity is -75dBm RSS average values, and y-axis represents the corresponding membership function of different field intensity values, with the intersection point of x-axis from origin It is defined as to the right [- 65, -55, -45, -35, -25], it is assumed that the mobile-terminated average signal strength received is -45dBm, then corresponds to Membership function vector be [0,0,1,0,0].

(3.2) number of satellite is received：The membership function of number of satellite, x-axis represents to receive number of satellite, and y-axis represents satellite The membership function of quantity, the satellite situation measured according to software, by the different corresponding membership functions of number of satellite It is defined as to the right from origin [4,5,6,7,8], it is assumed that the mobile-terminated number of satellite that receives is 5, then corresponding membership function Vector is [0,1,0,0,0].

(3.3) minimum signal strength Euclidean distance：It is subordinate in the reference point minimum Euclideam distance of finger-print region mobile terminal In category degree function, x-axis represents the Euclidean distance of minimum signal strength, and y-axis represents its membership function, with corresponding minimum Signal intensity Euclidean distance is defined as to the right [0,2,4,6,8] from origin, it is assumed that mobile terminal is calculated in certain time period Euclidean distance be 5, then corresponding membership function vector be [0,0,0.5,0.5,0].

(3.4) position error：Among the membership function of position error, x-axis represents placement error value, and y-axis represents position error Membership function, and at present GPS in position fixing process, its position error is at 10 meters or so, and WIFI alignment systems make indoors Membership function corresponding to position error is defined as to the right [0,4,8,12,16], no by used time position error at 2 meters or so Same position error can correspond to different membership function values.

(3.5) stability of the mobile-terminated collection of letters number：Gps satellite is positioned from the ground farther out relative to other, therefore it comes to user Say and relatively stablize.Indoors among environment, the factor such as noise, multipath effect, wall, for WIFI systems positioning measurement away from Had from actual distance than larger gap.

Third layer is fuzzy rule layer.

By above by after parameters membership function, then a fuzzy operation is performed, equivalent to come in many will be inputted Individual parameter membership function value product.The output valve is exactly corresponding node fitness value.It is expressed as follows with mathematical formulae 1：

${\mathrm{\α}}_{ij}=\left\{\begin{array}{cc}0,& {\mathrm{\α}}_{ij}\≥{\mathrm{\α}}_{max}\\ \frac{{\mathrm{\α}}_{max}-{\mathrm{\α}}_{ij}}{{\mathrm{\α}}_{max}-{\mathrm{\α}}_{min}},& {\mathrm{\α}}_{min}\≤{\mathrm{\α}}_{ij}\≤{\mathrm{\α}}_{max}\\ 1,& {\mathrm{\α}}_{ij}\≤{\mathrm{\α}}_{min}\end{array}\right.---\left(1\right)$

α_ijIt is i-th of parameter value for receiving j systems, α_max、α_minRespectively i parameter values are represented in the application in j systems Maximum threshold and minimum threshold.

4th layer is output layer.The layer is each regular node fitness value α_ijAdd a connection weight β_ij, it is therefore an objective to it will swear The single value that value is converted into suitably carrying out system switching is judged that basic thought is in system according to parameters In shared significance level, be weighted.Cost function is set up, selection functional value highest system is switched over, it is assumed that current The functional value of system is high, just without switching.It is expressed as follows with mathematical formulae 2：

${\mathrm{Score}}_j=\underset{i}{\Σ}{\mathrm{\β}}_{ij}\·{\mathrm{\α}}_{ij}---\left(2\right)$

Score_jFor the positioning target function value of j systems, α_ijRegular node fitness value in j systems is represented, it is defeated by third layer Go out to obtain.β_ijFor connection weight, and it is 1 to meet set weights sum.

Verified by many experiments, finally obtaining weight vectors defined formula is：

β_j=[β₁, β₂, β₃]_j=[0.8,0.1,0.1]

β₁Equivalent to number of satellite or the weighted value of average signal strength RSS or the minimum Euclideam distance of signal intensity, β₂The weighted value of positioning precision, β₃Received signal strength the evaluation value of stability.

(4) switch decision

Switch and be primarily generated at GPS and WIFI signal overlapping region, starting stage, mobile terminal is opened and receives parameter, Two alignment systems work simultaneously.By the calculating of neuro-fuzzy algorithm and object function, initial location system is judged.If mobile When equipment initial alignment is gps system, next step judges whether to meet WIFI location condition, and one is introduced in the algorithm and is resident Time timer, starts timing during handover trigger, if in the time restriction of setting, trigger condition is met, then performs switching, Otherwise return and re-start judgement.Equally, if mobile device initial alignment is WIFI systems, next step judges that what is received defends Whether star number amount meets GPS location condition, introduces residence time timer, and meeting trigger condition in time restriction then switches, Otherwise return and re-start judgement.

According to the number of satellite or the situation of change of signal intensity received, it can judge that the motion of mobile terminal becomes substantially The parameters received, are then input among the handoff algorithms of fuzzy neural network by gesture, by it is described mention it is fuzzy Change criterion and simple weighted, obtain the functional value of system.If the functional value of current alignment system is more than to be switched alignment system, Then without hand-off process, if the functional value of current alignment system is less than to be switched alignment system, then switch over.

For AP deployment, the optimization deployment principle of the AP based on signal Euclidean distance is totally divided into two steps：The first step, Among given WIFI indoor environments, it is considered to after actual size of the space and layout, the transmission characteristic for determining signal, further according to it Characteristic determines AP numbers.Second step, is established after number, and distribution planning is carried out to these AP position.Environment is worked as indoors In, when specifying AP numbers, first need predetermined threshold value S, the size of threshold value can directly affect AP number how much.To AP number It is on the increase, the signal Europe between test point in being disposed by calculating the AP under varying number and neighbouring reference point is several In distance average, take threshold value S of its maximum again with setting to compare, draw the minimum AP number bigger than threshold value S, this is just It is AP deployment numbers optimal among this environment., not only will be to signal Euclid in reference point when specifying AP optimum positions Distance average is maximum, also to allow all reference point locations error within limits.Two step numbers Method is expressed as follows：

$\left\{\begin{array}{c}\underset{p}{\min }\left\{\underset{\mathrm{\α}}{\max }{f}_1\left(p\right)\right\}\≥S\\ f_1\left(p\right)=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\stackrel{\‾}{d}}_i\\ {\stackrel{\‾}{d}}_i=\underset{j\∈N\left(i\right)}{\mathrm{\Σ}}\left|\right|{\stackrel{\‾}{RSS}}_i-{\stackrel{\‾}{RSS}}_j|{|}_2/length\left(N\right)\\ \mathrm{\α}=\{(X_1,Y_1),(X_2,Y_2),\mathrm{...},(X_p,Y_p)\}\end{array}\right.$

$\max f_2=f_1-{\[\underset{i=1}{\overset{n}{\Σ}}{({\stackrel{\‾}{d}}_i-f_1)}^2\]}^{1/2}$

So that the average signal Euclidean distance of each reference point and standard deviation f₂Maximum is reached, AP optimum positions are obtained, so that Ensuring the positioning performance of each reference point among room area can be more or less the same, and obtain the high positioning precision of comparison.