CN104796866A - Indoor positioning method and device - Google Patents

Abstract

The invention discloses an indoor positioning method and device and relates to the field of positioning. The method includes: selecting a preset amount of particles from the surrounding of a position at which a user is at a current moment; calculating a new position of each particle at a next moment according to walking condition of the user; at the new position of each particle, calculating the probability that the particle serves as the position of the user, and using the probability as a weight of the particle at the next moment; according to coordinates of the new positions of the particles and their weights, subjecting a particle cluster to weighted calculation to obtain its average position, using the average position of the particle cluster as the position of the user at the next moment, resampling all particles according to the weights of the particles at the next moment so as to obtain the preset amount of new particles, and calculating the position of the user at the next moment according to the positions of the new particles so as to acquire the positions of the user at various moments. Positioning precision is improved.

Description

Indoor orientation method and device

Technical field

The present invention relates to positioning field, particularly a kind of indoor orientation method and indoor positioning device.

Background technology

Along with the development of Intelligent mobile equipment, the service application of position-based information is constantly expanded.Along with the location information service of outdoor reaches its maturity, the trend of the oriented development that becomes more meticulous of such service, and to indoor expansion.But be different from outdoor environment, in indoor environment, GNSS (GPS (Global Position System)) signal is significantly decayed when penetrating building, and often has multi-path signal to disturb, very large in indoor error, substantially cannot use.

Current indoor positioning technologies mainly realizes location based on radio-frequency (RF) signal strength, generally has two kinds of localization methods:

Three limit positioning modes, the radio-frequency (RF) signal strength calculating user received according to user and the distance of radio-frequency transmissions end.The position of comprehensive at least three transmitting terminals, and the distance of the user calculated and each transmitting terminal, can estimate the position of user.

Fingerprint location method, needs to select some reference points in the environment in advance, collects the intensity of each radio frequency transmissions in each reference point.User terminal is by comparing the similarity of the signal strength signal intensity of each reference point in the signal strength signal intensity of oneself real-time collecting and database, and similarity is larger, thinks that reference point corresponding to distance is more close, comes estimating user position with this.

Above-mentioned two kinds of localization methods respectively have some shortcomings.Three limit positioning modes are comparatively large by excessive routing influence in indoor environment, and positioning precision is poor.Fingerprint location method needs to scan environment in advance, collect reference point signal data, positioning precision is larger by the impact of the quantity of reference point and data accuracy, and the few or indoor environment change of reference point quantity causes the data of reference point no longer accurately, all can affect positioning precision.

Summary of the invention

In order to solve above-mentioned traditional indoor positioning technologies Problems existing, the present invention proposes a kind of indoor orientation method and indoor positioning device.

First aspect of the present invention provides a kind of indoor orientation method, comprising: the particle choosing predetermined number around user's current time position; The reposition of each particle at subsequent time is calculated according to user's situation of walking; The reposition of each particle calculates the probability of particle as customer location, and using this probability as the weight of this particle at subsequent time; The mean place of population is calculated according to the reposition coordinate of each particle and Weight, and using the mean place of population as the position of user at subsequent time; In the weight of subsequent time, the new particle that resampling obtains predetermined number is carried out to all particles according to each particle, according to the position of the position calculation user of new particle subsequent time again, thus obtain the position in each moment of user.

Second aspect of the present invention provides a kind of indoor positioning device, comprising: particle chooses unit, for choosing the particle of predetermined number around user's current time position; Particle position computing unit, for calculating the reposition of each particle at subsequent time according to user's situation of walking; Granular Weights Computing unit, calculates the probability of particle as customer location on the reposition at each particle, and using this probability as the weight of this particle at subsequent time; Customer location determining unit, for calculating the mean place of population, and using the mean place of population as the position of user at subsequent time according to the reposition coordinate of each particle and Weight; According to the position of the position calculation user of new particle subsequent time again, thus obtain the position in each moment of user; Resampling unit, for carrying out to all particles the new particle that resampling obtains predetermined number in the weight of subsequent time according to each particle.

The present invention uses particle filter to carry out indoor positioning, can provide high position precision service, has stronger adaptability to indoor environment complicated and changeable.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Fig. 1 is the schematic flow sheet of the embodiment that the present invention is based on the indoor orientation method that particle filter technology realizes.

Fig. 2 is the principle schematic of the present invention three limit positioning mode.

Fig. 3 is the process of establishing schematic diagram of fingerprint database of the present invention.

Fig. 4 is the structural representation of the embodiment that the present invention is based on the indoor positioning device that particle filter technology realizes.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described.

Fig. 1 is the schematic flow sheet of the embodiment that the present invention is based on the indoor orientation method that particle filter technology realizes.As shown in Figure 1, the indoor orientation method of the present embodiment comprises:

S102, chooses the particle of predetermined number around user's current time position.

As a rule, number of particles is more, and positioning precision is higher, but amount of calculation also can be larger, therefore can set number of particles flexibly according to the computing capability of the requirement of positioning precision and equipment.Such as can choose 100 ~ 300 particles.

S104, calculates the reposition of each particle at subsequent time according to the walking situation meter of user between current time and subsequent time.

S106, the reposition of each particle calculates the probability of particle as customer location, and using this probability as the weight of this particle at subsequent time.

S108, calculates the mean place of population according to the reposition coordinate of each particle and Weight, and using the mean place of population as the position of user at subsequent time.Such as formula can be adopted be weighted, wherein for the mean place of population, Z _ibe the coordinate figure of i-th particle, w _iit is the weight of i-th particle.

S110, carries out to all particles the new particle that resampling obtains predetermined number in the weight of subsequent time according to each particle.Follow-up can according to the position of the position calculation user of new particle subsequent time again.When specific implementation, first can be normalized the weight of each particle, carry out resampling according to the particle weights after normalization, thus obtain the position in each moment of user.

After first time, location performed step S102, above-mentioned steps S104 ~ S110 is repeated in follow-up location, can obtain the position in each moment of user, and new population.

Weight according to particle carries out resampling to the particle chosen, and the particle that after resampling, weight is large can spread in population, and the little particle of weight can reduce or disappear, thus ensures that the customer location calculated approaches to real customer location.

In addition, owing to considering the error of dead reckoning in state transition equation (8), i.e. stochastic variable δ _z+1, make the new particle of same predecessor extraction in resampling, also can dissimilate in a following state transitions, ensure that population can not be degenerated to a certain local optimum, ensure the robustness of particle filter.

User's current time position in step S102, also referred to as the initial position of user, can adopt the mode of random selecting to select the particle of predetermined number at the initial position of user.When determining the initial position of user, the position that can estimate according to basic fixed position method is as the initial position of user.

Basic fixed position method can be such as three limit positioning modes, fingerprint location method, terrestrial reference identification positioning mode, localization method that specifically can be different according to the different choice of indoor environment.When indoor environment is comparatively complicated, in such as office building, radiofrequency signal is in non-line-of-sight propagation substantially, can adopt fingerprint location method.When indoor environment is comparatively spacious, block fewer, such as hall, market, underground garage, three limit positioning modes can be adopted.Meanwhile, in order to improve positioning precision, in the environment of radio-frequency transmissions source deficiency (such as garage does not almost have radio-frequency transmissions source), the radio-frequency transmissions source of some can be increased, such as Bluetooth Low Energy emission source.The basic fixed position method that the present invention selects is described respectively below.

Three limit positioning modes:

With reference to figure 2, user terminal can receive the signal of some emission sources, and existing signal attenuation model is as follows:

RSS＝A-10n log ₁₀d (1)

Wherein, RSS is receiving end signal intensity; A is the signal strength signal intensity that under reference range, (being generally 1m) receives; N is decay factor, generally 2 ~ 4; D is the distance between receiving terminal and transmitting terminal.

According to above-mentioned formula, when known receiving end signal intensity, the distance of terminal and respective signal emission source can be calculated.The distance of known each signal emitting-source of detecting and receiver is d ₁, d ₂..., d _n, each emission source coordinate is (x ₁, y ₁), (x ₂, y ₂) ..., (x _n, y _n), utilize least square method to get final product the estimated value (x of the coordinate of computing terminal _u, y _u), computing formula is as follows:

$\left\{\begin{array}{c}{(x_1-x_u)}^2+{(y_1-y_u)}^2=d_1^2\\ \·\\ \·\\ \·\\ {(x_n-x_u)}^2+{(y_n-y_u)}^2=d_n^2\end{array}\right.---\left(2\right)$

A front n-1 formula deducts the n-th formula respectively and can obtain:

$\left\{\begin{array}{c}{x}_1^2-x_n^2-2(x_1-x_n)x_u+y_1^2-y_n^2-2(y_1-y_n)y_u=d_1^2-d_n^2\\ \·\\ \·\\ \·\\ x_{n-1}^2-x_n^2-2(x_{n-1}-x_n)x_u+y_{n-1}^2-y_n^2-2(y_{n-1}-y_n)y_u=d_{n-1}^2-d_n^2\end{array}\right.$

If establish:

$A=\left[\begin{array}{cc}2(x_1-x_n)& 2(y_1-y_n)\\ \·& \·\\ \·& \·\\ \·& \·\\ 2(x_{n-1}-x_n)& 2(y_{n-1}-y_n)\end{array}\right]B=\left[\begin{array}{c}{x}_1^2-x_n^2+y_1^2-y_n^2+d_n^2-d_1^2\\ \·\\ \·\\ \·\\ x_{n-1}^2-x_n^2+y_{n-1}^2-y_n^2+d_n^2-d_{n-1}^2\end{array}\right]X=\left[\begin{array}{c}{x}_u\\ y_u\end{array}\right]$

Then obtain:

AX＝B(3)

Conversion above formula obtains formula X=A ^-1b, result of calculation X are the coordinate of user.In order to calculate the coordinate of user, minimum needs 3 information sources, as shown in Figure 2, can receive the signal of information source 1, information source 2, information source 3, utilize these signals can determine the position at user place at customer location place.

The above-mentioned computational methods being three limit positioning modes.

Fingerprint location method:

Before realizing fingerprint location function, need first to set up fingerprint database.With reference to figure 4, first, map is determined the region realizing location, and in region, plans some sweep circuits, make sweep circuit substantially can cover the region of this location.Scanning staff needs hand-held data gathering equipment, along the route walking of planning, equipment constantly records the signal strength signal intensity of each radiofrequency signal received on the way, according to path, calculate the position often organizing radio-frequency (RF) signal strength Data Collection, thus, can generate positional information and signal strength data database one to one, this database will can be used for location.

In addition, in the present invention, the data acquisition equipment magnetic flux density data that also need continuous record to receive on the way.According to the acceleration transducer in the built-in sensors of data acquisition equipment, can the horizontal direction of judgment device and vertical direction, and the magnetic field data collected is preserved the component of horizontal direction and the component of vertical direction.For magnetic field data, consider that distinct device can apply intrinsic deviation to magnetic field sensor, therefore the present invention also proposes the magnetic field data collected continuously to ask poor, also namely to vertical data and the horizontal data M in the magnetic field of any time _hi, M _vicalculate $\left\{\begin{array}{c}{D}_{\mathrm{Hi}}=M_{\mathrm{Hi}}-M_{\mathrm{Hi}-1}\\ D_{\mathrm{Vi}}=M_{\mathrm{Vi}}-M_{\mathrm{Vi}-1}\end{array}\right.,$ And by D _hi, D _vistored in database.Because the sampling interval of magnetic field data is less, magnetic field fingerprint by being distributed on route that scanning staff walked more thick and fast, so the D thinking and try to achieve can be similar to _hi, D _viposition be M _hi, M _vicorresponding position.

In addition, because the change of environment will likely affect propagation path and the magnetic field intensity of radiofrequency signal, therefore fingerprint database needs regular update, and renewal process can realize with reference to the process of establishing of above-mentioned fingerprint database.

At positioning stage, user terminal receives the signal strength information (RSSI of some information sources at any time ₁, RSSI ₂..., RSSI _n) carry out the calculating of signal space distance with the signal strength information of each reference point in fingerprint database.Signal space is regarded as actual geographic apart from little apart from little.Signal space distance can with reference to following formulae discovery:

$d_k={\mathrm{\Σ}}_{i=1}^n{\left({|{\mathrm{RSSI}}_i-{\mathrm{RSSI}}_{k_i}|}^q\right)}^{1/q}---\left(4\right)$

Wherein d _kthe signal space distance of a kth reference point and user's current location, the signal strength signal intensity of the information source i that reference point k collects, RSSI _iit is the signal strength signal intensity of the information source i that user receives.Q is space factor, during q=1, and d _kfor manhatton distance, during q=2, d _kfor Euclidean distance.

According to each reference point of calculating and the signal space distance of user's current location, filter out signal space apart from the alternative reference point of minimum several reference points as user's current location, then current according to calculated with weighted average method user position:

$Z_U=\frac{{\mathrm{\Σ}}_{j=1}^m\frac{1}{d_i}{Z}_{\mathrm{RPj}}}{{\mathrm{\Σ}}_{j=1}^{m}1/d_i}---\left(5\right)$

Wherein, Z _ube the current position of user calculated, m is the number of alternative reference point, represent the coordinate of a jth reference point.

The above-mentioned fingerprint location method being the present invention and adopting.

Terrestrial reference identification positioning mode:

A lot of radio-frequency transmissions source is scattered with in locating area, some signals from emission source can be received from the position at user place, because each emission source is different from the distance between user, as a rule the signal strength signal intensity from different emission source that receives of user is also different, and the position at the emission source place that signal strength signal intensity user received is the strongest is as the position at the current place of user.Terrestrial reference identification positioning mode ratio is easier to realize, and in radio-frequency transmissions source than in the indoor environment of comparatively dense, also can reach certain positioning precision.

In order to improve the reliability that position differentiates further, above-mentioned basic fixed position methods combining can be got up position.Such as, land used identifies the auxiliary three limit positioning modes of other positioning mode or fingerprint location method positions.

In step S104, calculate the reposition of each particle at subsequent time according to user's situation of walking, also namely determine the state transitions of particle, such as, following methods can be adopted to realize:

S104A, determine the step-length of user according to the height of user and cadence, formula is as follows:

L _s＝H(af+b) (6)

In formula, L _sbe the user's step-length estimated, H is the height of user, and f is the cadence of user, a and b is coefficient.Coefficient can pre-set, experimentally or experience obtain.According to the step-length calculated and step number, the displacement of user can be calculated.

Wherein, the height of user can be inputted by user, and the cadence of user and step number can by calculating.The present invention proposes a kind of method according to the acceleration information determination cadence on vertical direction and step number.First, obtain the acceleration information of vertical direction, concrete, the acceleration transducer in user terminal can obtain 3-axis acceleration data, 3-axis acceleration data is projected on gravity direction, can obtain the acceleration information on vertical direction.Secondly, in people's walking process, health can fluctuate thereupon up and down, when terminal and health relative position substantially constant (such as terminal is held in front), the fluctuation of vertical direction is generally because paces produce, substantially not by the impact of other interference, therefore the acceleration information of vertical direction can be used for detecting paces, within the scope of normal cadence, (as 1-2Hz) can be considered to a step at the one-period fluctuated up and down of vertical direction, thus identifies single paces.After being consecutively detected certain step number by said method, Fourier transform is carried out to the acceleration information of this section of vertical direction simultaneously, obtain peak point between 1-2Hz as cadence f.

In addition, although intelligent terminal also can obtain the data of direct acceleration and gyroscope attitude, but due to the restriction of sensor accuracy, direct acceleration information and gyro data are carried out user location and are had very large error, then can ensure certain precision by the method for material calculation of the present invention.

S104B, the angle rotated between two moment of the direction value that the magnetic field sensor held according to the direction value of user's current time, user records at subsequent time and the current time that gyroscope is determined and subsequent time determines the direction value of user's subsequent time.

The direction of travel of user and the top of intelligent terminal are towards identical, and consider terminal in use and not exclusively level, according to the gravity direction calculated, user side is that terminal top is towards projection in the horizontal direction to actual correction.The direction of travel of user can rely on the data of the magnetic field sensor of intelligent terminal inside and gyroscope two transducers to determine.Magnetic field sensor can the direction of magnetic north pole definitely, and because the precision in magnetic field is poor, the present invention adds the sensitivity that gyro data improves direction.Such as adopt the method for Kalman filtering, utilize following formula to determine the direction value of user's subsequent time:

${\mathrm{\θ}}_{k+1}=\widetilde{{\mathrm{\θ}}_{k+1}}-\frac{\mathrm{\α}}{\mathrm{\α}+\mathrm{\β}}(\widetilde{{\mathrm{\θ}}_{k+1}}-({\mathrm{\θ}}_k+{\mathrm{\ω}}_{k+1}))---\left(7\right)$

Wherein, θ _k+1represent the direction value of user's subsequent time, θ _krepresent the direction value of user's current time, represent the direction value that magnetic field sensor records at subsequent time, ω _k+1represent the angle rotated between two moment of the current time that gyroscope is determined and subsequent time, α represents the variance of magnetic field sensor, and β represents gyrostatic variance, rule of thumb or can test and determine variance yields.In the present invention, select α >10 β, to ensure that gained angle value determines primarily of magnetic field.

S104C, according to the position of each particle current time, step number that between two moment, user walks and each particle of Vector operation that often the walks reposition at subsequent time.Such as, each particle of following formulae discovery can be utilized at the reposition of subsequent time:

$\stackrel{\→}{P_{z+1}}=\stackrel{\→}{P_z}+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\→}{L_i}+\stackrel{\→}{{\mathrm{\δ}}_{z+1}}---\left(8\right)$

Wherein, represent the reposition of particle at subsequent time, represent the position of particle current time, represent current time and the vector representation that between subsequent time two moment, user i-th grows step by step, the size of this vector is the step-length determined, the direction of this vector is the direction value of the user's subsequent time determined, for the error of the position according to paces direction and displacement calculating, rule of thumb or can test and obtain.

In step s 106, the reposition of each particle calculates the probability of particle as customer location, and using this probability as the weight of this particle at subsequent time.A kind of exemplary implementation method, the probability density of the signal strength signal intensity of each information source that particle detects at reposition can be calculated, particle to collect this group signal strength signal intensity probability at reposition is calculated according to the probability density of the signal strength signal intensity of each information source and joint probability distribution, and as the weight of this particle at subsequent time.The computational methods of particle weights are described with three limit positioning modes and fingerprint location method below.

In three limit positioning modes, user at the particle weights of any position Z according to the signal strength information (RSSI of current information source ₁, RSSI ₂..., RSSI _n) calculate.

First, owing to having noise jamming in reality, because contemplated that the signal attenuation model of noise is:

RSSI＝A-10nlog ₁₀d+X _δ(9)

Wherein, X _δfor Gaussian distributed average is the noise of zero.For known position Z, d is the distance of Z and information source, and now measured signal intensity can not be a definite value, but in Gaussian Profile, signal strength signal intensity RSSI Arbitrary Information Sources i being detected at position Z can be calculated _iprobability density:

$f\left({\mathrm{RSSI}}_i\right)=\frac{1}{\mathrm{\σ}\sqrt{2\mathrm{\π}}}{e}^{-\frac{{({\mathrm{RSSI}}_i-A+10n{\log }_{10}{d}_I)}^2}{2{\mathrm{\σ}}^2}}---\left(10\right)$

In formula, d _ifor the distance of given position Z and i-th information source, σ is variance, and the implication of A and n, see the explanation in formula (1), repeats no more here.

Thus, this position can be collected the probability of one group of signal strength signal intensity according to the formula of joint probability distribution, be the particle weights w at position Z _z.

w _z＝F(RSSI ₁,RSSI ₂,..,RSSI _n)＝f(RSSI ₁)f(RSSI ₂)...f(RSSI _n) (11)

In fingerprint location method, owing to have collected radio-frequency (RF) signal strength and Magnetic Field in fingerprint base generation simultaneously.When calculating the weight of particle, magnetic field and radio-frequency (RF) signal strength can be used to judge.

First, according to radio-frequency (RF) signal strength, if given position is not reference point, then need the signal strength signal intensity of first carrying out Interpolate estimation given position according to known reference point.Choose the reference point in given position certain limit, to arbitrary information source k, calculate its signal strength signal intensity mean value in the method for given position linearly interpolation:

$\stackrel{\‾}{{\mathrm{RSSI}}_k}=\frac{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\frac{1}{d_i}{\mathrm{RSSI}}_{\mathrm{ki}}}{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\frac{1}{d_i}}---\left(12\right)$

Then detect that i-th its signal strength signal intensity of information source is RSSI _iprobability be:

$f\left({\mathrm{RSSI}}_i\right)=\frac{1}{\mathrm{\σ}\sqrt{2\mathrm{\π}}}{e}^{-\frac{{({\mathrm{RSSI}}_i-\stackrel{\‾}{{\mathrm{RSSI}}_i})}^2}{2{\mathrm{\σ}}^2}}---\left(13\right)$

Thus, according to the formula of the probability of each information source and joint probability distribution, the probability a certain position Z collecting one group of signal strength signal intensity can be calculated:

w _z＝F(RSSI ₁,RSSI ₂,..,RSSI _n)＝f(RSSI ₁)f(RSSI ₂)...f(RSSI _n) (14)

Further, following methods can also be adopted to revise the probability that particle collects this group signal strength signal intensity at reposition:

First, it is poor that the continuous magnetic field data between calculating current time and subsequent time is sampled.By the magnetic field in this section of path vertically and the data of horizontal component, (M is set to _h1..., M _hn) and (M _v1..., M _vn).For eliminating the intrinsic magnetic field deviation of distinct device, calculate the sampling of continuous magnetic field data poor $\left\{\begin{array}{c}{D}_{\mathrm{Hi}}=M_{\mathrm{Hi}+1}-M_{\mathrm{Hi}}\\ D_{\mathrm{Vi}}=M_{\mathrm{Vi}+1}-M_{\mathrm{Vi}}\end{array}\right.$ Obtain (D _h1..., D _hn-1) and (D _v1..., D _vn-1).

Secondly, by the poor (D of continuous magnetic field data sampling _h1..., D _hn-1) and (D _v1..., D _vn-1) compare with the similarity of each fragment of magnetic field finger print data stream, choose the position Z of the maximum magnetic field finger print data flow section correspondence of similarity _mas most probable customer location.

Finally, with this position Z _mcentered by, according to most probable customer location and similarity thereof, the probability that particle collects this group signal strength signal intensity at reposition is revised.Following formula such as can be adopted to revise:

$W_Z=F({\mathrm{RSSI}}_1,{\mathrm{RSSI}}_2,...,{\mathrm{RSSI}}_n)e^{-\mathrm{\δ}|Z-Z_M|}---\left(15\right)$

Wherein δ is the similarity of the most similar pattern obtained according to pattern matching primitives.Similarity is larger, Z _mthe contribution of position to granular Weights Computing is larger.

Above-mentioned granular Weights Computing method has merged the probability calculation based on signal strength signal intensity, and the result of the magnetic field data flow graph sample coupling of collecting continuously, and the weight of COMPREHENSIVE CALCULATING particle, the weight of calculating is more accurate.

In the present invention, basic fixed position part can have been calculated by terminal, also can have been calculated by server, and the amount of calculation that the particle filter in hi-Fix service needs is larger, to the computing of handheld terminal and battery power consumption demand larger, therefore can select the signal strength signal intensity of collection and built-in sensors data upload high in the clouds, and calculate beyond the clouds, and then result feedback is returned terminal.

Fig. 4 is the structural representation that the present invention is based on another embodiment of indoor positioning device that particle filter technology realizes.As shown in Figure 4, the indoor positioning device 400 of the present embodiment comprises:

Particle chooses unit 402, for choosing the particle of predetermined number around user's current time position;

Particle position computing unit 404, for calculating the reposition of each particle at subsequent time according to user's situation of walking;

Granular Weights Computing unit 406, calculates the probability of particle as customer location on the reposition at each particle, and using this probability as the weight of this particle at subsequent time;

Customer location determining unit 408, for calculating the mean place of population, and using the mean place of population as the position of user at subsequent time according to the reposition coordinate of each particle and Weight.

Resampling unit 410, for carrying out to all particles the new particle that resampling obtains predetermined number in the weight of subsequent time according to each particle.Then customer location determining unit, for the position according to the position calculation user of new particle subsequent time again, thus obtains the position in each moment of user.

In one embodiment, particle position computing unit 404, specifically for: the step-length determining user according to the height of user and cadence; The angle rotated between two moment of the direction value that the magnetic field sensor held according to the direction value of user's current time, user records at subsequent time and the current time that gyroscope is determined and subsequent time determines the direction value of user's subsequent time; According to the position of each particle current time, step number that between two moment, user walks and each particle of Vector operation that often the walks reposition at subsequent time.

In one embodiment, particle position computing unit 404 is according to the acceleration information determination cadence on vertical direction and step number.

In one embodiment, particle position computing unit 404 utilizes following formula to determine the direction value of user's subsequent time:

${\mathrm{\θ}}_{k+1}=\widetilde{{\mathrm{\θ}}_{k+1}}-\frac{\mathrm{\α}}{\mathrm{\α}+\mathrm{\β}}(\widetilde{{\mathrm{\θ}}_{k+1}}-({\mathrm{\θ}}_k+{\mathrm{\ω}}_{k+1}))$

Wherein, θ _k+1represent the direction value of user's subsequent time, θ _krepresent the direction value of user's current time, represent the direction value that magnetic field sensor records at subsequent time, ω _k+1represent the angle rotated between two moment of the current time that gyroscope is determined and subsequent time, α represents the variance of magnetic field sensor, and β represents gyrostatic variance.

In one embodiment, particle position computing unit 404 utilizes each particle of following formulae discovery at the reposition of subsequent time:

$\stackrel{\→}{P_{z+1}}=\stackrel{\→}{P_z}+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\→}{L_i}+\stackrel{\→}{{\mathrm{\δ}}_{z+1}}$

Wherein, represent the reposition of particle at subsequent time, represent the position of particle current time, represent current time and the vector representation that between subsequent time two moment, user i-th grows step by step, the size of this vector is the step-length determined, the direction of this vector is the direction value of the user's subsequent time determined, for the error of the position according to paces direction and displacement calculating.

In one embodiment, granular Weights Computing unit 406, specifically for: the probability density calculating the signal strength signal intensity of each information source that particle detects at reposition; Particle to collect this group signal strength signal intensity probability at reposition is calculated according to the probability density of the signal strength signal intensity of each information source and joint probability distribution, and as the weight of this particle at subsequent time.

In one embodiment, granular Weights Computing unit 406, also for adopting following methods to revise the probability that particle collects this group signal strength signal intensity at reposition: when there being the fingerprint database previously gathered in the environment applied, the continuous magnetic field data calculated between current time and subsequent time is sampled poor; Continuous magnetic field data sampling difference is compared with the similarity of each fragment of magnetic field finger print data stream, choose the magnetic field finger print data flow section that similarity is maximum, if similarity is greater than a threshold value ψ, then think that the final position of this magnetic field data flow section correspondence is the most probable customer location Z calculated according to magnetic field _m; According to Z _mmost probable customer location and similarity weights W that particle is calculated with the signal strength signal intensity of radiofrequency signal _zrevise:

wherein F (RSSI ₁, RSSI ₂..., RSSI _n) be the joint probability calculated according to signal strength signal intensity, δ is the similarity of the most similar pattern arrived obtained according to pattern matching primitives.Illustrate that magnetic field pattern similarity is larger, Z _mparticle weights around position is larger.

User's current time position is also referred to as the initial position of user, therefore indoor positioning device can also comprise: initial user position determination unit, determines user's current time position for utilizing three limit positioning modes, fingerprint location method or terrestrial reference identification positioning mode.

One of ordinary skill in the art will appreciate that all or part of step realizing above-described embodiment can have been come by hardware, the hardware that also can carry out instruction relevant by program completes, described program can be stored in a kind of computer-readable recording medium, the above-mentioned storage medium mentioned can be read-only memory, disk or CD etc.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (16)

1. an indoor orientation method, is characterized in that, comprising:

The particle of predetermined number is chosen around user's current time position;

The reposition of each particle at subsequent time is calculated according to user's situation of walking;

The reposition of each particle calculates the probability of particle as customer location, and using this probability as the weight of this particle at subsequent time;

The mean place of population is calculated according to the reposition coordinate of each particle and Weight, and using the mean place of population as the position of user at subsequent time;

In the weight of subsequent time, the new particle that resampling obtains predetermined number is carried out to all particles according to each particle, according to the position of the position calculation user of new particle subsequent time again, thus obtain the position in each moment of user.

2. the method for claim 1, is characterized in that, describedly calculates each particle according to user's situation of walking and comprises at the reposition of subsequent time:

The step-length of user is determined according to the height of user and cadence;

The angle rotated between two moment of the direction value that the magnetic field sensor held according to the direction value of user's current time, user records at subsequent time and the current time that gyroscope is determined and subsequent time determines the direction value of user's subsequent time;

According to the position of each particle current time, step number that between two moment, user walks and each particle of Vector operation that often the walks reposition at subsequent time.

3. method as claimed in claim 2, is characterized in that, according to the acceleration information determination cadence on vertical direction and step number.

4. method as claimed in claim 2, is characterized in that, utilize following formula to determine the direction value of user's subsequent time:

${\mathrm{\θ}}_{k+1}=\widetilde{{\mathrm{\θ}}_{k+1}}-\frac{\mathrm{\α}}{\mathrm{\α}+\mathrm{\β}}\left(\widetilde{{\mathrm{\θ}}_{k+1}}({\mathrm{\θ}}_k+{\mathrm{\ω}}_{k+1})\right)$

Wherein, θ _k+1represent the direction value of user's subsequent time, θ _krepresent the direction value of user's current time, represent the direction value that magnetic field sensor records at subsequent time, ω _k+1represent the angle rotated between two moment of the current time that gyroscope is determined and subsequent time, α represents the variance of magnetic field sensor, and β represents gyrostatic variance.

5. method as claimed in claim 2, is characterized in that, utilize each particle of following formulae discovery at the reposition of subsequent time:

$\stackrel{\→}{P_{z+1}}=\stackrel{\→}{P_z}+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\→}{L_i}+\stackrel{\→}{{\mathrm{\δ}}_{z+1}}$

Wherein, represent the reposition of particle at subsequent time, represent the position of particle current time, represent current time and the vector representation that between subsequent time two moment, user i-th grows step by step, the size of this vector is the step-length determined, the direction of this vector is the direction value of the user's subsequent time determined, for the error of the position according to paces direction and displacement calculating.

6. the method for claim 1, is characterized in that, describedly on the reposition of each particle, calculates the probability of particle as customer location, and is comprised as the weight of this particle at subsequent time by this probability:

The probability of the signal strength signal intensity instantly collected from each information source that estimation particle detects at reposition;

The probability corresponding according to the signal strength signal intensity instantly of each information source and joint probability calculation method, calculate the probability collecting this group signal strength signal intensity at reposition, and as the weight of this particle at subsequent time.

7. method as claimed in claim 6, is characterized in that, adopts following methods to revise the probability that particle collects this group signal strength signal intensity at reposition:

The continuous magnetic field data calculated between current time and subsequent time is sampled poor;

Continuous magnetic field data sampling difference is compared with the similarity of each fragment of magnetic field finger print data stream, chooses the position of the maximum magnetic field finger print data flow section correspondence of similarity as most probable customer location;

According to most probable customer location and similarity thereof, the probability that particle collects this group signal strength signal intensity at reposition is revised.

8. the method for claim 1, is characterized in that, utilizes three limit positioning modes, fingerprint location method or terrestrial reference identification positioning mode to determine user's current time position.

9. an indoor positioning device, is characterized in that, comprising:

Particle chooses unit, for choosing the particle of predetermined number around user's current time position;

Particle position computing unit, for calculating the reposition of each particle at subsequent time according to user's situation of walking;

Granular Weights Computing unit, calculates the probability of particle as customer location on the reposition at each particle, and using this probability as the weight of this particle at subsequent time;

Customer location determining unit, for calculating the mean place of population, and using the mean place of population as the position of user at subsequent time according to the reposition coordinate of each particle and Weight; Also for the position of the position calculation user of new particle that obtains according to resampling unit subsequent time again, thus obtain the position in each moment of user;

Resampling unit, for carrying out to all particles the new particle that resampling obtains predetermined number in the weight of subsequent time according to each particle.

10. device as claimed in claim 9, is characterized in that, described particle position computing unit, specifically for:

The step-length of user is determined according to the height of user and cadence;

The angle rotated between two moment of the direction value that the magnetic field sensor held according to the direction value of user's current time, user records at subsequent time and the current time that gyroscope is determined and subsequent time determines the direction value of user's subsequent time;

According to the position of each particle current time, step number that between two moment, user walks and each particle of Vector operation that often the walks reposition at subsequent time.

11. devices as claimed in claim 10, it is characterized in that, described particle position computing unit is according to the acceleration information determination cadence on vertical direction and step number.

12. devices as claimed in claim 10, it is characterized in that, described particle position computing unit utilizes following formula to determine the direction value of user's subsequent time:

${\mathrm{\θ}}_{k+1}=\widetilde{{\mathrm{\θ}}_{k+1}}-\frac{\mathrm{\α}}{\mathrm{\α}+\mathrm{\β}}\left(\widetilde{{\mathrm{\θ}}_{k+1}}({\mathrm{\θ}}_k+{\mathrm{\ω}}_{k+1})\right)$

Wherein, θ _k+1represent the direction value of user's subsequent time, θ _krepresent the direction value of user's current time, represent the direction value that magnetic field sensor records at subsequent time, ω _k+1represent the angle rotated between two moment of the current time that gyroscope is determined and subsequent time, α represents the variance of magnetic field sensor, and β represents gyrostatic variance.

13. devices as claimed in claim 10, is characterized in that, described particle position computing unit utilizes each particle of following formulae discovery at the reposition of subsequent time:

$\stackrel{\→}{P_{z+1}}=\stackrel{\→}{P_z}+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\→}{L_i}+\stackrel{\→}{{\mathrm{\δ}}_{z+1}}$

Wherein, represent the reposition of particle at subsequent time, represent the position of particle current time, represent current time and the vector representation that between subsequent time two moment, user i-th grows step by step, the size of this vector is the step-length determined, the direction of this vector is the direction value of the user's subsequent time determined, for the error of the position according to paces direction and displacement calculating.

14. devices as claimed in claim 9, is characterized in that, described granular Weights Computing unit, specifically for:

The probability of the signal strength signal intensity instantly collected from each information source that estimation particle detects at reposition;

The probability corresponding according to the signal strength signal intensity instantly of each information source and joint probability calculation method, calculate the probability collecting this group signal strength signal intensity at reposition, and as the weight of this particle at subsequent time.

15. devices as claimed in claim 14, is characterized in that, described granular Weights Computing unit, also for adopting following methods to revise the probability that particle collects this group signal strength signal intensity at reposition:

The continuous magnetic field data calculated between current time and subsequent time is sampled poor;

Continuous magnetic field data sampling difference is compared with the similarity of each fragment of magnetic field finger print data stream, chooses the position of the maximum magnetic field finger print data flow section correspondence of similarity as most probable customer location;

According to most probable customer location and similarity thereof, the probability that particle collects this group signal strength signal intensity at reposition is revised.

16. devices as claimed in claim 9, is characterized in that, also comprise: initial user position determination unit, determine user's current time position for utilizing three limit positioning modes, fingerprint location method or terrestrial reference identification positioning mode.