CN103411607B - Pedestrian's step-size estimation and dead reckoning method - Google Patents

Abstract

The invention discloses a kind of pedestrian's step-size estimation and dead reckoning method, the method comprises the steps: a) to utilize the portable terminal with acceleration transducer to measure pedestrian by individual's step parameter K during normal step-length walking; B) detect lines People's Bank of China makes a move the maximal value a of total acceleration amplitude in process _maxwith minimum value a _min; C) formula is adopted calculate pedestrian to walk the step-length step of a described step _size.A kind of non-linear step-length model newly that the present invention proposes, illustrate more accurately step-length that pedestrian's row makes a move and period total acceleration amplitude maximum, relation between minimum value.The pedestrian's step-size estimation proposed and dead reckoning method are convenient to computing machine and are realized, and can calculate the position of pedestrian in real time, to realize indoor positioning, tool has significant practical applications.

Description

Pedestrian's step-size estimation and dead reckoning method

Technical field

The present invention relates to pedestrian's reckoning field.Specifically, the present invention relates to a kind of new pedestrian's step-size estimation and dead reckoning method.

Background technology

Walking reckoning refers to step number to pedestrian's walking, step-length, direction are measured and estimate, thus extrapolates the information such as the track of pedestrian's walking and position, is one of technology realizing indoor positioning.Wherein pedestrian's direction of travel can adopt existing heading measure instrument to measure, as electronic compass and gyroscope.But in people's walking process of how being expert at, real-time estimation goes out the step-length of pedestrian is difficult problem, step-length estimation directly will affect the precision of walking reckoning.In order to estimating step length, the model existed at present can be divided into four classes: constant/pseudo-constant step-length model, linear step-length model, non-linear step-length model and artificial intelligence step-length model.Consider convenience and the real-time of algorithm realization, take into account the accuracy of step-size estimation simultaneously, often choose non-linear step-length model.At present, conventional non-linear step-length model mainly contains following several:

Documents " H.Weinberg; " Using the ADXL202in Pedometer and Personal Navigation Applications; " Analog Devices AN-602Application Note, 2002. " disclose Weinberg approach model:

${\mathrm{step}}_{\mathrm{size}}=K^{*}\sqrt[4]{a_{\max }-a_{\min }}$

Documents " J.Scarlet; " Enhancing the Performance of Pedometers Using a Single Accelerometer; " Analog Devices AN-900Application Note, 2005. " disclose Scarlet approach model:

${\mathrm{step}}_{\mathrm{size}}=K^{*}\frac{\frac{\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\left|a_j\right|}{N}-a_{m\mathrm{in}}}{a_{\max }-a_{\min }}$

Documents " J.W.Kim; H.J.Jang; D-H.Hwang; and C.Park; " A Step, Stride and Heading Determination for the Pedestrian Navigation System, " Journal of Global Positioning Systems; pp.273-279,2004. " discloses Kim approach model:

${\mathrm{step}}_{\mathrm{size}}=K^{*}\sqrt[3]{\frac{\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\left|a_j\right|}{N}}$

In these formula, a _maxand a _minthe maximum total acceleration amplitude produced when representing that pedestrian's row makes a move respectively and minimum total acceleration amplitude; a _jrepresent in the process that pedestrian's row makes a move, gather the jth total acceleration amplitude in N number of total acceleration amplitude altogether; K represents the parameter of the separate equations for asynchronous passerby, and namely asynchronous passerby estimates by statistical actual step size and step-length the parameter that result is given according to during normal step-length walking.Due to for asynchronous passerby, above-mentioned formula all gives corresponding parameter, and when therefore pedestrian is with normal step-length walking, the step sizes adopting above-mentioned step-size estimation algorithm to estimate is less with actual step size phase ratio error.

But found through experiments, when same pedestrian walks with asynchronous long row, step-length application condition compared with actual step size that existing non-linear step-size estimation algorithm estimates is large.

Patent of invention content

An object of the present invention is to overcome the deficiencies in the prior art, a kind of new pedestrian's step-size estimation method is provided, this method propose a kind of non-linear step-length model newly, make it accurately can not only estimate the normal step-length of asynchronous passerby, to the different step-lengths of same pedestrian, also there is better adaptability, while not increasing calculated amount, reduce the error of the different step-length estimation of same pedestrian.

For realizing this object, the present invention adopts following technical scheme:

A kind of pedestrian's step-size estimation method, the method comprises the steps:

A) portable terminal with acceleration transducer is utilized to measure pedestrian by individual's step parameter K during normal step-length walking;

B) detect lines People's Bank of China makes a move the maximal value a of total acceleration amplitude in process _maxwith minimum value a _min;

C) formula is adopted ${\mathrm{step}}_{\mathrm{size}}=K^{*}((a_{\max }-a_{\min })+\sqrt[4]{a_{\max }-a_{\min }}),$ Calculate pedestrian to walk the step-length step of a described step _size.

Preferably, the method measuring individual step parameter K in described step a) can comprise the steps:

A-1) setup parameter Ki is arbitrary constant;

A-2) pedestrian's taking portable terminal tries away n step by normal step-length, records actual walking X rice, and in the process, this portable terminal keeps stationary state relative to the upper part of the body of the pedestrian carrying it;

A-3) detect pedestrian walk n step every a one-step process in the maximal value a of total acceleration amplitude _maxwith minimum value a _min, adopt formula ${\mathrm{step}}_{\mathrm{size}}={\mathrm{Ki}}^{*}((a_{\max }-a_{\min })+\sqrt[4]{a_{\max }-a_{\min }})$ Estimate pedestrian to walk the every long step by step of n step, describedly often longly step by step will to add up, estimate the distance Y rice that pedestrian n walking is walked;

A-4) formula is adopted calculate individual step parameter K.

Preferably, in described step b), detect lines People's Bank of China makes a move the maximal value a of total acceleration amplitude in process _maxwith minimum value a _minmethod can comprise the steps:

B-1) pedestrian carries make a move capable relative to the portable terminal of its upper part of the body maintenance stationary state, total acceleration amplitude accl is obtained by the acceleration transducer being embedded into portable terminal, People's Bank of China of being expert at makes a move in process, and the frequency of acceleration transducer sense acceleration is not less than 15Hz;

B-2) peak value and the valley of total acceleration amplitude in a one-step process is detected;

B-3) difference between the peak value and valley of total acceleration amplitude exceedes threshold value, and their time interval within the specific limits, then this peak value is the maximal value a of the total acceleration amplitude that pedestrian's row makes a move _max, valley is the minimum value a of the total acceleration amplitude that pedestrian's row makes a move _min.

Preferably, at described step b-1) after also can comprise step b-1-1):

B-1-1) adopt sliding-window filtering device to carry out signal filtering to the total acceleration amplitude accl obtained, make total acceleration amplitude signal curve smoothing.

Preferably, the acceleration transducer of described portable terminal can be 3-axis acceleration sensor, at described step b-1) in the amplitude (accl_x, accl_y, accl_z) of 3-axis acceleration sensor sense acceleration three axles, then adopt formula $\mathrm{accl}=\sqrt{{(\mathrm{accl}\_x)}^2+{(\mathrm{accl}\_y)}^2+{(\mathrm{accl}\_z)}^2}$ Calculate total acceleration amplitude accl.

Preferably, portable terminal can comprise acceleration transducer and processor, and the acceleration information detected is sent to processor by acceleration transducer, carries out calculating realize this method by processor, or data upload to far-end backstage, is carried out calculating and is realized this method by processor on backstage.

Two of object of the present invention is to overcome the deficiencies in the prior art, provides a kind of new pedestrian's dead reckoning method, is convenient to utilize computing machine or other electronic equipments estimate pedestrian's step-length in real time and carry out dead reckoning.

For realizing this object, the present invention adopts following technical scheme:

Utilize the pedestrian's step-size estimation method described in claim 1 to carry out pedestrian's dead reckoning method, the method comprises the steps:

A) pedestrian's initial position is set;

B) step-length utilizing the pedestrian's step-size estimation method described in claim 1 to estimate pedestrian's row to make a move; Utilize heading measure instrument measurement row people to walk the course angle of this step simultaneously;

C) according to estimating the step-length and course angle that pedestrian's row makes a move, the position that pedestrian is new is estimated.

Preferably, the described pedestrian's initial position that arranges obtains by input manually or use WIFI or GPS locator meams.

Preferably, the course angle that pedestrian's row makes a move records by electronic compass or gyroscope.

The present invention proposes a kind of non-linear step-length model newly:

${\mathrm{step}}_{\mathrm{size}}=K^{*}((a_{\max }-a_{\min })+\sqrt[4]{a_{\max }-a_{\min }}),$

This model illustrate more accurately step-length that pedestrian's row makes a move and period total acceleration amplitude maximum, relation between minimum value.Experiment shows, adopts pedestrian's step-size estimation method of this non-linear step-length model while not increasing calculated amount, has better adaptability, make step-size estimation more accurate to the different step-lengths of same pedestrian.Pedestrian's dead reckoning method that the present invention proposes is convenient to computing machine and is realized, and calculate the position of pedestrian in real time, to realize indoor positioning, tool has significant practical applications.

By following description also by reference to the accompanying drawings, the present invention will become more clear, and these accompanying drawings are for explaining embodiments of the invention.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the embodiment of the present invention 1;

Fig. 2 is the process flow diagram of the step 100 of the embodiment of the present invention 1;

Fig. 3 is the acceleration amplitude oscillogram of the embodiment of the present invention 1;

Fig. 4 to 6 is design sketchs of the embodiment of the present invention 1;

Fig. 7 is the schematic diagram of the embodiment of the present invention 2;

Fig. 8 is the process flow diagram of the embodiment of the present invention 2.

Embodiment

Embodiment 1, the present embodiment relates to a kind of pedestrian's step-size estimation method, and its process flow diagram is with reference to Fig. 1.First, pedestrian's taking portable terminal tries away by normal step-length, and this portable terminal comprises 3-axis acceleration sensor and processor, and this portable terminal keeps stationary state relative to the upper part of the body of the pedestrian carrying it.Specifically, this portable terminal can be held in the fixing hand do not waved of pedestrian, also can be fixed in the waist of pedestrian, and be not limited to this, identify individual step parameter K in the process and store (step 100).Identify that the method for individual step parameter K will specifically describe at further part.

After obtaining the individual step parameter K of pedestrian, can start to estimate the step-length of this pedestrian.When pedestrian carries the walking of this portable terminal, the frequency of 3-axis acceleration sensor sense acceleration is not less than 15Hz, and 3-axis acceleration sensor uses formula after the amplitude (accl_x, accl_y, accl_z) of acceleration three axles being detected $\mathrm{accl}=\sqrt{{(\mathrm{accl}\_x)}^2+{(\mathrm{accl}\_y)}^2+{(\mathrm{accl}\_z)}^2}$ Calculate total acceleration amplitude accl(step 101).

Then, sliding-window filtering device also can be adopted to carry out signal filtering to the total acceleration amplitude accl obtained, make waveform level and smooth.

Fig. 3 is the total acceleration amplitude oscillogram of the embodiment of the present invention.As shown in the figure, horizontal ordinate is the sampled point numbering of acceleration amplitude, and ordinate is total acceleration amplitude, line style is that pedestrian walks the original total acceleration signal waveform of four steps, line style is by the total acceleration signal waveform after sliding-window filtering.

Can find from Fig. 3, pedestrian often makes a move, and total acceleration amplitude all can occur maximal value and minimum value, and acceleration signal presents periodic feature along with pedestrian walks.Therefore adopt peak detection method, detect crest value and the trough value of total acceleration amplitude, if the difference between crest value and trough value exceedes threshold value, and their time interval within the specific limits, just can judge that pedestrian has walked a step.This crest value is the maximal value a that pedestrian walks total acceleration amplitude in this step _max, trough value is the minimum value a that pedestrian walks total acceleration amplitude in this step _min(step 102).Between described crest value and trough value, the threshold value of difference can walk the maximal value of the total acceleration amplitude of each step in process according to statistics pedestrian examination and minimum value determines, threshold value can equal pedestrian's examination and walks the maximal value of total acceleration amplitude of each step and the mean value of the difference of minimum value in process and be multiplied by 0.8 specifically.The scope of described crest value and trough value time interval can determine according to the cadence of pedestrian, and the cadence of usual pedestrian is at about 1 ~ 3Hz, then the scope of time interval can be 0.3 second to 1 second.In practical application, detect lines People's Bank of China makes a move the maximal value a of acceleration amplitude in process _maxwith minimum value a _minmethod be not limited only to said method, also can adopt other effective ways.

Finally, formula is adopted ${\mathrm{step}}_{\mathrm{size}}=K^{*}((a_{\max }-a_{\min })+\sqrt[4]{a_{\max }-a_{\min }})$ Calculate pedestrian to walk the step-length step of this step _size(step 103).This formula is a kind of non-linear step-length model newly that the present invention proposes,

With reference to Fig. 2, it is the process flow diagram of the step 100 of the embodiment of the present invention 1.Identify in step 100 that the method for individual step parameter K is as follows: first setup parameter Ki is arbitrary constant (step 200), pedestrian's taking portable terminal tries away n step by normal step-length, record actual walking X rice, in the process, this portable terminal keeps stationary state (step 201) relative to the upper part of the body of the pedestrian carrying it.Detect pedestrian walk n step every a one-step process in the maximal value a of total acceleration amplitude _maxwith minimum value a _min, adopt formula ${\mathrm{step}}_{\mathrm{size}}={\mathrm{Ki}}^{*}((a_{\max }-a_{\min })+\sqrt[4]{a_{\max }-a_{\min }})$ Estimate pedestrian to walk the every long step by step of n step, describedly often longly step by step will to add up, estimate the distance Y rice (step 202) that pedestrian n walking is walked.Adopt formula calculate the individual step parameter K(step 203 of this pedestrian).

In the present embodiment, the acceleration information detected is sent to the processor of portable terminal by the 3-axis acceleration sensor of portable terminal, is undertaken calculating the estimation realizing pedestrian's step-length by this processor.In practical application, also can by data upload to far-end backstage, carry out calculating on backstage and realize this method, the hardware device realizing this method does not limit.

With reference to Fig. 4 to Fig. 6, it is the design sketch of the embodiment of the present invention 1.XIAOMI M1 intelligent mobile phone platform achieves Weinberg, Kim, Scarlet tri-kinds of step-size estimation methods in step-size estimation method in this paper and background technology.Same pedestrian's handheld mobile phone walks 10m, 30m and 50m with asynchronous long row respectively, experimental result as shown in Figure 4,5, 6, wherein with four kinds of line styles represent the result of Weinberg, Kim, Scarlet tri-kinds of step-size estimation methods in step-size estimation method in this paper and background technology respectively, and in figure, ordinate represents the distance of the pedestrian's walking adopting step-size estimation method to estimate.The total acceleration amplitude that wherein Kim step-size estimation method uses eliminates acceleration of gravity.

Found through experiments:

When the normal step-length of pedestrian is walked (walk 10m 14 steps, walk 30m and walk 41 steps, walk 50m 68 steps), as shown in Figure 3, four kinds of algorithms estimation results are basically identical, and error is little;

(10m 10 steps are walked when the large walking of same pedestrian is walked, walk 30m 30 steps, walk 50m 51 steps), as shown in Figure 4, pedestrian actual walk 50m, step-size estimation method of the present invention estimation result is 51.94m, and Weinberg, the estimation result of Kim and Scarlet step-size estimation method is respectively 41.62m, about 45.41m and 40.32m, and error is larger;

(10m 20 steps are walked when same pedestrian's small step long row is walked, walk 30m 60 steps, walk 50m 100 steps), as shown in Figure 5, pedestrian actual walk 50m, step-size estimation method of the present invention estimation result is 52.44m, and Weinberg, the estimation result of Kim and Scarlet step-size estimation method is respectively 60.57m, about 61.05m and 76.12m, and error is larger.

Therefore, prove by experiment, the present invention preferably resolves when same pedestrian walks with asynchronous long row, and pedestrian's step-length that existing non-linear step-size estimation method estimates exists the problem compared with big error, and tool has significant practical applications.

Embodiment 2, the embodiment of the present invention relates to a kind of method that the pedestrian's of utilization dead reckoning method realizes indoor positioning, and process flow diagram is with reference to Fig. 8.Pedestrian, in indoor positioning, first will obtain initial position, and the reference position of pedestrian can manually input or use the auxiliary acquisition of other locator meamss, such as WIFI location (step 300).Then, pedestrian's row makes a move, estimate pedestrian according to a kind of pedestrian's step-size estimation method described in embodiment 1 to walk the step-length (step 301) of this step, utilize electronic compass or gyroscope or other instrument sensors to obtain pedestrian to walk the course angle (step 302) of this step simultaneously, the step-length that last basis estimates and the course angle recorded, estimate the position (step 303) that pedestrian is new, then, with the initial position (getting back to step 300) of the new location updating pedestrian estimated, pedestrian again row makes a move, just according to step 300, 301, 302, 303 estimate the new position of pedestrian again, by that analogy, thus realize pedestrian's indoor positioning.

With reference to Fig. 7, for utilizing the schematic diagram of the method for the step-length estimated and the course angle recorded estimation pedestrian reposition in the present embodiment.If a certain moment pedestrian is in P _n-1(X _n-1, Y _n-1) place, direction of travel and E axle clamp angle are α, row make a move (step-length be d) after arrive P _n(X _n, Y _n), then P _n(X _n, Y _n) and P _n-1(X _n-1, Y _n-1) relation between 2 is as follows: $\left\{\begin{array}{c}{X}_n=X_{n-1}+d\×\cos \mathrm{\α}\\ Y_n=Y_{n-1}+d\×\sin \mathrm{\α}\end{array}\right..$

More than in conjunction with most preferred embodiment, invention has been described, but the present invention is not limited to the embodiment of above announcement, and should contain various carry out according to essence of the present invention amendment, equivalent combinations.

Claims (9)

1. pedestrian's step-size estimation method, the method comprises the steps:

A) portable terminal with acceleration transducer is utilized to measure pedestrian by individual's step parameter K during normal step-length walking;

B) detect lines People's Bank of China makes a move the maximal value a of total acceleration amplitude in process _maxwith minimum value a _min;

C) formula is adopted ${\mathrm{step}}_{size}=K*\left(\right(a_{max}-a_{min})+\sqrt[4]{a_{max}-a_{min}}),$ Calculate pedestrian to walk the step-length step of a described step _size.

2. a kind of pedestrian's step-size estimation method according to claim 1, is characterized in that, described step a) in measure the method for individual step parameter K and comprise the steps:

A-1) setup parameter Ki is arbitrary constant;

A-2) pedestrian's taking portable terminal tries away n step by normal step-length, records actual walking X rice, and in the process, this portable terminal keeps stationary state relative to the upper part of the body of the pedestrian carrying it;

A-3) detect pedestrian walk n step every a one-step process in the maximal value a of total acceleration amplitude _maxwith minimum value a _min, adopt formula ${\mathrm{step}}_{size}=Ki*\left(\right(a_{max}-a_{min})+\sqrt[4]{a_{max}-a_{min}})$ Estimate pedestrian to walk the every long step by step of n step, describedly often longly step by step will to add up, estimate the distance Y rice that pedestrian n walking is walked;

A-4) formula is adopted calculate individual step parameter K.

3. a kind of pedestrian's step-size estimation method according to claim 1, is characterized in that, at described step b) in detect lines People's Bank of China to make a move the maximal value a of total acceleration amplitude in process _maxwith minimum value a _minmethod comprise the steps:

B-1) pedestrian carries make a move capable relative to the portable terminal of its upper part of the body maintenance stationary state, total acceleration amplitude accl is obtained by the acceleration transducer being embedded into portable terminal, People's Bank of China of being expert at makes a move in process, and the frequency of acceleration transducer sense acceleration is not less than 15Hz;

B-2) peak value and the valley of total acceleration amplitude in a one-step process is detected;

B-3) difference between the peak value and valley of total acceleration amplitude exceedes threshold value, and their time interval within the specific limits, then this peak value is the maximal value a of the total acceleration amplitude that pedestrian's row makes a move _max, valley is the minimum value a of the total acceleration amplitude that pedestrian's row makes a move _min.

4. a kind of pedestrian's step-size estimation method according to claim 3, is characterized in that, at described step b-1) after also comprise step b-1-1):

B-1-1) adopt sliding-window filtering device to carry out signal filtering to the total acceleration amplitude accl obtained, make total acceleration amplitude signal curve smoothing.

5. a kind of pedestrian's step-size estimation method according to claim 3, it is characterized in that, the acceleration transducer of portable terminal is 3-axis acceleration sensor, at described step b-1) in the amplitude (accl_x of 3-axis acceleration sensor sense acceleration three axles, accl_y, accl_z), then formula is adopted $accl=\sqrt{{(accl\_x)}^2+{(accl\_y)}^2+{(acc\_z)}^2}$ Calculate total acceleration amplitude accl.

6. a kind of pedestrian's step-size estimation method according to claim 1, it is characterized in that, portable terminal comprises acceleration transducer and processor, the acceleration information detected is sent to processor by acceleration transducer, carry out calculating by processor and realize this method, or data upload to far-end backstage, is carried out calculating and is realized this method by processor on backstage.

7. utilize the pedestrian's step-size estimation method described in claim 1 to carry out pedestrian's dead reckoning method, the method comprises the steps:

A) pedestrian's initial position P is set _n-1(X _n-1, Y _n-1);

B) the step-length d utilizing the pedestrian's step-size estimation method described in claim 1 to estimate pedestrian's row to make a move; Utilize heading measure instrument measurement row people to walk the course angle α of this step simultaneously;

C) formula is adopted $\left\{\begin{array}{c}{X}_n=X_{n-1}+d\×\cos \mathrm{\α}\\ Y_n=Y_{n-1}+d\×\sin \mathrm{\α}\end{array}\right.$ Estimate the position that pedestrian is new.

8. a kind of pedestrian's dead reckoning method according to claim 7, is characterized in that, the described pedestrian's of setting initial position is by inputting manually or using WIFI or GPS locator meams to obtain.

9. a kind of pedestrian's dead reckoning method according to claim 7, it is characterized in that, the course angle that pedestrian's row makes a move is recorded by electronic compass or gyroscope.