CN110455316B - Self-adaptive zero-speed interval detection method - Google Patents

Abstract

The invention provides an adaptive zero-speed interval detection method, which belongs to the technical field of pedestrian navigation and positioning. Including: S1: obtain the output data of the inertial device in real time, and the output data includes the angular velocity information and acceleration information of the inertial device; S2: determine the initial angular velocity energy threshold for angular velocity zero velocity detection according to the angular velocity information; S3: according to the The initial angular velocity energy threshold value performs angular velocity zero velocity detection on the sliding window of angular velocity zero velocity detection, and determines the number of detection points in the sliding window and the interval state corresponding to the sliding window; S4: According to the number of detection points and the The interval state corresponding to the sliding window adaptively adjusts the initial angular velocity energy threshold to obtain an initial zero velocity interval based on the initial angular velocity energy threshold; S5: perform acceleration and zero velocity on the initial zero velocity interval according to the acceleration information Detecting and correcting the initial zero-speed interval to obtain a corrected zero-speed interval.

Description

An adaptive zero-speed interval detection method

technical field

The invention relates to the technical field of pedestrian navigation and positioning, in particular to an adaptive zero-speed interval detection method.

Background technique

The zero-speed update method is an error correction method widely used in the pedestrian navigation system. It periodically clears the position error of the navigation solution according to the characteristics of the zero-speed interval in the pedestrian's gait. The premise of zero-speed update is to accurately and effectively detect the zero-speed interval in the process of pedestrian movement. Commonly used zero-speed interval detection methods include acceleration modulo method, angular velocity modulo method, acceleration sliding standard deviation method, or a combination of multiple methods. However, the existing zero-speed interval detection methods have fixed or single threshold settings, and cannot The changes of gait are adaptively adjusted, and the adaptability to multi-gait movements is poor.

SUMMARY OF THE INVENTION

The problem solved by the present invention is that the existing zero-speed interval detection method has poor adaptability to multiple gaits.

One aspect of the present invention provides an adaptive zero-speed interval detection method, comprising:

S1: Obtain output data of the inertial device in real time, where the output data includes angular velocity information and acceleration information of the inertial device;

S2: Determine the initial angular velocity energy threshold for angular velocity zero velocity detection according to the angular velocity information;

S3: perform zero-speed angular velocity detection on the sliding window of zero-speed angular-velocity detection according to the initial angular-velocity energy threshold, and determine the number of detection points in the sliding window and the interval state corresponding to the sliding window;

S4: adaptively adjust the initial angular velocity energy threshold according to the number of detection points and the interval state corresponding to the sliding window to obtain an initial zero velocity interval based on the initial angular velocity energy threshold;

S5: Perform acceleration zero-speed detection on the initial zero-speed interval according to the acceleration information, and correct the initial zero-speed interval to obtain a corrected zero-speed interval.

Optionally, the S4 step includes:

S410: Calculate the time t corresponding to the sliding window according to the number of detection points;

S420: Adaptively adjust the initial angular velocity energy threshold according to the time t and the interval state corresponding to the sliding window to obtain the initial zero velocity interval based on the initial angular velocity energy threshold.

Optionally, the step S420 includes:

When the interval state corresponding to the sliding window is a motion interval, determine whether the size of the time t satisfies t ₁ ≤t≤t ₂ , where t ₁ is the first preset time, and t ₂ is the second preset time : Yes, go back to step S3, and perform zero-speed angular velocity detection on the next sliding window; No, perform adaptive adjustment on the initial angular velocity energy threshold T _ω , and carry out the zero-speed interval according to the adjustment times of the initial angular velocity energy threshold judge;

When the interval state corresponding to the sliding window is a zero-speed interval, determine whether the size of the time t satisfies t≥t ₃ , where t ₃ is the third preset time: yes, the sliding window is determined as the initial Zero-speed interval; if no, adaptively adjust the initial angular velocity energy threshold, and judge the zero-velocity interval according to the adjustment times of the initial angular velocity energy threshold.

Optionally, when the interval state corresponding to the sliding window is a motion interval, the step of adaptively adjusting the initial angular velocity energy threshold includes:

When the size of the time t satisfies t<t ₁ , reducing the initial angular velocity energy threshold Tω;

When the size of the time t satisfies t>t ₂ , the initial angular velocity energy threshold is increased.

Optionally, when the interval state corresponding to the sliding window is a zero-speed interval, the adaptive adjustment to the initial angular velocity energy threshold is:

Increase the initial angular velocity energy threshold.

Optionally, performing the zero-speed interval judgment according to the adjustment times of the initial angular velocity energy threshold includes:

Count the adjustment times of the initial angular velocity energy threshold, and judge whether the adjustment times of the initial angular velocity energy threshold exceed the preset value: No, go back to step S3; Yes, judge whether the sliding window is in the zero-speed range: Yes, so The sliding window is determined to be the initial zero-velocity interval; if not, go back to step S3, and perform zero-velocity detection of angular velocity on the next sliding window.

Optionally, the S5 step includes:

S510: Determine an acceleration amplitude threshold for acceleration zero-speed detection according to the acceleration information corresponding to the detection points in the initial zero-speed interval in the initial zero-speed interval;

S520: Perform acceleration zero-speed detection on all the initial zero-speed interval detection points according to the acceleration amplitude threshold, and correct the initial zero-speed interval to obtain a corrected zero-speed interval.

Optionally, the step S510 includes:

S511: Calculate acceleration amplitudes corresponding to all the initial zero-speed interval detection points according to the acceleration information corresponding to the initial zero-speed interval detection points;

S512: Sort the acceleration amplitudes corresponding to all the initial zero-speed interval detection points, and generate a temporary interval according to the sorted acceleration amplitudes;

S513: Determine the acceleration amplitude threshold according to the acceleration amplitude corresponding to the detection point in the temporary interval in the temporary interval.

Optionally, the step S520 includes:

Perform acceleration zero-speed detection on the initial zero-speed interval detection points in sequence according to the acceleration amplitude threshold value, and obtain all the initial zero-speed interval detection points in the zero-acceleration state;

The modified zero-speed interval is generated according to the initial zero-speed interval detection points of all zero-acceleration states.

Optionally, the step S2 includes:

S210: Calculate the angular velocity energy of the first n (n≥1) detection points according to the angular velocity information;

S220: Calculate the mean value of angular velocity energy and the standard deviation of angular velocity energy according to the angular velocity energy;

S230: Determine the initial angular velocity energy threshold according to the angular velocity energy mean value and the angular velocity energy standard deviation.

In the self-adaptive zero-speed interval detection method of the present invention, the initial angular velocity energy threshold is determined by the angular velocity information output by the inertial device, so that the initial angular velocity energy threshold matches the pedestrian's initial gait; Sliding the window to perform zero-speed angular velocity detection, determining the number of detection points in the sliding window and the interval state corresponding to the sliding window, and adaptively adjusting the initial angular velocity energy threshold according to the interval state and the number of detection points, The initial angular velocity energy threshold is adjusted in a gradual and urgent manner to obtain an initial zero velocity interval based on the initial angular velocity energy threshold, which reduces the probability of missed detection of the zero velocity state detection point; The zero-speed interval is corrected to obtain the more accurate corrected zero-speed interval. The adaptive zero-speed interval detection method of the present invention can gradually adjust the initial angular velocity energy threshold during the pedestrian movement process, and can realize multi-gait. Zero-speed interval detection, high detection reliability and strong practicability.

Description of drawings

Fig. 1 is the overall flow chart of the self-adaptive zero-speed interval detection method according to the present invention;

FIG. 2 is a flowchart of steps S2 of one embodiment of the adaptive zero-speed interval detection method according to the present invention;

3 is a flowchart of step S4 in one embodiment of the adaptive zero-speed interval detection method according to the present invention;

FIG. 4 is a schematic diagram of a zero-speed interval and a motion interval in one embodiment of the adaptive zero-speed interval detection method according to the present invention.

Detailed ways

In order to make the above objects, features and advantages of the present invention more clearly understood, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

Please refer to Figure 4, taking the left foot as an example in the pedestrian movement: Pose1-Pose2 is the movement interval, that is, the gait process from the left foot leaving the ground to the contact with the ground is the movement interval, and Pose3-Pose4 is the zero-speed interval , that is, the gait process corresponding to the left foot always in contact with the ground is in the zero-speed interval; in pedestrian navigation, the motion parameters are generally corrected by obtaining motion observations in the zero-speed interval. The speed range also includes the case where the speed is zero and the acceleration is not zero. Therefore, the speed and acceleration are generally combined to make a comprehensive judgment. However, due to individual differences and the existence of multiple gaits, the adaptability of zero-speed interval detection is poor.

The inventor of the present invention proposes a progressively adjusted adaptive zero-speed interval detection method based on long-term research. The main basis is that no matter what gait a pedestrian is in, the zero-speed interval will not be too short, and the motion interval will not be too long. The invention firstly processes the output data of the inertial device to obtain the initial angular velocity energy threshold for angular velocity zero-velocity detection, and then adjusts the initial angular velocity energy threshold in a progressive approximation manner until the pedestrian gait initially conforms to the basic law of pedestrian motion, and obtains An initial zero-speed interval based on the initial angular velocity energy threshold. The main purpose of this step is to not leak the zero-speed detection point as much as possible. At this time, the initial zero-speed interval is larger than the real zero-speed interval, and then zero-speed detection is performed through the acceleration zero-speed detection. Speed correction to get the corrected zero-speed range.

It should be noted that, in this specification, a data group output by the inertial device at the same time point is used as a detection point for analysis; the sliding window represents an interval composed of multiple detection points, and the sliding window is an interval length, also It is the interval in which the number of detection points will change. When the state corresponding to the detection point changes, the sliding window ends, and the initial starting point of the next sliding window is the first detection point after the sliding window. In this specification, the interval state corresponding to the sliding window is a zero-speed interval or, and the sliding window is a zero-speed interval, which means that the interval composed of the detection points in the sliding window is a zero-speed interval; the sliding window The corresponding interval state is a motion interval or, and the sliding window is a motion interval, both indicating that the interval composed of detection points in the sliding window is a motion interval.

Please refer to Figure 1, an adaptive zero-speed interval detection method, including:

S1: Obtain output data of the inertial device in real time, where the output data includes angular velocity information and acceleration information of the inertial device;

S2: Determine the initial angular velocity energy threshold T _ω for angular velocity zero velocity detection according to the angular velocity information;

S3: carry out angular velocity zero velocity detection on the sliding window of angular velocity zero velocity detection according to the initial angular velocity energy threshold T _ω , and determine the number L of detection points in the sliding window and the interval state corresponding to the sliding window;

S4: adaptively adjust the initial angular velocity energy threshold T _ω according to the number of detection points L and the interval state corresponding to the sliding window to obtain an initial zero velocity interval based on the initial angular velocity energy threshold;

S5: Perform acceleration zero-speed detection on the initial zero-speed interval according to the acceleration information, and correct the initial zero-speed interval to obtain a corrected zero-speed interval.

Referring to FIG. 2, in some embodiments, the step S2 includes:

S210: Calculate the angular velocity energy of the first n (n≥1) detection points according to the angular velocity information.

Specifically, from the beginning of detection and calculation, the angular velocity information of the first n detection points is obtained, and the corresponding angular velocity energy is calculated; the calculation formula of angular velocity energy ω _i is:

和

为三轴陀螺仪在第i(1≤i≤n，n＜30)个检测点获取到的角速度信息。in

and

is the angular velocity information obtained by the three-axis gyroscope at the i-th (1≤i≤n, n<30) detection point.

S220: Calculate the angular velocity energy mean value μ _ω and the angular velocity energy standard deviation σ _ω according to the angular velocity energy.

Specifically, the calculation formula of the average angular velocity energy μ _ω is:

The formula for calculating the standard deviation of angular velocity energy σ _ω is:

S230: Determine the initial angular velocity energy threshold T _ω according to the angular velocity energy mean value μ _ω and the angular velocity energy standard deviation σ _ω .

Specifically, take out the angular velocity _energies satisfying the following formula in _{ω 1} , ω ₂ ,...

ω _i ≤ μ _ω +λ ₁ σ _ω

where λ ₁ is a constant, which is calibrated by experimental data. In this embodiment, λ ₁ =5,

Take the initial angular velocity energy threshold T _ω as:

T _ω =max{ω _i }

Of course, it should be understood that, in some embodiments, the initial angular velocity energy threshold T _ω can be taken as:

T _ω = μ _ω +λ ₁ σ _ω

In this way, firstly calculate the mean angular velocity energy μ _ω and the standard deviation σ _ω of angular velocity energy of the first n (n>1) detection points according to the angular velocity information, and then calculate the mean value μ _ω of angular velocity energy and the standard deviation σ of angular velocity energy according to the angular velocity information. _ω determines the initial angular velocity energy threshold T _ω , no matter what gait the pedestrian is at when the detection starts, the initial angular velocity energy threshold T _ω is associated with the angular velocity energy mean value μ _ω and the angular velocity energy standard deviation σ _ω , the initial angular velocity energy threshold T _ω can achieve the preliminary adaptation of the gait with pedestrians, which provides a basis for the subsequent adaptive adjustment of the initial angular velocity energy threshold T _ω , which meets the needs of multi-gait zero-speed interval detection, High reliability and strong practicability.

In some embodiments, the S3 step includes:

S310: Perform angular velocity zero-speed detection on the detection points in the sliding window sequentially from front to back, until the angular velocity zero-speed detection data satisfies the first end condition.

Specifically, the angular velocity zero velocity detection is performed according to the angular velocity zero velocity detection formula, and the angular velocity zero velocity detection formula is:

Among them, j is the j-th detection point calculated from the beginning of detection, C _j ^ω is the motion state corresponding to the j-th detection point, when C _j ^ω = 1, the motion state corresponding to the j-th detection point is zero-speed state, When C _j ^ω =0, the motion state corresponding to the jth detection point is the motion state;

The first end condition is:

Wherein, P is the first detection point in the sliding window, and L ₀ is an integer greater than 0;

S320: When the angular velocity zero velocity detection data satisfies the first end condition, determine the number L of the detection points in the sliding window.

Specifically, the number of detection points L: L=L ₀ ; the (P+L ₀ -1)th detection point is the last detection point in the sliding window;

S330: When the first detection point in the sliding window is in the zero-speed state, the interval state corresponding to the sliding window is in the zero-speed interval, and when the first detection point in the sliding window is in the motion state, the The interval state corresponding to the sliding window is the motion interval.

时，所述滑动窗口对应的区间状态为零速区间，当

时，所述滑动窗口对应的区间状态为运动区间。Specifically, when

, the interval state corresponding to the sliding window is the zero-speed interval, and when

, the interval state corresponding to the sliding window is a motion interval.

It should be understood that, in other embodiments, the step S310 further includes:

When the first detection point in the sliding window is in a zero-velocity state, the angular velocity zero-velocity detection is performed on the detection points before the sliding window from the back to the front of the first detection point until the angular velocity is zero. The fast detection data satisfies the second end condition.

时，根据所述角速度零速检测公式以P为起点从后至前依次对所述滑动窗口之前的检测点进行角速度零速检测，直至所述角速度零速检测数据满足第二结束条件。That is, when

, according to the angular velocity zero velocity detection formula, with P as the starting point, angular velocity zero velocity detection is performed on the detection points before the sliding window sequentially from the back to the front, until the angular velocity zero velocity detection data satisfies the second end condition.

The second end condition is:

or

Wherein, L ₁ is an integer greater than 0.

Correspondingly, in the step S320, the number of detection points L: L=L ₀ +L ₁ . It should be understood that, at this time, the interval corresponding to the sliding window changes, and the first detection point P in the sliding window is: P=PL ₁ +1.

In this way, when the interval state corresponding to the sliding window is zero-speed state, after the initial angular velocity energy threshold T _ω is adaptively adjusted, the range corresponding to the sliding window is correspondingly based on the initial angular velocity energy threshold T _ω Forward and backward adjustment avoids the occurrence of leaking zero-speed state detection points, with high reliability and strong practicability.

Referring to Figure 3, the step S4 includes:

S410: Calculate the time t corresponding to the sliding window according to the number L of detection points.

Specifically, the time t=(L-1)*t ₀ , where t ₀ is the detection period of the inertial device. In this embodiment, the detection frequency of the inertial device is 100 Hz, that is, t ₀ is 0.01s.

S420: Adaptively adjust the initial angular velocity energy threshold T _ω according to the time t and the interval state corresponding to the sliding window to obtain the initial zero velocity interval based on the initial angular velocity energy threshold T _ω .

Specifically, in some embodiments, the step S420 includes:

When the interval state corresponding to the sliding window is a motion interval, determine whether the size of the time t satisfies t ₁ ≤t≤t ₂ , where t ₁ is the first preset time, and t ₂ is the second preset time : Yes, go back to step S3, and perform zero-speed angular velocity detection on the next sliding window; No, perform adaptive adjustment on the initial angular velocity energy threshold T _ω , and perform zero-speed adjustment according to the adjustment times of the initial angular velocity energy threshold T _ω speed interval judgment;

When the interval state corresponding to the sliding window is a zero-speed interval, determine whether the size of the time t satisfies t≥t ₃ , where t ₃ is the third preset time: yes, the sliding window is determined as the initial Zero-speed interval; if no, adaptively adjust the initial angular velocity energy threshold T _ω , and judge the zero-velocity interval according to the adjustment times of the initial angular velocity energy threshold T _ω .

It should be understood that when the sliding window is a movement interval, according to the movement characteristics of a person, in any gait, the movement interval is neither too short nor too long. By judging whether the size of the time t satisfies the t ₁ ≤ t ≤ t ₂ , it is determined whether the initial angular velocity energy threshold value T _ω needs to be adjusted, and the judgment basis is reliable and practical. The first preset time t ₁ and the second preset time t ₂ are calibrated according to experimental data. In some embodiments, the first preset time t ₁ =0.05s, the second preset time t ₂ =3s, in this embodiment, the first preset time t ₁ =0.1s, and the second preset time t ₂ =2s.

It should be understood that the starting point of the next sliding window is the first detection point after the sliding window.

Specifically, when the interval state corresponding to the sliding window is a motion interval, the step of adaptively adjusting the initial angular velocity energy threshold T _ω includes:

When the size of the time t satisfies t<t ₁ , reducing the initial angular velocity energy threshold T _ω ;

When the size of the time t satisfies t>t ₂ , the initial angular velocity energy threshold T _ω is increased.

Specifically, in some embodiments, when the size of the time t satisfies t<t ₁ , the adjustment method of the initial angular velocity energy threshold T _ω is: T _ω =k ₁ *T _ω , where k ₁ < 1, k ₁ is a constant, which is calibrated by experimental data; in some embodiments, k ₁ =0.85, in this embodiment, the k ₁ =0.95; in other implementations, when the size of the time t satisfies t When <t ₁ , the adjustment method of the initial angular velocity energy threshold value T _ω is: T _ω =min _{j=P, . . . , P+L} {ω _j }.

Specifically, in some embodiments, when the size of the time t satisfies t>t ₂ , the adjustment method of the initial angular velocity energy threshold T _ω is: T _ω =k ₂ *T _ω , where k ₂ > 1, k ₂ is a constant, which is calibrated by experimental data; in some embodiments, k ₂ =1.15, and in this embodiment, the k ₂ =1.05.

In this way, when the sliding window is a motion interval, it is judged whether the time t satisfies t ₁ ≤t≤t ₂ , and when the size of the time t does not satisfy t ₁ ≤t≤t ₂ , it is determined whether the time t satisfies t 1 ≤t≤t 2 . The initial angular velocity energy threshold T _ω is adaptively adjusted, and the initial angular velocity energy threshold T _ω is adjusted by asymptotic approximation. When the size of the time t satisfies t ₁ ≤ t ≤ t ₂ , it is determined that the sliding window meets the The law of the motion interval, the sliding window is determined as the motion interval, and the detection of the next sliding window is performed, which has high reliability and strong practicability.

Specifically, when the interval state corresponding to the sliding window is the zero-speed interval, according to the movement characteristics of the person, in any gait, the zero-speed interval will not be too short, and by judging whether the size of the time t satisfies the third The preset time determines whether the initial angular velocity energy threshold T _ω needs to be adjusted, and the judgment basis is reliable. The third preset time t ₃ is calibrated according to experimental data. In some embodiments, the third preset time t ₃ : t ₃ =1.5s, and in this embodiment, the time t ₃ =0.5s .

Specifically, when the interval state corresponding to the sliding window is a zero-speed interval, the adaptive adjustment to the initial angular velocity energy threshold T _ω is:

Increase the initial angular velocity energy threshold T _ω .

Specifically, in some embodiments, when the size of the time t does not satisfy: t≥t ₃ , the adjustment method of the initial angular velocity energy threshold T _ω is: T _ω =k ₃ *T _ω , where k ₃ > 1, k ₃ is a constant, which is calibrated by experimental data; in other embodiments, the adjustment method of the initial angular velocity energy threshold T _ω is:

Wherein, k ₄ and k ₅ are integers, which are calibrated by experimental data. Preferably, k ₄ +k ₅ ≤L. Specifically, in some embodiments, k ₄ =L, k ₅ =0; The initial _angular velocity energy threshold T _ω is adjusted by asymptotic approximation, which has high reliability and strong practicability.

In this way, when the interval state corresponding to the sliding window is in the zero-speed interval, it is judged whether the time t satisfies t≥t ₃ , and when the size of the time t does not satisfy t ≥ t ₃ , it is adjusted by means of asymptotic approximation The initial angular velocity energy threshold T _ω avoids the occurrence of a leaking zero velocity state detection point. When the time t satisfies t ≥ t ₃ , it is determined that the sliding window conforms to the law of the zero velocity interval, and the sliding window is determined to be The initial zero-speed interval has high reliability and strong practicability.

The determination of the zero-speed interval according to the adjustment times of the initial angular velocity energy threshold T _ω includes:

Count the adjustment times of the initial angular velocity energy threshold T _ω , and judge whether the adjustment times of the initial angular velocity energy threshold T _ω exceed the preset value: No, go back to step S3; Yes, judge whether the sliding window is in the zero-speed interval : Yes, the sliding window is determined to be the initial zero-speed interval; No, go back to step S3, and perform zero-speed angular velocity detection on the next sliding window.

Specifically, the preset value may be determined according to an experimental value, or may be designated manually. In this way, when the number of times of adjustment of the initial angular velocity energy threshold T _ω exceeds a preset value, the adjustment of the initial angular velocity energy threshold T _ω is stopped, avoiding a large number of operations, reducing the difficulty of data processing, and at the same time, it can also be adjusted to a certain extent. In order to ensure the reliability of the initial zero-speed interval data, the reliability is high and the practicability is strong. It should be understood that, in the detection of the same sliding window, regardless of whether the sliding window is in the motion interval or the zero velocity interval, the adjustment times of the initial angular velocity energy threshold T _ω are counted.

In some embodiments, the S5 step includes:

S510: Determine an acceleration amplitude threshold for acceleration zero-speed detection according to the acceleration information corresponding to the detection points in the initial zero-speed interval in the initial zero-speed interval;

S520: Perform acceleration zero-speed detection on all the initial zero-speed interval detection points according to the acceleration amplitude threshold, and correct the initial zero-speed interval to obtain a corrected zero-speed interval.

Specifically, the step S510 includes:

S511: Calculate acceleration amplitudes corresponding to all the initial zero-speed interval detection points according to the acceleration information corresponding to the initial zero-speed interval detection points.

Specifically, the acceleration amplitude calculation formula is:

和

分别为所述惯性器件在第i个检测点获取到的加速度信息。in,

and

are the acceleration information obtained by the inertial device at the i-th detection point, respectively.

S512: Sort the acceleration amplitudes corresponding to all the detection points in the initial zero-speed interval, and generate a temporary interval according to the sorted acceleration amplitudes.

1/4≤k₆＜1/2，k₆由实验数据标定，在一些实施例中，k₆＝1/4，也就是说，Q为大于等于L/2的最小整数，R为大于等于L/4的最小整数。Specifically, the sorting may be in ascending/descending order, the L is the number of detection points in the initial zero-speed interval, and the acceleration amplitudes corresponding to all the detection points in the initial zero-speed interval are in ascending/descending order Arranged, the corresponding L acceleration amplitudes after sorting are respectively denoted as f′ ₁ , f′ ₂ ,...,f′ _L , and the acceleration amplitudes corresponding to the temporary interval are respectively denoted as f′ _Q-R+1 , f′ _Q-R+2 ,…,f′ _Q+R , where,

1/4≤k ₆ <1/2, k ₆ is calibrated by experimental data, in some embodiments, k ₆ =1/4, that is, Q is the smallest integer greater than or equal to L/2, R is greater than or equal to The smallest integer of L/4.

S513: Determine the acceleration amplitude threshold according to the acceleration amplitude corresponding to the detection point in the temporary interval in the temporary interval.

Specifically, the step S513 includes:

Calculate acceleration amplitude mean μ _f and acceleration amplitude standard deviation σ _f according to the acceleration amplitude corresponding to the detection point of the temporary interval in the temporary interval;

The acceleration amplitude threshold is calculated according to the acceleration amplitude mean μ _f and the acceleration amplitude standard deviation σ _f .

Specifically, the calculation formula of the mean value of acceleration amplitude μ _f is:

The calculation formula of the standard deviation σ _f of the acceleration amplitude is:

Specifically, the acceleration amplitude threshold includes an upper acceleration amplitude threshold T _{f above} and an acceleration amplitude upper threshold T _{f below} , specifically,

The acceleration amplitude upper threshold T _f : T _{f upper} = μ _f +λ ₂ σ _f ,

The acceleration amplitude upper threshold T _{f is below} : T _{f below} = μ _f -λ ₃ σ _f ,

Wherein, λ ₂ and λ ₃ are constants, which are calibrated by experimental data, in some embodiments, λ ₂ =3, λ ₃ =3.

Specifically, the step S520 includes:

Acceleration zero-speed detection is performed on the initial zero-speed interval detection points in sequence according to the acceleration amplitude threshold value, and all the initial zero-speed interval detection points in the zero-acceleration state are obtained.

Specifically, according to the acceleration amplitude upper threshold T _f and the acceleration amplitude upper threshold T _f , the acceleration zero-speed detection is performed on the initial zero-speed interval detection points in sequence from front to back, and the acceleration zero-speed detection formula is: :

为j时刻的运动状态，1表示零加速状态，0表示运动状态。of which, of which

is the motion state at time j, 1 represents the zero acceleration state, and 0 represents the motion state.

The modified zero-speed interval is generated according to the initial zero-speed interval detection points of all zero-acceleration states.

1/8≤k₇＜1/4。但是在后续进行行人定位导航时，采用处于零加速状态的连续所述初始零速区间检测点进行定位导航。It should be noted that, under normal circumstances, the continuous initial zero-speed interval detection points of all zero-acceleration states constitute the corrected zero-speed interval, but, in some extreme cases, the initial zero-speed intervals of all zero-acceleration states The detection points are not completely continuous. At this time, if the initial zero-speed interval detection points of the motion state are separated from the initial zero-speed interval detection points of the adjacent zero acceleration states, only the initial zero-speed interval detection points of the motion state are required. The number of detection points in the speed interval does not exceed the preset value S, and it is considered that the initial zero-speed interval detection points in the motion state belong to the corrected zero-speed interval, and the preset value S is calibrated according to experimental data. In some embodiments, the The preset value S:

1/8≤k ₇ <1/4. However, in the subsequent pedestrian positioning and navigation, the continuous initial zero-speed interval detection points in the zero-acceleration state are used for positioning and navigation.

It should be noted that after the correction of the zero-speed interval, return to step S3 to perform the angular velocity zero-velocity detection of the next sliding window. At this time, the starting point of the next sliding window is after the initial zero-velocity interval. The first detection point.

In this embodiment, the IMU is tied to the instep, and the Raspberry Pi is fixed to the waist to exercise on the 80m*60m rectangular playground. Four groups of data were collected in the experiment, of which two groups were collected during normal walking, moving 193 steps and 204 steps respectively, denoted as W_193 and W_204; the other two groups were collected during mixed exercise, moving 133 steps and 145 steps respectively. Steps are recorded as Hunhe_133 and Hunhe_145. At the same time, the three-condition method is used for zero-speed detection. The detection results are shown in Table 1. Analysis of the data in Table 1 shows that when the motion gait is a single walking, the adaptive algorithm and the three conditions The average missed detection rates of the two methods were 0.505% and 1%, respectively; when the motion gait was mixed, the average missed detection rates of the two methods were 0.375% and 28.925%, respectively.

Whether it is a single movement or a mixed movement, the missed detection rate of the number of steps of the adaptive algorithm is lower than that of the three-condition method. When the gait is relatively simple, both can accurately detect the number of steps, but the accuracy of the adaptive algorithm is higher than that of the three-condition method. higher; when the gait is changeable, the adaptive algorithm can still detect the number of steps more accurately, but the three-condition rule has many missed steps and cannot accurately detect the number of steps; it proves that this method can not only accurately detect the number of steps The number of steps, and the adaptability to the gait is strong.

Table 1 Comparison of the step number detection results between the adaptive algorithm and the three-condition method

Note:

The self-adaptive zero-speed interval detection method of the present invention first determines the initial angular velocity energy threshold T _ω through the angular velocity information output by the inertial device, so that the initial angular velocity energy threshold T _ω matches the pedestrian's initial gait; The initial angular velocity energy threshold T _ω performs angular velocity zero-velocity detection on the sliding window, determines the number L of detection points in the sliding window and the interval state corresponding to the sliding window, and according to the interval state and the number of detection points L The initial angular velocity energy threshold T _ω is adaptively adjusted, and the initial angular velocity energy threshold T _ω is adjusted in a progressive and urgent manner to obtain the initial zero velocity interval based on the initial angular velocity energy threshold, which reduces the leakage of the zero velocity state detection point. Then, the initial zero-speed interval is corrected through the acceleration zero-speed detection to obtain the more accurate corrected zero-speed interval. The present invention can gradually adjust the initial angular velocity energy threshold T _ω in the process of pedestrian movement, can realize multi-gait zero-speed interval detection, has high reliability and strong practicability.

Although the present disclosure is disclosed above, the scope of protection of the present disclosure is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present disclosure, and these changes and modifications will fall within the protection scope of the present invention.

Claims (8)

1. An adaptive zero-speed interval detection method is characterized by comprising the following steps:

s1: acquiring output data of an inertial device in real time, wherein the output data comprises angular velocity information and acceleration information of the inertial device;

s2: determining an initial angular velocity energy threshold T of angular velocity zero-speed detection according to the angular velocity information_ω；

S3: according to the initial angular velocity energy threshold value T_ωCarrying out angular velocity zero-speed detection on a sliding window for angular velocity zero-speed detection, and determining the number L of detection points in the sliding window and an interval state corresponding to the sliding window;

s4: calculating the time T corresponding to the sliding window according to the number L of the detection points, and carrying out energy threshold T on the initial angular velocity according to the time T and the interval state corresponding to the sliding window_ωCarrying out self-adaptive adjustment to obtain an initial zero-speed interval based on the initial angular speed energy threshold value:

when the interval state corresponding to the sliding window is a motion interval, judging whether the time t meets t or not₁≤t≤t₂Wherein, t₁Is a first predetermined time, t₂For a second preset time: returning to the step of S3, performing angular velocity zero-speed detection on the next sliding window; no, for the initial angular velocity energy threshold T_ωPerforming self-adaptive adjustment and according to the initial angular velocity energy threshold value T_ωThe zero-speed interval is judged according to the adjustment times;

when the interval state corresponding to the sliding window is a zero-speed interval, judging whether the time t meets the condition that t is more than or equal to t₃Wherein t is₃For a third preset time: if yes, the sliding window is determined as the initial zero-speed interval; no, for the initial angular velocity energy threshold T_ωPerforming self-adaptive adjustment and according to the initial angular velocity energy threshold value T_ωThe zero-speed interval is judged according to the adjustment times;

s5: and carrying out acceleration zero-speed detection on the initial zero-speed interval according to the acceleration information, and correcting the initial zero-speed interval to obtain a corrected zero-speed interval.

2. The adaptive zero-velocity interval detection method as claimed in claim 1, wherein the energy threshold T for the initial angular velocity is set when the interval status corresponding to the sliding window is a motion interval_ωThe step of performing adaptive adjustment comprises:

when the time t satisfies the condition that t is less than t₁While decreasing said initial angular velocity energy threshold T_ω；

When the size of the time t satisfies t > t₂While increasing the initial angular velocity energy threshold T_ω。

3. The adaptive zero-velocity interval detection method of claim 1, wherein the energy threshold T for the initial angular velocity is set when the interval status corresponding to the sliding window is a zero-velocity interval_ωThe self-adaptive adjustment is as follows:

increasing the initial angular velocity energy threshold T_ω。

4. The adaptive interval zero-velocity detection method of claim 1, wherein said detecting is based on said initial angular velocity energy threshold T_ωThe zero-speed interval judgment of the adjustment times comprises the following steps:

counting the initial angular velocity energy threshold T_ωJudging the initial angular velocity energy threshold value T_ωWhether the adjustment times exceed a preset value: otherwise, go back to step S3; if yes, judging whether the sliding window is a zero-speed interval: if yes, the sliding window is determined as the initial zero-speed interval; otherwise, the process returns to step S3, and angular velocity zero speed detection is performed for the next sliding window.

5. The adaptive zero-speed interval detection method of claim 1, wherein the step of S5 comprises:

s510: determining an acceleration amplitude threshold value of acceleration zero-speed detection according to the acceleration information corresponding to the initial zero-speed interval detection point in the initial zero-speed interval;

s520: and performing acceleration zero-speed detection on all the initial zero-speed interval detection points according to the acceleration amplitude threshold, and correcting the initial zero-speed interval to obtain the corrected zero-speed interval.

6. The adaptive zero-speed interval detection method of claim 5, wherein the step S510 comprises:

s511: calculating acceleration amplitudes corresponding to all the initial zero-speed interval detection points according to the acceleration information corresponding to the initial zero-speed interval detection points;

s512: sequencing the acceleration amplitudes corresponding to all the initial zero-speed interval detection points, and generating a temporary interval according to the sequenced acceleration amplitudes;

s513: and determining the acceleration amplitude threshold according to the acceleration amplitude corresponding to the detection point of the temporary interval in the temporary interval.

7. The adaptive zero-velocity interval detection method of claim 6, wherein the step S520 comprises:

sequentially carrying out acceleration zero-speed detection on the initial zero-speed interval detection points according to the acceleration amplitude threshold value to obtain all the initial zero-speed interval detection points in a zero-acceleration state;

and generating the corrected zero-speed interval according to the initial zero-speed interval detection points in all zero acceleration states.

8. The adaptive zero-speed interval detection method according to claim 1, wherein the step S2 includes:

s210: calculating the angular velocity energy of the first n (n is more than or equal to 1) detection points according to the angular velocity information;

s220: calculating the angular velocity energy mean value mu according to the angular velocity energy_ωSum angular velocity energy standard deviation sigma_ω；

S230: according to the angular velocity energy mean value mu_ωAnd the angular velocity energy scaleTolerance sigma_ωDetermining the initial angular velocity energy threshold T_ω。

Images