KR102583899B1 - Apparatus and method for indoor positioning of pedestrians - Google Patents

Abstract

The present invention includes an IMU movement estimation unit that determines the pedestrian movement pattern based on a sensing signal that detects the pedestrian's steps and rotation angles, and transmits and receives wireless signals to multiple APs to measure the distance to each of the multiple APs at each step. A distance information acquisition unit that measures in a designated manner, sets a number of local coordinate systems with the location of each of the multiple APs as the origin, and uses the pedestrian movement pattern and the measured distance to each of the multiple APs on the multiple local coordinate systems. Based on the acquired local movement trajectory, the bias generated by NLOS is calculated from the stride length and measured distance according to the pedestrian's steps, and the local position determines the initial local step position of the local movement trajectory based on the calculated stride length and bias. Including an estimation unit and a global position estimation unit that converts the local movement trajectory obtained in each of a plurality of local coordinate systems into a movement trajectory in the global coordinate system to obtain the pedestrian movement trajectory in the global coordinate system, accurately determining the user's location and movement trajectory indoors. We provide a pedestrian indoor location positioning device and method that can not only estimate but also estimate the exact location and movement trajectory even when measurement errors exist.

Description

Apparatus and method for indoor positioning of pedestrians}

The present invention relates to an apparatus and method for indoor positioning of pedestrians, and to an apparatus and method for indoor positioning of pedestrians using wireless signals and inertial measurement sensors.

Pedestrian dead reckoning technology, also known as positioning, is increasingly important in modern smart cities in that it obtains information that can provide various services appropriate to the user's current environment. In particular, recently, there has been a demand for a positioning technique that can accurately detect not only the user's current location but also the user's mobility trajectory.

Previously, various methods were proposed for pedestrian location estimation technology, such as using GPS (Global Positioning System), using wireless signals, and using various sensors. However, most of these existing pedestrian location estimation technologies focus on guessing the location in an outdoor environment, and have the limitation that it is difficult to estimate the exact location of a pedestrian in an indoor environment.

In the case of the GPS method, since satellite signals must be used, not only can it not be used in indoor environments where it is difficult to receive satellite signals, but there is a problem that the error range is very large. The method of using wireless signals is based on the received signal strength (RSS) or round-trip time (Round-trip time) of the wireless signals between multiple access points (APs) and user equipment (UEs) held by pedestrians. Trip Time (hereinafter referred to as RTT) is measured to determine the location of the pedestrian based on the distance from each identified AP to the UE. However, in indoor environments, wireless signals are received not only through line-of-sight (LOS) but also through non-line-of-sight (NLOS), so accurate positioning is difficult. In particular, there is a problem that positioning becomes more difficult if the locations of multiple APs are densely located or are located in a similar direction to the UE. Additionally, even when using various sensors, errors in direction estimation occur in indoor environments due to various surrounding factors. Currently, magnetic field sensors are mainly used to estimate movement direction, but magnetic fields are frequently distorted indoors, which has the limitation of making accurate positioning difficult.

Korea Registered Patent No. 10-1622536 (registered on May 13, 2016)

The purpose of the present invention is to provide an indoor location positioning device and method for pedestrians that can accurately estimate the movement trajectory of pedestrians indoors.

Another object of the present invention is to provide an indoor location positioning device and method for pedestrians that can accurately estimate the location and movement trajectory of pedestrians even when measurement errors occur.

In order to achieve the above object, a pedestrian indoor location positioning device according to an embodiment of the present invention includes an IMU movement estimation unit that determines a pedestrian movement pattern based on a sensing signal that detects the pedestrian's steps and rotation angles; a distance information acquisition unit that transmits and receives wireless signals to and from multiple APs and measures the distance to each of the multiple APs in a predetermined manner at each step; Set a plurality of local coordinate systems with the location of each of the plurality of APs as the origin, and based on the local movement trajectory obtained using the pedestrian movement pattern and the measured distance to each of the plurality of APs on the plurality of local coordinate systems. a local position estimation unit that calculates a bias caused by NLOS from the stride length and measured distance according to the pedestrian's steps, and determines an initial local step position of the local movement trajectory based on the calculated stride length and bias; and a global position estimation unit that converts the local movement trajectory obtained in each of the plurality of local coordinate systems into a movement trajectory in the global coordinate system to obtain the pedestrian movement trajectory in the global coordinate system.

The IMU movement estimation unit includes an acceleration sensor that detects vertical acceleration according to the steps of the pedestrian and generates an acceleration signal; A gyro sensor that generates a gyro signal according to the pedestrian's rotation angle; And it may include a movement pattern estimation unit that receives the acceleration signal and the gyro signal as the sensing signal and analyzes the pedestrian's steps and direction changes at each step to obtain the pedestrian's movement pattern.

The local location estimation unit sets the location of each of the plurality of APs as an origin based on the distance to the plurality of APs, and uses the movement direction at the initial step position (p ₁ ) of the pedestrian movement pattern as the first axis of the coordinate system. a local coordinate system conversion unit that sets a plurality of local coordinate systems and converts the pedestrian movement pattern into a local movement trajectory of the local coordinate system; a trajectory estimation unit that calculates the stride length and the bias by distinguishing the local movement trajectory depending on whether it is a linear movement trajectory or a non-linear movement trajectory; and a local initial position calculation unit that determines an initial local step position in a local coordinate system based on the calculated stride length and the bias.

The trajectory estimation unit includes a local step selection unit that selects a step of a step position located within a predetermined threshold distance from the position of the corresponding AP among a plurality of local step positions of the local movement trajectory as a local step adjacent to the AP; If the local movement trajectory is a straight movement trajectory, two different random steps are selected in various combinations from the local steps adjacent to the AP, and the distance difference from the corresponding AP in the selected step and the distance difference between the corresponding AP and the remaining steps are calculated. a straight line trajectory estimation unit that estimates the stride length and bias according to the straight line movement trajectory in the local coordinate system based on the straight line trajectory; And if the local movement trajectory is a non-linear movement trajectory, setting an initial local angle that is the distance from the AP corresponding to the local coordinate system to the initial local step position and the angle from the first axis of the local coordinate system to the initial local step position direction, Two different random steps are selected in various combinations from the local steps adjacent to the AP, and a linear movement trajectory is formed in the local coordinate system based on the distance difference between the AP and the corresponding AP in the selected step and the distance difference between the corresponding AP and the remaining steps. It may include a non-linear trajectory estimation unit that estimates the stride length and bias according to the step length and calculates and obtains the initial local angle based on the estimated stride length and bias.

A pedestrian indoor location positioning method according to another embodiment of the present invention for achieving the above object includes determining a pedestrian movement pattern based on a sensing signal that detects the pedestrian's steps and rotation angles; Transmitting and receiving wireless signals with multiple APs and measuring the distance to each of the multiple APs in a predetermined manner at each step; Set a plurality of local coordinate systems with the location of each of the plurality of APs as the origin, and based on the local movement trajectory obtained using the pedestrian movement pattern and the measured distance to each of the plurality of APs on the plurality of local coordinate systems. Calculating a bias generated by NLOS from the stride length and measured distance according to the step of the pedestrian, and determining an initial local step position of the local movement trajectory based on the calculated stride length and bias; and converting the local movement trajectory obtained in each of the plurality of local coordinate systems into a movement trajectory in the global coordinate system to obtain the pedestrian movement trajectory in the global coordinate system.

Therefore, the pedestrian indoor location positioning device and method according to an embodiment of the present invention utilizes wireless signals and inertial measurement sensors with multiple APs in an indoor environment where errors are likely to occur due to surrounding factors to determine the user's location and movement trajectory. Not only can it be accurately estimated, but it is also possible to estimate the exact location and movement trajectory even when measurement errors exist.

Figure 1 shows a schematic configuration of a pedestrian indoor positioning system according to an embodiment of the present invention.
Figure 2 shows a schematic structure of a pedestrian indoor positioning device according to an embodiment of the present invention.
Figure 3 shows an example of a sensing signal in which the IMU movement tracking unit detects the pedestrian steps and rotation angle using the IMU sensor.
Figure 4 shows an example of a pedestrian's movement trajectory estimated using the IMU sensor of Figure 3.
FIG. 5 shows an example of the detailed configuration of the local location estimation unit of FIG. 1.
Figure 6 shows an example of a local movement trajectory converted to a local coordinate system.
FIG. 7 is a diagram illustrating a method by which the global position estimation unit estimates the initial position and initial movement direction in a linear movement trajectory.
Figure 8 shows a method for determining the indoor location of a pedestrian according to an embodiment of the present invention.

In order to fully understand the present invention, its operational advantages, and the objectives achieved by practicing the present invention, reference should be made to the accompanying drawings illustrating preferred embodiments of the present invention and the contents described in the accompanying drawings.

Hereinafter, the present invention will be described in detail by explaining preferred embodiments of the present invention with reference to the accompanying drawings. However, the present invention may be implemented in many different forms and is not limited to the described embodiments. In order to clearly explain the present invention, parts that are not relevant to the description are omitted, and the same reference numerals in the drawings indicate the same members.

Throughout the specification, when a part "includes" a certain element, this does not mean excluding other elements, unless specifically stated to the contrary, but rather means that it may further include other elements. In addition, terms such as "... unit", "... unit", "module", and "block" used in the specification refer to a unit that processes at least one function or operation, which is hardware, software, or hardware. and software.

Figure 1 shows a schematic configuration of a pedestrian indoor positioning system according to an embodiment of the present invention.

Referring to FIG. 1, the pedestrian indoor positioning system according to this embodiment includes at least one user terminal (UE) and a plurality of access points (AP1 to AP4). In this embodiment, at least one user terminal (UE) is a mobile device that a pedestrian can carry and move, and may include a pedestrian indoor location determination device. The pedestrian indoor positioning device is equipped with an inertial momentum unit (IMU) sensor to track the movement trajectory of the pedestrian. In addition, multiple APs (AP1 to AP4) each transmit and receive various wireless signals such as WiFi with the user terminal (UE), detect the distance (r) between each AP and the user terminal (UE), and calculate the detected distance (r). The global coordinate information of the and AP can be transmitted to the pedestrian indoor location determination device provided in the user terminal (UE).

In this embodiment, the pedestrian indoor location positioning device uses the pedestrian's movement trajectory obtained from the IMU sensor, the global coordinate information of each AP obtained from multiple APs (AP1 to AP4), and the distance (r) to the user terminal (UE). Based on this, it accurately tracks the pedestrian's indoor location and movement trajectory according to location movement.

As mentioned above, when only information obtained from a pedestrian indoor location determination device is used, there is a problem that errors occur in direction estimation due to magnetic distortion, etc., and the distance (r) obtained from multiple APs (AP1 to AP4) This is because if only information is used, large errors may occur due to the placement location of APs (AP1 to AP4) and the NLOS path. Therefore, in this embodiment, the location information obtained from the pedestrian indoor location determination device and the location information obtained through multiple APs (AP1 to AP4) are integrated to estimate the location and movement trajectory of the pedestrian indoors very accurately.

Figure 2 shows a schematic structure of a pedestrian indoor positioning device according to an embodiment of the present invention, and Figure 3 shows an example of a sensing signal in which the IMU movement tracking unit detects the pedestrian steps and rotation angles using the IMU sensor, Figure 4 shows an example of a pedestrian's movement trajectory estimated using the IMU sensor of Figure 3.

Referring to FIG. 2, the pedestrian indoor location positioning device according to this embodiment includes an IMU movement estimation unit 100, a distance information acquisition unit 200, a local location estimation unit 300, and a global position estimation unit 400. can do.

The IMU movement estimation unit 100 is configured to track the movement trajectory of a pedestrian based on an IMU sensor and may include a sensor unit 110 and a movement pattern estimation unit 120. In this embodiment, the sensor unit 110 detects the position and direction change for each step as the pedestrian moves and outputs a sensing signal. The sensor unit 110 is an IMU sensor configuration and may include an acceleration sensor 111 and a gyro sensor 112, and may additionally include other sensors such as a magnetic field sensor or a GPS sensor.

The acceleration sensor 111 detects the pedestrian's steps according to the repetitive up and down acceleration pattern that occurs when the pedestrian walks, and the gyro sensor 112 detects the change in direction when the pedestrian walks and outputs this as a sensing signal. .

In FIG. 3 , (a) shows an example of an acceleration signal output from the acceleration sensor 111, and (b) shows an example of a gyro signal output from the gyro sensor 112. And FIG. 4 shows the pedestrian's movement pattern estimated based on the acceleration and gyro signals detected by the acceleration sensor 111 and the gyro sensor 112.

As shown in (a) of FIG. 3, the acceleration sensor 111 can output an acceleration signal with a peak corresponding to the vertical acceleration that occurs repeatedly as the pedestrian walks, and (b) of FIG. 3. As shown, the gyro sensor 112 is not greatly affected by the pedestrian's walking, but can output a gyro signal that significantly changes when the direction changes according to the pedestrian's turn.

The movement pattern estimation unit 120 receives the sensing signal including the acceleration signal output from the acceleration sensor 111 and the gyro signal output from the gyro sensor 112 and analyzes it according to a predetermined method to determine whether the pedestrian steps have occurred. Detect whether the direction changes due to rotation, and if a change of direction occurs, check the rotation angle (μ).

As shown in FIG. 4, the movement pattern estimation unit 120 determines only whether a step has occurred based on the acceleration signal, and can estimate the pedestrian's movement pattern by assuming that the stride length (d) in each step is uniform. there is. In the case of the gyro sensor 112, since it operates regardless of magnetic distortion, etc., it can accurately detect the pedestrian's change of direction, that is, the rotation angle (μ), even indoors. However, it is difficult to accurately detect changes in stride length with the acceleration sensor 111. However, in general, the stride length (d) of a pedestrian is almost constant. Accordingly, the movement pattern estimation unit 120 assumes that the stride length (d) is uniform, and while the pedestrian moves from the initial position (p ₁ ) to the position for each step (p ₂ to p 8), the predetermined same stride length (p ₈ ) is assumed to be uniform. The pedestrian's movement pattern can be estimated by assuming that it moves as much as d). In the movement pattern estimation unit 120, the stride length d may be designated as an arbitrary distance and may be obtained through subsequent calculation.

And _the movement pattern estimation unit ₁₂₀ has _a predetermined angle range (e.g., If it is within -3°≤ μ≤3°), it is judged to be within the error range and the pedestrian is determined to have moved on a straight trajectory, and as shown in FIG. 4, a specific step position (the fourth and sixth positions in FIG. 4) is determined to be within the error range. If the rotation angle (μ) at p ₄ , p ₆ )) exceeds the angle range, it can be determined that the pedestrian has moved in a non-straight trajectory.

Therefore, if the pedestrian's initial position (p ₁ ) and the initial movement direction (w) and stride length (d) based on the x axis can be confirmed in the global coordinate system with x and y axes, the position for each step (p ₂ ~ p ₈ ) can be determined. It can be estimated accurately. In other words, the pedestrian's movement trajectory can be accurately determined.

The distance information acquisition unit 200 transmits and receives wireless signals with each of the plurality of surrounding APs (AP ₁ to AP _M ) in a predetermined manner, and transmits and receives wireless signals from each of the plurality of APs (AP ₁ to AP _M ) to the pedestrian indoor location determination device. Check the distance (r ₁ ~ r _M ) and obtain location information in the global coordinate system for each of multiple APs (AP ₁ ~ AP _M ). Each of multiple APs (AP ₁ to AP _M ) is placed at a fixed location. Therefore, the location information in the global coordinate system of each of the multiple APs (AP ₁ to AP _M ) can be stored in advance in each AP (AP ₁ to AP _M ) and transmitted to the pedestrian indoor location determination device. In addition, each of multiple APs (AP ₁ to AP _M ) transmits and receives wireless signals to and from pedestrian indoor positioning devices to measure Received Signal Strength (RSS) or Round-Trip Time (RTT). By measuring, you can check the distance (r ₁ ~ r _M ) between each AP (AP ₁ ~ AP _M ) and the pedestrian indoor positioning device. Here, it is assumed that the distance (r ₁ ~ r _M ) is confirmed using RTT, but in some cases, RSS may be used. And each of the multiple APs (AP ₁ to AP _M ) transmits the confirmed distance (r ₁ to r _M ) to the pedestrian indoor location determination device. However, in some cases, the pedestrian indoor location determination device may be configured to directly measure the RTT to check the distance (r ₁ to r _M ) to multiple APs (AP ₁ to AP _M ).

The distance information acquisition unit 200 may check the distance (r ₁ to r _M ) from each step position (p ₁ to p ₈ ) of the pedestrian for each of the multiple APs (AP ₁ to AP _M ).

At this time, the distance (r ₁ to r _M ) between each AP (AP ₁ to AP _M ) and the pedestrian indoor location determination device may include a bias (b) due to NLOS, as shown in FIG. 1. And the distance (r ₁ ~ r _M ) including the bias (b) does not represent the exact distance between the AP (AP ₁ ~ AP _M ) and the pedestrian indoor location determination device.

Accordingly, the local position estimation unit 300 converts the pedestrian movement pattern estimated by the movement pattern estimation unit 120 into a local movement trajectory in a local coordinate system based on the positions of each of the multiple APs (AP ₁ to AP _M ). Thus, the bias (b) due to NLOS is calculated along with the arbitrarily designated stride length (d).

The local position estimation unit 300 converts the pedestrian's movement pattern estimated by the movement pattern estimation unit 120 into a local movement trajectory on a local coordinate system with the positions of each of the multiple APs (AP ₁ to AP _M ) as the origin, Stride length (d) and bias (b) are calculated in different ways depending on whether the local movement trajectory is a straight or non-linear trajectory.

FIG. 5 shows an example of the detailed configuration of the local position estimation unit of FIG. 1, and FIG. 6 shows an example of a local movement trajectory converted to a local coordinate system.

Referring to FIG. 5 , the local position estimation unit 300 may include a local coordinate system conversion unit 310, a trajectory estimation unit 320, and a local initial position calculation unit 330.

First, the local coordinate system conversion unit 310 uses the pedestrian movement pattern estimated by the movement pattern estimation unit 120 at the location of each of the multiple APs (AP ₁ to AP _M ) as the origin, instead of the x and y axes of the global coordinate system. In the pedestrian's movement trajectory, a number of local coordinate systems are set, with the initial movement direction in the q-axis direction and the u-axis perpendicular to the q-axis, and the initial position (p ₁ ) of the pedestrian's movement pattern corresponds to the AP (AP _m ) The pedestrian's movement pattern is converted into a local movement trajectory in the local coordinate system so that it is placed at a location spaced apart by a distance (r _m,1 ) from ).

Since the local coordinate system conversion unit 310 configures a local coordinate system for each of a plurality of APs (AP ₁ to AP _M ), the local movement trajectory is also a trajectory on the local coordinate system corresponding to the number of APs (AP ₁ to AP _M ). It appears as That is, M local coordinate systems and M local movement trajectories corresponding thereto are obtained.

Here, for convenience of explanation, the local coordinate system according to one AP (AP _m ) among M local coordinate systems is described as an example. Hereinafter, the identifier (m) that distinguishes multiple APs (AP ₁ to AP _M ), that is, The identifier (m) for the local coordinate system is omitted.

At this time, as shown in (a) and (b) of FIG. 6, the local coordinate system conversion unit 310 ensures that the movement direction at the initial position (p ₁ ) in the pedestrian's movement pattern is the q axis of each local coordinate system. , the local coordinate system can be viewed as a configuration rotated by an angle corresponding to the initial movement direction (w) so that the initial movement direction (w) in the global coordinate system of Figure 4 is the q-axis direction, and the N step positions in the global coordinate system (p ₁ ~ p _N ) can be re-expressed as N local step positions (z ₁ ~ z _N ) in the local coordinate system. Here, the local coordinates of each local step position (z ₁ ~ z _N ) can be expressed as ((q ₁ , u ₁ ) ~ (q _N , u _N )) along the q-axis and u-axis.

Therefore, each local step position (z ₁ ~ z _N ) according to the local movement trajectory is based on the local coordinates (q ₁ , u ₁ ) of the initial local step position (z ₁ ), stride length (d) and bias (b), and the q axis. It can be organized as shown in Equation 1 by considering the rotation angle (θ) in each of the total N steps.

here ego, am.

When M local coordinate systems are configured by the local coordinate system conversion unit 310, the pedestrian movement pattern is converted, and a local movement trajectory is applied, the trajectory estimation unit 320 determines that the pedestrian movement pattern estimated by the movement pattern estimation unit 120 is Stride length (d) and bias (b) are calculated by re-estimating the local movement trajectory on the local coordinate system by categorizing it according to whether it is a straight trajectory or a non-straight trajectory.

The trajectory estimation unit 320 may include a local step selection unit 321, a linear trajectory estimation unit 322, and a non-linear trajectory estimation unit 323.

The local step selection unit 321 first selects a predetermined threshold distance (r _k ) from the APs (AP ₁ to AP _M ) corresponding to the local coordinate system among all local steps (z ₁ to z _N ) in the local movement trajectory converted to the local coordinate system. ) are grouped into an adjacent step group, and two local steps are selected from all possible combinations from multiple local steps included in the adjacent step group. Here, the local step selection unit 321 selects and outputs two local steps from adjacent step groups in various combinations, and the linear trajectory estimation unit 322 and the non-linear trajectory estimation unit 323 determine the stride length (d) and bias. This is to enable more accurate calculation of (b), and a detailed explanation will be provided later.

When the pedestrian movement pattern is a straight trajectory, the linear trajectory estimation unit 322 calculates the stride length (d) and bias (b) by estimating the local movement trajectory on the local coordinate system, and the non-straight trajectory estimation unit 323 calculates the pedestrian movement trajectory. If the pattern is a non-linear trajectory, the stride length (d) and bias (b) are calculated by estimating the local movement trajectory on the local coordinate system.

The straight-line trajectory estimation unit 322 is activated when the pedestrian movement pattern determined by the movement pattern estimation unit 120 is a straight trajectory, as shown in (a) of FIG. 6, and estimates the pedestrian movement trajectory on the local coordinate system to calculate the stride length (d). ) and bias (b) are obtained.

When the local movement trajectory is a straight trajectory as shown in (a) of FIG. 6, the pedestrian movement direction in the local coordinate system is the q-axis direction, so c _n = (n-1) and s _n = 0. Therefore, Equation 1 can be rewritten as Equation 2.

Equation 2 is a non-linear equation, and to solve it, non-linear items can be removed by selecting two different steps (n), substituting them, and subtracting them from each other. If the equation obtained by substituting the value (a ₁ ) for an arbitrary local step for n in Equation 2 is subtracted from Equation 2, it is calculated as in Equation 3.

Similarly, if the value (a ₂ ) for an arbitrary local step other than a ₁ is substituted for n in Equation 2 and subtracted from Equation 2, an equation obtained similar to Equation 3 is subtracted from Equation 3, It is calculated as in Equation 4.

Equation 4 is a function of two arbitrary local steps (a ₁ , a ₂ ) among local steps (z ₁ to z _N ) in the local coordinate system for a specific AP (AP _m ), and is the square of the stride length (d) (d ² ) and is a linear function for bias (b). Therefore, linear equations for the square of the stride (d ² ) and the bias (b) can be formed in the form of a determinant as shown in Equation 5.

The matrix of Equation 5 is the distance (r 1 to r N ) from each local step (z ₁ to z _N ) as two random local steps (a ₁ , a ₂ ) are selected and the distance (r ₁ to r _N ) from the two selected local steps (a ₁ , a ₂ ) represents the difference between the distances (r _a1 , r _a2 ). Since the stride length (d) is said to be uniform, Equation 5 can be calculated and solved as a matrix for x according to Equation 6.

In Equation 6, x( = [d ² , b] ^T ) is a matrix for stride length (d) and bias (b) in the local coordinate system based on the corresponding AP (AP _m ), so the right side of Equation 6 is Once calculated, the stride length (d) and bias (b) can be obtained. At this time, the element of x obtained is the square of the stride length (d ² ), so the stride length (d) can be calculated as two values, but since the stride length (d) has a positive value, it can be obtained as one value.

However, the determinant of Equation 5 is two arbitrary local steps (a ₁ , a ₂ ), and changes may occur in the calculated x ( = [d ² , b] ^T ) depending on which step is selected. This is because the distance (r ₁ to r _N ) at each local step (z ₁ to z _N ) may include errors due to noise, etc. Therefore, in order to calculate x( = [d ² , b] ^T ) as accurately as possible, an appropriate local step (z ₁ ~ z _N ) that can measure the accurate distance (r ₁ ~ r _N ) is a random local step (a ₁ , a ₂ ) needs to be selected.

Accordingly, in this embodiment, the local step selection unit 321 selects a plurality of local steps included in the adjacent step group within a predetermined threshold distance (r _k ) from the corresponding AP (AP _m ) so that a relatively accurate distance can be obtained. Two local steps (a ₁ , a ₂ ) are extracted and applied in various combinations from the steps, and the straight line trajectory estimation unit 322 performs math on each of the two local steps (a ₁ , a ₂ ) applied in various combinations. Calculate Equation 6 to calculate the stride length (d) and bias (b) for each combination. And by calculating the median value of the stride length (d) and bias (b) calculated for each combination, the representative stride length (d ^* ) and representative bias (b ^* ) for each of multiple APs (AP _m ) can be obtained. there is.

Meanwhile, the non-straight trajectory estimation unit 323 is activated when the pedestrian movement pattern determined by the movement pattern estimation unit 120 is a non-straight trajectory, as shown in (b) of FIG. 6, and estimates the local movement trajectory on the local coordinate system. This obtains the stride length (d) and bias (b).

Even when the pedestrian movement pattern is a non-linear trajectory as shown in (b) of Figure 6, the initial movement direction at the pedestrian's initial local step position (z ₁ ) in the local coordinate system is the q-axis direction, but then the local step position (z ₂ ~ z The direction of movement changes depending on _N ). Here, if the angle from the q-axis direction to the initial local step position (z ₁ ) direction is called the initial local angle (γ), , and the distance from AP (AP _m ), which is the origin of the local coordinate system, to the initial local step position (z ₁ ) is am. Therefore, Equation 1 can be rewritten as Equation 7 by additionally reflecting the local rotation angle (θ _n ) based on the q-axis according to the rotation at each local step (z ₂ ~ z _N ).

Afterwards, like the linear trajectory estimation unit 322, the non-linear trajectory estimation unit 323 also subtracts the equation obtained by substituting the value for an arbitrary local step (a ₁ ) for n in equation 7 from equation 7, , Equation 8 is calculated.

here ego, am.

In addition, if the equation obtained by substituting the value (a ₂ ) for an arbitrary local step other than a ₁ for n in equation 7 is subtracted from equation 8, the nonlinear term 2dRf _n,a (γ) is canceled Equation 9 is obtained.

Equation 9 is also a function of two random local steps (a ₁ , a ₂ ) among a plurality of local steps (z ₁ to z _N ) in the local coordinate system for a specific AP (AP _m ), the square of the stride length (d ² ) and bias (b) can be formed in the form of a determinant such as Equation 10.

The solution to Equation 10 is calculated using Equation 11.

However, in Equation 6, it is a determinant for calculating two variables, stride length (d) and bias (b), whereas in Equation 11, not only stride length (d) and bias (b) but also the initial local angle (γ) must be calculated. do. Accordingly, the nonlinear trajectory estimation unit 323 changes and substitutes the initial local angle γ to obtain an initial local angle γ(S) that minimizes the error e ₂ according to Equation 12.

In equation 12 As, std() is the standard deviation function. and , and R _n,a (S,γ) for each random local step (a ₁ , a ₂ ) can be calculated according to Equation 13 from Equation 8.

Equation 13 is calculated n-1 times for each random local step (a ₁ , a ₂ ), so R(S,γ) for each random local step (a ₁ , a ₂ ) is 2(n-1) It is calculated as , and in Equation 12, the error (e ₂ (S,γ)) is the variance of R(S,γ) for each random local step (a ₁ , a ₂ ), so Equation 12 is the error (e ₂ By searching for γ that minimizes (S, γ)) and substituting it into Equation 11, the stride length (d) and bias (b) can be obtained.

In addition, the non-linear trajectory estimation unit 323 also selects the local step 321 in various combinations from a plurality of local steps included in the adjacent step group within a predetermined threshold distance (r _k ) from the corresponding AP (AP _m ). Calculate Equation 11 for the two extracted local steps (a ₁ , a ₂ ) to obtain the stride length (d) and bias (b) for each combination, and the stride length (d) and bias for each combination obtained. By calculating the median value of (b), the representative stride length (d ^* ) and representative bias (b ^* ) for each of multiple APs (AP _m ) can be obtained.

That is, the trajectory estimation unit 320 calculates a random local step from the AP (AP _m ) located at the origin in the local movement trajectory in which the pedestrian movement pattern is expressed as a trajectory on the local coordinate system _of each of the multiple APs (AP 1 to AP _M ). Stride length (d ^* ) and bias (b ^* ) are obtained based on the distance change of (a ₁ , a ₂ ) and the remaining local steps.

The local initial position calculation unit 330 determines the initial local step position of the local movement trajectory based on the representative stride length (d ^* ) and representative bias (b ^* ) according to the straight and non-linear trajectories estimated by the trajectory estimation unit 320 ( It is obtained by calculating the coordinate values (q ₁ ^* , u ₁ ^* ) of z ₁ ).

When the representative stride length (d ^* ) and representative bias (b ^* ) are obtained from the straight line trajectory estimation unit 322, the local initial position calculation unit 330 calculates the initial local step position (z ₁ ) of the trajectory from Equation 3. The coordinate values (q ₁ ^* , u ₁ ^* ) can be obtained by calculating according to Equations 14 and 15, respectively.

In Equation 15, u ₁ ^* is a value calculated from u ₁ ² , so ambiguity in the sign occurs, but this can be resolved by the global position estimation unit 400, which will be described later.

And when the representative stride length (d ^* ) and representative bias (b ^* ) are obtained from the non-linear trajectory estimation unit 323, the local initial position calculation unit 330 calculates the representative stride length (d ^* ) and representative bias in Equation 8. By substituting (b ^* ), the linear equation for the coordinate values (q ₁ ^* , u ₁ ^* ) of the initial local step position (z ₁ ) of the trajectory is obtained as shown in Equation 16.

And when Equation 16 is converted to determinant form, it is expressed as Equation 17.

From Equation 17, the initial local step position (z ₁ ) in a non-linear trajectory can be calculated according to Equation 18.

The local location estimation unit 300 configures a plurality of local coordinate systems with each of a plurality of APs (AP ₁ to AP _M ) as the origin, and converts the pedestrian movement pattern into a local movement trajectory in each of the plurality of local coordinate systems. After that, it is classified according to whether it is a straight trajectory or a non-straight trajectory, and the distance (r 1 ~ r) between each AP ₍ AP ₁ ~ AP _M ) and each local step position (z ₁ ~ z _N ) on the corresponding local movement trajectory Based on _N ), the stride length (d), bias (b), and coordinate values (q ₁ , u ₁ ) of the initial local step position (z ₁ ) were calculated.

However, each local step position (z ₁ to z _N ) calculated by the local position estimation unit 300 is a position coordinate in the local coordinate system for each AP (AP ₁ to AP _M ). Accordingly, the global position estimation unit 400 converts each of the local step positions (z ₁ to z _N ) calculated as coordinate values on a plurality of local coordinate systems into position coordinate values on the global coordinate system.

Since the initial local step position (z ₁ ) and stride length (d) in each of the plurality of local coordinate systems were calculated in the local position estimation unit 300, the mth AP (AP _m ) among the plurality of APs (AP ₁ to AP _M ) The nth local step position (z _n ) measured based on can be called z _n ^(m)* . And, if the position of the mth AP (AP _m ) in the global coordinate system is P _AP ^(m) and the initial movement direction of the pedestrian is w, the global position estimation unit 400 determines the position according to the mth AP (AP _m ). The local step position (z _n ^(m)* ) on the local coordinate system can be calculated as the step position (p _n ^(m) (w)) according to the pedestrian movement pattern on global coordinates according to Equation 19.

Equation 19 is the position (p _n ^(m) (w)) according to the pedestrian movement trajectory estimated based on the mth AP (AP _m ), and is the location of each local step in each of the multiple APs (AP ₁ to AP _M ) If (z ₁ ~ z _N ) is estimated correctly, the position (p _n ^(m) (w)) along the pedestrian movement trajectory calculated according to Equation 19 should be the same for all APs (AP ₁ ~ AP _M ). , one w ^* that satisfies Equation 20 is calculated.

If Equation 20 is satisfied, the global position estimation unit 400 substitutes the initial direction (w ^* ) into Equation 19 and determines the calculated position (p _n ^(m) (w ^* )) as the pedestrian movement trajectory. .

However, errors may exist in each local step position (z ₁ to z _N ) calculated according to each AP (AP ₁ to AP _M ), so w may be calculated differently.

In this case, the global position estimation unit 400 again distinguishes between cases where the pedestrian movement pattern is a straight movement trajectory and a non-straight movement trajectory, and accurately estimates the initial position (p ₁ ) and initial movement direction (w) in the global coordinate system. do.

FIG. 7 is a diagram illustrating a method by which the global position estimation unit estimates the initial position and initial movement direction in a linear movement trajectory.

The error generated in the global coordinate system with the other two APs (AP ₂ and AP ₃ ) based on the first AP (AP ₁ ) in FIG. 7 is shown. In a straight-line movement trajectory, as explained in Equation 15, ambiguity in the sign may occur at u ₁ ^* among the initial positions (q ₁ , u ₁ ).

That is, the local step position (z ₁ ~ z _N ) in the linear movement trajectory can be obtained by dividing into Z ₊ and Z _- expressed in Equation 21 according to the ambiguity of the sign in u ₁ ^* .

Because of this, as shown in (a) and (b) of FIG. 7, w may be calculated not to be the same.

In order to resolve the above-mentioned ambiguity, the number of APs must be at least 3, and if n APs exist, 2n ambiguities occur because there are 2 ambiguities for each AP. However, Equation 22 states that the relative distance between the two APs (AP _m1 , AP _m2 ) on the global coordinate system and the relative distance between the local step positions (z _n ^(m1) , z _n ^(m2) ) on the corresponding local coordinate system must be the same. This ambiguity can be resolved by using

From Equation 22, the distance between the position (P _AP ( _r ^{)) of the reference AP (AP r} ) and the position (P _AP ^(m) ) of the other AP (AP _r ) is the local step position (P AP (m)) for the reference AP (AP _r ). Since z _n ^(r) ) must be the same as the local step position (z _n ^(m) ) for another AP (AP _r ), the variables with ambiguity according to Equation 23 and Equation 24 are y ₊ ^(r) and It can be reduced to two: y _- ^(r) .

If Equation 23 holds, the global position estimation unit 400 adds z ₊ ⁽ r) to the y ₊ ^(r) set and z _- ^(r) to the y _- ^(r) set. However, if Equation 22 holds, z _- ^(r) is added to the y ₊ ^(r) set, and z ₊ ^(r) is added to the y _- ⁽ r) set. _In this _case , the initial direction of movement ( ^w ) is determined by ^the error ( ) is obtained as an initial movement direction (w ^* ) that satisfies Equation 25 so that is minimized.

On the other hand, for non-linear movement trajectories _, the global position estimator ₄₀₀ calculates the cumulative _error ( ) Obtain the initial movement direction (w ^* ) that satisfies Equation 26 so that is minimized.

here am.

As a result, the pedestrian indoor location positioning device according to this embodiment uses the pedestrian movement pattern acquired based on the IMU sensor provided by the pedestrian indoor location location device to be acquired based on wireless signals with multiple APs (AP ₁ to AP _M ). Convert to a local movement trajectory on a local coordinate system based on distance, calculate the stride length (d) and bias (b) caused by NLOS, etc. in each local coordinate system, and calculate the stride length (d) and bias (b) based on the calculated stride length (d) and bias (b). Accordingly, the accurate pedestrian movement trajectory can be estimated by precisely re-estimating the local movement trajectory in each local coordinate system and then matching the estimated local movement trajectory in each local coordinate system with the global coordinate system.

Figure 8 shows a method for determining the indoor location of a pedestrian according to an embodiment of the present invention.

Referring to FIGS. 1 to 7 , when explaining the pedestrian indoor positioning method of FIG. 8 , first, a sensing signal is acquired using an IMU sensor including an acceleration sensor 111 and a gyro sensor 112 (S11). And based on the acquired sensing signal, the pedestrian movement pattern is determined by analyzing the multiple step positions (p ₁ ~ p _N ) and the rotation angle (μ) at each step according to the pedestrian's steps (S12). At this time, it is assumed that the pedestrian's stride length (d), which represents the interval between each step position (p ₁ to p _N ), is a uniform predetermined interval at each predetermined step.

In addition, wireless signals are transmitted and received with each of the multiple APs (AP ₁ to AP _M ) nearby, and the distance (r ₁ to r _M ) from each of the multiple APs (AP ₁ to AP _M ) to the pedestrian indoor location determination device is measured. (S13). Here, the distance (r ₁ to r _M ) at each step position (p ₁ to p _N ) can also be measured.

Once the pedestrian movement pattern and the distance (r ₁ to r _M ) for each of the multiple APs (AP ₁ to AP _M ) are obtained, each of the multiple APs (AP ₁ to AP _M ) is set as the origin and the initial step position (p ₁₎ ) is set as the first axis of the coordinate system, and the pedestrian movement pattern is converted into a local movement trajectory of the local coordinate system (S14).

And in each local coordinate system, among the multiple local step positions (z ₁ ~ z _N ) of the local movement trajectory, the step (n) corresponding to the step position located within a predetermined threshold distance (r _k ) from the AP, that is, the origin, is selected as a local step adjacent to the AP. Select by step (S15).

Afterwards, it is determined whether the pedestrian movement pattern or local movement trajectory is a linear movement trajectory or a non-straight movement trajectory (S16).

If it is determined to be a straight-line movement trajectory, two different random steps (a ₁ , a ₂ ) are selected in various combinations from local steps adjacent to the AP, and the distance difference from the corresponding AP at the selected step and the correspondence with the remaining steps are determined. Based on the distance difference from the AP, the stride length (d) and bias (b) according to the linear movement trajectory in the local coordinate system are estimated according to Equation 6 (S17). At this time, if multiple strides (d) and bias (b) are obtained depending on the combination of the two selected steps (a ₁ and a ₂ ), the median value of the obtained strides (d) and bias (b) is obtained. You can also calculate the representative stride length (d ^* ) and representative bias (b ^* ).

And when the stride length (d) and bias (b) in the straight-line movement trajectory in the local coordinate system are obtained, the obtained stride length (d) and bias (b), at least one random step (a ₁ , a ₂ ) and each local step Based on the difference between the distance from the AP at the position (z ₁ to z _N ), the initial local step position (z ₁ ) on the local coordinate system of the linear movement trajectory is estimated according to Equations 14 and 15 (S18).

On the other hand, if it is determined to be a non-linear movement trajectory, the distance (R) from the corresponding AP to the initial local step position (z ₁ ) and the initial local step position (z ₁ ) angle from the 1 axis of the local coordinate system are Set the local angle (γ), select two different random steps (a ₁ , a ₂ ) in various combinations from local steps adjacent to the AP, and determine the distance difference from the corresponding AP at the selected step and the correspondence with the remaining steps. Based on the distance difference from the moving AP, the stride length (d) and bias (b) according to the linear movement trajectory in the local coordinate system are estimated according to Equation 11 (S19). At the same time, the initial local angle (γ(S)) is obtained so that the error calculated in a predetermined manner based on the estimated stride length (d) and bias (b) is minimized.

Afterwards, based on the obtained stride length (d), bias (b), and initial local angle (γ(S)), the initial local step position (z ₁ ) on the local coordinate system of the non-linear movement trajectory is estimated according to Equation 18 ( S20).

The initial local step position (z ₁ ), stride length (d) and bias (b), and initial local angle (γ(S)) are obtained in multiple local coordinate systems for each of multiple APs (AP 1 _to AP _M ), and the local Once the movement trajectory is confirmed, each local step position (z ₁ ~ z _N ) on the local movement trajectory according to each AP (AP _m ) is converted back to the step position (p ₁ ^(m) ~ p _N ^(m) ) on the global coordinate system. Convert (S21). Then, it is determined whether the positions of the step positions (p ₁ ^(m) ~ p _N ^(m) ) converted to the global coordinate system match (S22). If they match, the pedestrian's movement trajectory has been accurately estimated in the global coordinate system, and a sensing signal is obtained from the IMU sensor again (S11). However, if they do not match, it is determined that an error has occurred in the pedestrian's movement trajectory in the global coordinate system, and the error in the global step position is compensated (S23). At this time, the error in the global step position can also be compensated according to whether it is a linear or non-linear movement trajectory, and it is determined that an error in the rotation angle (w) occurred in the process of converting the local movement trajectory to the global coordinate system. This compensates for the error in the rotation angle (w). Here, the rotation angle (w) can be viewed as the pedestrian's initial position (p ₁ ) and the initial movement direction (w) based on the x-axis.

If it is a straight movement trajectory, it is judged to be caused by ambiguity of the sign when estimating the second axis coordinate value for the initial local step position (z ₁ ), and different APs (AP _r , AP) in the global coordinate system The distance between the positions (P _AP ^(r) , P _AP ^(m) ) of m ) and the local step position ((z ₁ ^(r) ~ z) for each sign in different local _coordinate systems according to AP (AP _r , AP _m ) Compare the distance between _N ^(r) ), (z ₁ ^(m) ~ z _N ^(m) )) and select a local step position (z ₁ ^(m) ~ z _N ^(m) ) that makes the distance as equal as possible. And, by setting the rotation angle (w) in each local coordinate system accordingly, the error in the global step position can be compensated.

On the other hand, in the case of a non-linear movement trajectory, the difference between the local step positions (z ₁ ^(m) ~ z _N ^(m) ) according to each local movement trajectory converted to the global coordinate system is minimized in each local coordinate system. You can compensate for the error in the global step position by setting the rotation angle (w).

The method according to the present invention can be implemented as a computer program stored on a medium for execution on a computer. Here, computer-readable media may be any available media that can be accessed by a computer, and may also include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules or other data, including read-only memory (ROM). It may include dedicated memory), RAM (random access memory), CD (compact disk)-ROM, DVD (digital video disk)-ROM, magnetic tape, floppy disk, optical data storage device, etc.

The present invention has been described with reference to the embodiments shown in the drawings, but these are merely illustrative, and those skilled in the art will understand that various modifications and other equivalent embodiments are possible therefrom.

Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the attached claims.

100: IMU movement estimation unit 110: sensor unit
111: Acceleration sensor 112: Gyro sensor
120: Movement pattern estimation unit 200: Distance information acquisition unit
300: Local position estimation unit 310: Local coordinate system conversion unit
320: Trajectory estimation unit 330: Local initial position calculation unit
400: Global location estimation unit

Claims (20)

IMU movement estimation unit that determines the pedestrian movement pattern based on the sensing signal that detects the pedestrian's steps and rotation angles;
a distance information acquisition unit that transmits and receives wireless signals to and from multiple APs and measures the distance to each of the multiple APs in a predetermined manner at each step;
Set a plurality of local coordinate systems with the location of each of the plurality of APs as the origin, and based on the local movement trajectory obtained using the pedestrian movement pattern and the measured distance to each of the plurality of APs on the plurality of local coordinate systems. a local position estimation unit that calculates a bias generated by NLOS from the stride length and measured distance according to the steps of the pedestrian, and determines an initial local step position of the local movement trajectory based on the calculated stride length and bias; and
A pedestrian indoor positioning device including a global position estimation unit that converts a local movement trajectory obtained in each of a plurality of local coordinate systems into a movement trajectory in a global coordinate system to obtain a pedestrian movement trajectory in the global coordinate system. The method of claim 1, wherein the IMU movement estimation unit
An acceleration sensor that generates an acceleration signal by detecting the vertical acceleration according to the pedestrian's steps;
A gyro sensor that generates a gyro signal according to the pedestrian's rotation angle; and
A pedestrian indoor location positioning device comprising a movement pattern estimation unit that receives the acceleration signal and the gyro signal as the sensing signal and analyzes the pedestrian's steps and direction changes at each step to obtain the pedestrian movement pattern. The method of claim 1, wherein the local location estimation unit
A plurality of local coordinate systems with the location of each of the plurality of APs as the origin based on the distance to the plurality of APs, and the movement direction at the initial step position (p ₁ ) of the pedestrian movement pattern as the first axis of the coordinate system. a local coordinate system conversion unit that converts the pedestrian movement pattern into a local movement trajectory in a local coordinate system;
a trajectory estimation unit that calculates the stride length and the bias by distinguishing the local movement trajectory depending on whether it is a linear movement trajectory or a non-linear movement trajectory; and
A pedestrian indoor positioning device comprising a local initial position calculation unit that determines an initial local step position in a local coordinate system based on the calculated stride length and the bias. The method of claim 3, wherein the trajectory estimation unit
a local step selection unit that selects a step of a step position located within a predetermined threshold distance from the position of the corresponding AP among a plurality of local step positions of the local movement trajectory as a local step adjacent to the AP;
If the local movement trajectory is a straight movement trajectory, two different random steps are selected in various combinations from the local steps adjacent to the AP, and the distance difference from the corresponding AP in the selected step and the distance difference between the corresponding AP and the remaining steps are calculated. a straight line trajectory estimation unit that estimates the stride length and bias according to the straight line movement trajectory in the local coordinate system based on the straight line trajectory; and
If the local movement trajectory is a non-linear movement trajectory, set the initial local angle, which is the distance from the AP corresponding to the local coordinate system to the initial local step position and the angle from the first axis of the local coordinate system to the initial local step position direction, Two different random steps are selected in various combinations from the local steps adjacent to the AP, and based on the distance difference between the corresponding AP in the selected step and the corresponding AP in the remaining steps, a linear movement trajectory is followed in the local coordinate system. A pedestrian indoor positioning device including a non-linear trajectory estimation unit that estimates stride length and bias, and obtains the initial local angle by calculating the initial local angle based on the estimated stride length and bias. The method of claim 4, wherein the linear trajectory estimation unit
Local coordinates (q ₁ , u ₁ ) of the initial local step position (z ₁ ) and a total of N steps (n = 1, 2, .. N) proceeding in the direction of the q-axis, which is the first axis of the local coordinate system, in each of the local Depending on the step position (z ₁ to z _N ) and the distance (r _n ) between the corresponding AP, Equation

Determinant obtained by substituting two different random steps (a ₁ , a ₂ ) obtained from local steps adjacent to the AP into the stride (d) and bias (b) functions expressed as and analyzing the difference between the remaining steps

Equation

A pedestrian indoor positioning device that calculates the stride length (d) and the bias (b) according to . The method of claim 5, wherein the non-linear trajectory estimation unit
Distance from the origin of the local coordinate system to the initial local step position (z ₁ ) and the initial local angle (which is the angle from the q-axis direction to the initial local step position (z ₁ ) direction) ) and the local rotation angle (θ _n ) about the q axis according to the rotation at each local step (z ₂ to z _N ), Equation

Substituting two different random steps (a ₁ , a ₂ ) obtained from local steps adjacent to the AP into the stride (d) and bias (b) functions expressed as

(here ego, am.)
Determinant converted from

Equation

A pedestrian indoor positioning device that calculates the stride length (d) and the bias (b) according to . The method of claim 6, wherein the non-linear trajectory estimation unit
By varying the initial local angle (γ) in the equation for calculating the stride length (d) and the bias (b), the equation

here ( , and std() is the standard deviation function, am. and am.)
A pedestrian indoor positioning device that obtains an initial local angle (γ(S)) such that the error (e ₂ ) is minimized. The method of claim 7, wherein the trajectory estimation unit
The stride length (d), bias (b), and initial local angle (γ(S)) for each of two different random steps (a ₁ , a ₂ ) obtained in different combinations from the local steps adjacent to the AP are multiple. Once obtained, a pedestrian indoor location location device that selects each median value as a representative value. The method of claim 7, wherein the local initial position calculation unit
In the case of a straight movement trajectory, the coordinate values (q ₁ ^* , u ₁ ^* ) of the initial local step position (z ₁ ) are expressed in the equation

and

Obtained by calculating according to
In the case of a non-linear movement trajectory, the coordinate values (q ₁ ^* , u ₁ ^* ) of the initial local step position (z ₁ ) are expressed in the equation

Determinant converted from

Equation

A pedestrian indoor location positioning device that calculates and obtains according to. The method of claim 9, wherein the global location estimation unit
The rotation angle indicated by the initial movement direction of the pedestrian movement pattern based on the local step position (z _n ^(m)* ) in the local coordinate system for each of the plurality of APs (AP ₁ to AP _M ) based on the x-axis in the global coordinate system According to (w) the equation

A pedestrian indoor positioning device that converts to a step position (p _n ^(m) (w)) in the global coordinate system. The method of claim 10, wherein the global location estimation unit
If the step positions (p _n ^(m) (w)) on the global coordinate system converted from each of the multiple local coordinate systems do not match, it is determined whether it is a straight movement trajectory, and if it is a straight movement trajectory, different APs (AP) in the global coordinate system are determined. The distance between the positions (P _AP ^(r) _, P _AP ^(m) ) of r , AP m ) and the local step position ((z ₁ ^(r ) ₎ for each sign in different local coordinate systems according to AP (AP _r , AP _m ) ⁾ to z _N ^(r) ), (z ₁ ^(m) to z _N ^(m) )) local step position to ensure the same distance (z ₁ ^(m) to z _N ^(m) )) Select and set the corresponding rotation angle (w) to compensate for the error in the global step position,
If it is a non-linear movement trajectory, the rotation angle in each local coordinate system is set so that the difference between the local step positions (z ₁ ^(m) ~ z _N ^(m) ) according to each local movement trajectory converted to the global coordinate system is minimized. A pedestrian indoor positioning device that compensates for errors in the global step position by setting (w). In the pedestrian indoor positioning method of the pedestrian indoor positioning device that tracks the indoor position of the pedestrian,
Determining a pedestrian movement pattern based on a sensing signal that detects the pedestrian's steps and rotation angles;
Transmitting and receiving wireless signals with multiple APs and measuring the distance to each of the multiple APs in a predetermined manner at each step;
Set a plurality of local coordinate systems with the location of each of the plurality of APs as the origin, and based on the local movement trajectory obtained using the pedestrian movement pattern and the measured distance to each of the plurality of APs on the plurality of local coordinate systems. Calculating a bias generated by NLOS from the stride length and measured distance according to the step of the pedestrian, and determining an initial local step position of the local movement trajectory based on the calculated stride length and bias; and
A pedestrian indoor location positioning method comprising the step of converting a local movement trajectory obtained in each of a plurality of local coordinate systems into a movement trajectory in a global coordinate system to obtain a pedestrian movement trajectory in the global coordinate system. The method of claim 12, wherein the step of determining the pedestrian movement pattern is
Obtaining an acceleration signal by detecting the vertical acceleration according to the pedestrian's steps using an acceleration sensor;
Obtaining a gyro signal according to the pedestrian's rotation angle using a gyro sensor; and
A pedestrian indoor location positioning method comprising the step of receiving the acceleration signal and the gyro signal as the sensing signal and analyzing the pedestrian's steps and direction changes at each step to obtain the pedestrian movement pattern. The method of claim 13, wherein determining the initial local step position
A plurality of local coordinate systems with the location of each of the plurality of APs as the origin based on the distance to the plurality of APs, and the movement direction at the initial step position (p ₁ ) of the pedestrian movement pattern as the first axis of the coordinate system. Converting the pedestrian movement pattern into a local movement trajectory in a local coordinate system;
calculating the stride length and the bias by classifying the local movement trajectory according to whether it is a linear movement trajectory or a non-linear movement trajectory; and
A pedestrian indoor positioning method comprising determining an initial local step position in a local coordinate system based on the calculated step length and the bias. The method of claim 14, wherein calculating the bias includes
selecting a step of a step position located within a predetermined threshold distance from the position of the corresponding AP among a plurality of local step positions of the local movement trajectory as a local step adjacent to the AP;
If the local movement trajectory is a straight movement trajectory, two different random steps are selected in various combinations from the local steps adjacent to the AP, and the distance difference from the corresponding AP in the selected step and the distance difference between the corresponding AP and the remaining steps are calculated. estimating the stride length and bias based on a linear movement trajectory on a local coordinate system; and
If the local movement trajectory is a non-linear movement trajectory, set the initial local angle, which is the distance from the AP corresponding to the local coordinate system to the initial local step position and the angle from the first axis of the local coordinate system to the initial local step position direction, Two different random steps are selected in various combinations from the local steps adjacent to the AP, and based on the distance difference between the corresponding AP in the selected step and the corresponding AP in the remaining steps, a linear movement trajectory is followed in the local coordinate system. A pedestrian indoor positioning method comprising the steps of estimating stride length and bias, and calculating and obtaining the initial local angle based on the estimated stride length and bias. The method of claim 15, wherein the step of estimating according to the linear movement trajectory is
Local coordinates (q ₁ , u ₁ ) of the initial local step position (z ₁ ) and a total of N steps (n = 1, 2, .. N) proceeding in the direction of the q-axis, which is the first axis of the local coordinate system, in each of the local Depending on the step position (z ₁ to z _N ) and the distance (r _n ) between the corresponding AP, Equation

Determinant obtained by substituting two different random steps (a ₁ , a ₂ ) obtained from local steps adjacent to the AP into the stride (d) and bias (b) functions expressed as and analyzing the difference between the remaining steps

Equation

A pedestrian indoor positioning method for calculating the stride length (d) and the bias (b) by calculating according to . The method of claim 16, wherein calculating and obtaining the initial local angle
Distance from the origin of the local coordinate system to the initial local step position (z ₁ ) and the initial local angle (which is the angle from the q-axis direction to the initial local step position (z ₁ ) direction) ) and the local rotation angle (θ _n ) about the q axis according to the rotation at each local step (z ₂ to z _N ), Equation

Substituting two different random steps (a ₁ , a ₂ ) obtained from local steps adjacent to the AP into the stride (d) and bias (b) functions expressed as

(here ego, am.)
Determinant converted from

Equation

A pedestrian indoor positioning method comprising calculating the stride length (d) and the bias (b) by calculating according to . The method of claim 17, wherein calculating and obtaining the initial local angle
By varying the initial local angle (γ) in the equation for calculating the stride length (d) and the bias (b), the equation

here ( , and std() is the standard deviation function, am. and am.)
A pedestrian indoor positioning method further comprising obtaining an initial local angle (γ(S)) such that the error (e ₂ ) is minimized. 19. The method of claim 18, wherein determining the initial local step position
In the case of a straight movement trajectory, the coordinate values (q ₁ ^* , u ₁ ^* ) of the initial local step position (z ₁ ) are expressed in the equation

and

calculating and obtaining according to; and
In the case of a non-linear movement trajectory, the coordinate values (q ₁ ^* , u ₁ ^* ) of the initial local step position (z ₁ ) are expressed in the equation

Determinant converted from

Equation

A pedestrian indoor location location method including the step of calculating and obtaining according to. The method of claim 19, wherein the step of acquiring the pedestrian movement trajectory is
The rotation angle indicated by the initial movement direction of the pedestrian movement pattern based on the local step position (z _n ^(m)* ) in the local coordinate system for each of the plurality of APs (AP ₁ to AP _M ) based on the x-axis in the global coordinate system According to (w) the equation

A pedestrian indoor positioning method that converts to a step position (p _n ^(m) (w)) in the global coordinate system.