CN111829516A - Autonomous pedestrian positioning method based on smart phone - Google Patents
Autonomous pedestrian positioning method based on smart phone Download PDFInfo
- Publication number
- CN111829516A CN111829516A CN202010721022.2A CN202010721022A CN111829516A CN 111829516 A CN111829516 A CN 111829516A CN 202010721022 A CN202010721022 A CN 202010721022A CN 111829516 A CN111829516 A CN 111829516A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- angle
- positioning
- quaternion
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 239000002245 particle Substances 0.000 claims abstract description 49
- 238000005259 measurement Methods 0.000 claims abstract description 24
- 238000001914 filtration Methods 0.000 claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 20
- 238000004364 calculation method Methods 0.000 claims abstract description 18
- 238000001514 detection method Methods 0.000 claims abstract description 18
- 238000011478 gradient descent method Methods 0.000 claims abstract description 12
- 210000000689 upper leg Anatomy 0.000 claims abstract description 12
- 238000005070 sampling Methods 0.000 claims description 20
- 230000008569 process Effects 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 230000001133 acceleration Effects 0.000 claims description 13
- 230000005484 gravity Effects 0.000 claims description 12
- 230000004927 fusion Effects 0.000 claims description 6
- 239000000126 substance Substances 0.000 claims description 6
- 230000007704 transition Effects 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000006467 substitution reaction Methods 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 description 13
- 230000005021 gait Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012937 correction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000035515 penetration Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010348 incorporation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000005309 stochastic process Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
- G01C21/12—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
- G01C21/16—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
- G01C21/165—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation combined with non-inertial navigation instruments
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
Abstract
The invention relates to the technical field of human navigation and positioning, and provides an autonomous pedestrian positioning method based on a smart phone, which comprises the following steps: step 1, carrying out error calibration on information acquisition equipment built in a mobile phone; step 2, acquiring attitude data of the human body in an initial state, and performing attitude calculation and system initial alignment; step 3, collecting thigh swing angle signal data when the pedestrian normally walks, and carrying out pedestrian step detection; step 4, fusing data of two sensors, namely a gyroscope and an accelerometer and solving a quaternion by using the obtained pedestrian stride detection result and based on a course estimation algorithm of a gradient descent method; step 5, utilizing the following step length estimation model based on the accelerometer measurement value to estimate the step length of the pedestrian; and 6, performing map matching based on particle filtering by using the obtained step-span estimation result. The invention can improve the pedestrian positioning precision and optimize the positioning track, so that the track after filtering is more in line with the real situation.
Description
Technical Field
The invention relates to the technical field of pedestrian navigation and positioning, in particular to an autonomous pedestrian positioning method based on a smart phone.
Background
Location-based services are vital to people's lives, where services related to location information are visible everywhere. The most widely used Positioning System is the Global Positioning System (GPS) in the united states, and the GPS-based Positioning technology needs to rely on satellites for Positioning. GPS can provide reliable location information for pedestrians in open air environments, but there are signal blockage and multipath problems in tunnels, high-rise building complexes, and indoor environments. In recent years, technologies for positioning based on Wi-Fi, Bluetooth (Bluetooth), Radio Frequency Identification (RFID), Near Field Communication (NFC), Ultra Wide Band (UWB), 5G, visible light, and sound source have been diversified. Some technologies need to be able to obtain a better positioning effect in a relatively stable and unchangeable environment, once a scene of deploying infrastructure changes, signals captured in the positioning technology are affected, and the final positioning effect is greatly reduced. Although the positioning technologies can provide positioning services in a situation of GPS failure, the technologies all depend on deployment and support of infrastructure, and even special positioning terminals need to be equipped, so that the manufacturing cost and the technical limitation limit the popularization of the technologies in society.
The position information is acquired by depending on positioning equipment, and the smart phone has many advantages as a positioning carrier. The intelligent mobile phone is high in popularization rate in society as an intelligent mobile terminal, and has a plurality of users, and pedestrians do not need extra capital investment. In addition, various sensors such as an accelerometer, a gyroscope, a magnetometer and the like are built in the smart phone, and relevant sensor data can be collected for positioning. The smart phone has abundant computing resources, can process and calculate data, and can perform convenient and friendly human-computer interaction.
Currently, a common pedestrian location technology such as a Wireless Local Area Network (WLAN) technology, an Access Point (AP) of the WLAN technology can broadcast a beacon frame, and the beacon frame can be usually sent once every 100ms in a certain area, and includes a Media Access Control (MAC) address. The mobile node receives these signals and identifies the corresponding AP based on its MAC address. The user position can be estimated using the measured Signal Strength (RSS) of the AP. When the WLAN is used for positioning, the signal intensity of three or more surrounding APs is received, the distance between the receiving equipment and the surrounding APs is estimated according to the RSS measured value, and positioning is realized through a triangulation method. However, the problem of large error in converting RSS measurement value into distance is particularly serious in indoor environment, WLAN signals are easily reflected by walls, and are also shielded by indoor building structures or human bodies to generate signal attenuation, so that the positioning result is not accurate enough.
At present, more and more research focuses on fixing MEMS devices (such as accelerometers and gyroscopes) at the waist or the foot of a pedestrian, and performing multiple integration on data collected by a sensor to locate the pedestrian. However, due to various factors such as the measurement principle and inherent defects of the sensor, the error accumulation after multiple integrations is too large, the yaw is severe, and the normal indoor positioning requirement cannot be met.
Disclosure of Invention
The invention mainly solves the technical problems that the pedestrian positioning method in the prior art relies on external auxiliary equipment to realize positioning and the pedestrian positioning tracking track passes through the wall, and provides the autonomous pedestrian positioning method based on the smart phone. The method has the characteristics of high stability, strong independence, good anti-interference performance and the like.
The invention provides an autonomous pedestrian positioning method based on a smart phone, which comprises the following processes:
step 1, carrying out error calibration on information acquisition equipment built in a mobile phone;
step 2, acquiring attitude data of the human body in an initial state, and performing attitude calculation and system initial alignment;
step 3, collecting thigh swing angle signal data when the pedestrian normally walks, and carrying out pedestrian step detection;
step 4, fusing data of two sensors, namely a gyroscope and an accelerometer and solving a quaternion by using the obtained pedestrian stride detection result and based on a course estimation algorithm of a gradient descent method;
and 5, utilizing the following step length estimation model based on the accelerometer measurement value to estimate the step length of the pedestrian:
wherein M is1Represents the step length estimation result, k, corresponding to the model one1Representing the parameter, acc, corresponding to the stride modelmIn order to eliminate the resultant acceleration value of the local gravity, N represents the number of acceleration sampling points in a stride period;
step 6, performing map matching based on particle filtering by using the obtained step-span estimation result to realize updating of pedestrian positioning;
step 601, under a pedestrian track calculation system, obtaining a measurement equation of particle filtering by using the following particle filtering state transfer equation, further obtaining pedestrian course angles and step length particles which may appear in the next iteration of the particle filtering, and further realizing updating of pedestrian positioning:
in the state transition equation of particle filtering, the state quantity of the system is
Variation quantity delta psi of course anglekAnd stride length SLkFor system input, the measurement equation of the particle filter can be deduced to be obtained by a course estimation algorithm and a stride length estimation algorithm:
wherein the content of the first and second substances,
and
respectively noise of step length and noise of course angle increment;
step 602, performing particle through-wall judgment on the new particles obtained in step 601 by using a line segment intersection model, connecting the particles of two adjacent steps, and if the line segment is intersected with a line segment formed by a wall body and an intersection point is on the line segment, determining that the particles are invalid particles; and updating the particle weight to complete the updating of the pedestrian positioning.
Further, the information acquisition apparatus includes: accelerometers, gyroscopes, and magnetometers.
Further, step 2, collecting posture data of the human body in an initial state, and performing posture calculation and system initial alignment, wherein the method comprises the following processes:
step 201, acquiring posture data of a human body in an initial state;
step 202, determining values of a pitch angle and a roll angle in an Euler angle according to a projection relation of a gravity vector under a navigation system under a sensor coordinate system;
and 203, obtaining a complete initial Euler angle by using the obtained course angle according to the values of the pitch angle and the roll angle in the determined Euler angles, and further realizing the attitude calculation and the initial alignment of the system.
Further, step 3, collect thigh swing angle signal data when the pedestrian normally walks to carry out pedestrian step detection, including following process:
step 301, inputting an angle signal γnDetecting the auxiliary threshold lim corresponding to the thigh angle signal of the nth sampling point, the signal sampling frequency SRpAnd limv；
Step 302, if the angle signal satisfies γn>limp&&γn>γn-1&&γn+1<γnThen record as the first wavePeak value, otherwise if the condition gamma is satisfiedn<limv&&γn<γn-1&&γn+1>γnThen record as the first wave trough value, record the first detected wave peak value or wave trough value as (n)b,γb)，nbDenotes the corresponding sample point, γbRepresenting the corresponding angle signal magnitude; if the angle signal satisfies gamman>limp&&γn>γn-1&&γn+1<γnIf the distance between the sampling point meeting the condition and the sampling point corresponding to the previous wave peak value is greater than 0.3. SR, recording the point as the next wave peak value, otherwise, if the condition gamma is metn<limv&&γn<γn-1&&γn+1>γnRecording as the next wave valley value;
step 303, repeating steps 301 to 302 until all wave peak values and wave trough values are detected and recorded;
step 304, if the sampling point of the peak value is larger than the last valley value, the last peak value is recorded as (n)e,γe)，neRepresents the corresponding sampling point, γeRepresenting the magnitude of the corresponding angle signal, otherwise the last valley value is recorded as (n)e,γe)。
Further, step 4, performing data fusion of two sensors, namely a gyroscope and an accelerometer and solving a quaternion by using the obtained pedestrian stride detection result and a course estimation algorithm based on a gradient descent method, and comprising the following processes:
the objective function in the quaternion estimation process can be expressed as follows:
wherein f (-) is expressed as an error function, UNAnd USRespectively representing vectors under a navigation system and a mobile phone coordinate system, wherein q represents a quaternion, and q represents a conjugate quaternion;
solving the optimal attitude quaternion by the following formula:
wherein the content of the first and second substances,
denotes f (q)k,UN,US) At qkThe gradient is positioned, the opposite direction of the position is the direction in which the target function descends most quickly, and lambda represents the size of the step;
the accelerometer reading needs normalization processing, and the accelerometer measurement value in the sensor coordinate system is normalized
Then there should be:
the corresponding jacobian matrix can be found:
the gradient formula of the final objective function can be obtained as follows:
substitution of the above formula
And obtaining the optimal quaternion after multiple iterations, and further finishing course angle estimation.
The autonomous pedestrian positioning method provided by the invention can realize pedestrian positioning with high stability, strong independence and good anti-interference performance under the conditions of not depending on infrastructure and other external auxiliary equipment by utilizing the built-in sensor of the smart phone. A stride detection algorithm based on a thigh swing angle is realized from a gait analysis angle, the detection result is accurate, and the error rate is low; a simplified step size model is provided, and calculation is easy; the map matching of the pedestrian track is realized based on the particle filtering, the track wall-through phenomenon is corrected, and the track is more in line with the real situation.
Drawings
Fig. 1 is a flowchart of an implementation of the autonomous pedestrian positioning method based on the smart phone provided in the present invention.
Detailed Description
In order to make the technical problems solved, technical solutions adopted and technical effects achieved by the present invention clearer, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some but not all of the relevant aspects of the present invention are shown in the drawings.
Fig. 1 is a flowchart of an implementation of an autonomous pedestrian positioning method based on a smart phone according to an embodiment of the present invention. As shown in fig. 1, the autonomous pedestrian positioning method based on a smart phone provided in the embodiment of the present invention includes:
step 1, carrying out error calibration on information acquisition equipment built in the mobile phone.
The information acquisition apparatus includes: the system comprises an accelerometer, a gyroscope and a magnetometer, wherein the accelerometer is used for acquiring specific force acceleration data; collecting angular velocity data by using a gyroscope; magnetic field strength data is collected using a magnetometer. Because the embodiment needs to collect the intelligent information, the mobile phone referred to in the embodiment is a smart phone.
In this embodiment, the accelerometer error calibration adopts a six-position calibration method, which does not require special measurement auxiliary equipment, and only needs to turn the accelerometer over at six different positions to acquire local gravitational acceleration. The collected specific acceleration of the three sensitive axes is 1g or-1 g respectively, and the standard value of the local acceleration of gravity is used as a reference to identify and calibrate errors. During the measurement, the acceleration is continuously acquired for a period of time and averaged to obtain the measurement value at rest.
The gyroscope error calibration adopts an Allan variance method, which is a method for describing a random process. The Allan variance can provide direct information of the underlying stochastic process and achieves a better development momentum due to its computational simplicity and easy adaptation to various noise types. Although there is a certain limitation in mapping from the Allan variance to the frequency spectrum, the method is still powerful and has become an IEEE standard specification and test method for inertial sensors. The method first separates statically measured gyroscope data into subsets and then calculates the statistical variance between the equal dimensional subsets.
The error calibration of the magnetometer adopts a splayed ellipsoid fitting method. In the absence of measurement errors and interfering magnetic fields, the projected trajectory of the magnetometer measurements should be a sphere with a radius equal to the local magnetic field strength. Based on the theoretical premise, the mobile phone can be rotated for a plurality of times according to the splayed pattern, the mobile phone needs to be rotated to the starting point when the rotation action is finished every time, and then the rotation times are increased. The magnetic field intensity obtained while rotating is transmitted to a computer by wireless and then the data is processed. The more the rotation times are, the more the model fitted to the magnetic field intensity points acquired by the mobile phone is similar to the ellipsoid model.
And 2, acquiring attitude data of the human body in an initial state, and performing attitude calculation and system initial alignment.
Step 201, posture data of a human body in an initial state is collected.
Because the position of the mobile phone in the pocket is arbitrary, the coordinate axis of the mobile phone is not coincident with the coordinate axis in the navigation system in the initial state, and the course of the mobile phone is not consistent with the actual moving direction of the pedestrian. Aiming at the problem, the system needs to be initially aligned, and meanwhile, the influence on the system caused by different placement positions of the mobile phone can be avoided. For initial alignment, data is acquired that the pedestrian is stationary for a predetermined period of time, for example set to 2-3 seconds.
In the embodiment, specific force acceleration data in an initial state of a human body is collected by an accelerometer; acquiring angular velocity data in an initial state by using a gyroscope; and collecting magnetic field intensity data of an initial state by using a magnetometer.
And 202, determining values of a pitch angle and a roll angle in an Euler angle according to the projection relation of the gravity vector under the navigation system under the sensor coordinate system.
Because the pedestrian is in the initial state (in the static state), the data collected by the accelerometer in the mobile phone is the reading of the gravity vector in the sensor coordinate system. The values of the pitch angle and the roll angle in the Euler angle can be determined according to the projection relation of the gravity vector under the navigation system under the sensor coordinate system. In the process of standing still, a person cannot be completely ensured to be in a still state, so that continuous data in the still state need to be averaged, and the accuracy of initial posture calculation can be improved.
And 203, obtaining a complete initial Euler angle by using the obtained course angle according to the values of the pitch angle and the roll angle in the determined Euler angles, and further realizing the attitude calculation and the initial alignment of the system.
And calculating the pitch angle and the roll angle through the mean value data obtained by final calculation. The complete initial euler angle cannot be solved using accelerometers, so a heading angle needs to be calculated using a magnetometer. When the heading angle is calculated, the geomagnetic vector in the sensor coordinate system needs to be converted from the mobile phone coordinate system to the horizontal plane in the navigation system. After the quaternion corresponding to the spatial attitude rotation is obtained, attitude calculation is realized; after the pitch angle, the roll angle and the initial course angle are obtained, the initial alignment of the system is realized.
And 3, collecting thigh swing angle signal data when the pedestrian normally walks, and carrying out pedestrian stride detection.
The variation of the thigh swing angle has a periodic and stable character during the gait cycle. The stride detection can be performed according to the change of the thigh swing angle. In a complete stride, the angle signal corresponds to four key event points of wave peak value, descending zero crossing, wave valley value and ascending zero crossing. After the wave peak value is reached again from the rising zero crossing, the gait enters the next period. When these four gait events are detected, a complete stride is deemed to have occurred and is recorded as a stride. In practice, the start and end of a pedestrian's walk do not meet a full stride. However, if the four states are simply discarded without being satisfied, and not recorded as a stride, the accuracy of the positioning is affected. And combining the analysis result with the actual situation to carry out the following pedestrian stride detection:
step 301, inputting an angle signal γnDetecting the auxiliary threshold lim corresponding to the thigh angle signal of the nth sampling point, the signal sampling frequency SRpAnd limv。
Step 302, if the angle signal satisfies γn>limp&&γn>γn-1&&γn+1<γnRecording as the first wave peak value, otherwise if the condition gamma is satisfiedn<limv&&γn<γn-1&&γn+1>γnThen record as the first wave trough value, record the first detected wave peak value or wave trough value as (n)b,γb)，nbDenotes the corresponding sample point, γbRepresenting the corresponding angle signal magnitude; if the angle signal satisfies gamman>limp&&γn>γn-1&&γn+1<γnIf the distance between the sampling point meeting the condition and the sampling point corresponding to the previous wave peak value is greater than 0.3. SR, recording the point as the next wave peak value, otherwise, if the condition gamma is metn<limv&&γn<γn-1&&γn+1>γnThe next valley value is recorded.
And 303, repeatedly executing the steps 301 to 302 until all the wave peak values and the wave valley values are detected and recorded.
Step 304, if the sampling point of the peak value is larger than the last valley value, the last peak value is recorded as (n)e,γe)，neRepresents the corresponding sampling point, γeRepresenting the magnitude of the corresponding angle signal, otherwise the last valley value is recorded as (n)e,γe)。
And 4, fusing data of two sensors, namely a gyroscope and an accelerometer and solving a quaternion by using the obtained pedestrian stride detection result and based on a course estimation algorithm of a gradient descent method. The objective function in the quaternion estimation process can be expressed as follows:
wherein f (-) is expressed as an error function, UNAnd USAnd the vectors in a navigation system and a mobile phone coordinate system are respectively expressed, q represents a quaternion, and q represents a conjugate quaternion.
When the quaternion has no error, the vector under the navigation system can obtain the same vector as the vector under the sensor coordinate system through the rotation of the attitude. The error function is a multivariate function taking quaternion as a dependent variable, the optimization aims to make the error function approach to 0, and the solution q meeting the objective function is obtained at the moment, namely the quaternion of the optimal attitude is solved through the following formula:
wherein the content of the first and second substances,
denotes f (q)k,UN,US) At qkThe gradient, where the opposite direction is the direction in which the objective function falls the fastest, and λ represents the size of the step.
When data fusion of two sensors of a gyroscope and an accelerometer is carried out, the earth gravity vector under a navigation system is normalized to be gN(0,0,0, -1). The accelerometer reading needs normalization processing, and the accelerometer measurement value in the sensor coordinate system is normalized
Then there should be:
the corresponding jacobian matrix can be found:
the gradient formula of the final objective function can be obtained as follows:
substitution of the above formula
And obtaining the optimal quaternion after multiple iterations, and further finishing course angle estimation.
In pedestrian heading estimation algorithms, it is common to base the assumption of parallel heading, i.e., there is a relatively stationary relationship between the direction the pedestrian is heading and the heading of the sensor. The angle between the mobile phone and the real heading is already corrected through the initial alignment of the system. When the mobile phone is placed in a pocket, even if a pedestrian is in a motion state, the pedestrian still has a stable relation with a human body, and does not move too violently. The course estimation algorithm based on the gradient descent method can realize the course estimation of the embodiment.
Due to various factors such as the measurement principle and the inherent defects of the sensor, the single sensor cannot meet the attitude calculation requirement of the system. By means of a sensor data fusion algorithm based on a gradient descent method, measurement data of a plurality of sensors can be fused to obtain reliable attitude information. The sensor data fusion algorithm based on the gradient descent method is essentially a process of optimizing a solved quaternion. The method is based on a gradient descent method for fusing sensor data and solving quaternion, and the idea is to optimize and estimate the solved quaternion by constructing a reasonable optimization target. Because of the incorporation of accelerometer and magnetometer data, in the quaternion estimation process, the magnetometer only measures the earth's magnetic field, assuming that only local gravity is measured by the accelerometer.
When the gradient descent method is used, an initial quaternion is required to be input, and the initial quaternion is obtained in the initial alignment stage of the system. The gravity vector and the magnetic field vector need to be normalized when quaternion updating is carried out, so that the influence of dimension can be eliminated.
And 5, estimating the pedestrian step length based on the step length estimation model of the accelerometer measurement value.
Logically, the pedestrian stride length estimation model should be influenced by some personalized parameters, such as the height, weight, sex, age, etc. of the pedestrian. In addition to this, errors in the model itself need to be taken into account. To some extent, too many factors are considered to make the model too complex to build. For the stride model to be simple and well-functioning, a simple linear model is often used, which ignores these personalization parameters. The best method is to train a cross-step model for a specific person to achieve a high-precision cross-step estimation.
The embodiment performs pedestrian step estimation by using the following step estimation model based on accelerometer measurement:
wherein M is1Represents the step length estimation result, k, corresponding to the model one1The parameter corresponding to the stride model is shown, and the reference value given in this embodiment is 0.98, accmTo reject the resultant acceleration value of local gravity. N represents the number of acceleration samples in a stride period. The step-span model uses data collected by the acceleration sensor as input of step-span estimation, and avoids influence on the step-span model due to personal difference.
And 6, performing map matching based on particle filtering by using the obtained step-span estimation result, and realizing updating of pedestrian positioning.
As a technology of relative positioning, the pedestrian dead reckoning has the problem that errors are accumulated continuously during walking. In an outdoor environment, the requirement on positioning accuracy is not high, and in a daily mobile phone positioning and navigation service, a positioning error of several meters cannot bring about too large problem on use experience; in an indoor environment, the building structure is compact, positioning targets are dense, and small positioning errors bring poor positioning experience. In particular, when the positioning result is drawn on a map, the problem of the track passing through the wall occurs. The invention improves the positioning result through particle filtering and improves the positioning precision.
Step 601, under a pedestrian track calculation system, obtaining a measurement equation of particle filtering by using the following particle filtering state transfer equation, and further obtaining pedestrian course angles and step length particles which may appear in the next iteration of the particle filtering. And further updating the pedestrian positioning:
in the state transition equation of particle filtering, the state quantity of the system is
Variation quantity delta psi of course anglekAnd stride length SLkFor system input, the measurement equation of the particle filter can be deduced to be obtained by a course estimation algorithm and a stride length estimation algorithm:
wherein the content of the first and second substances,
and
respectively, the step-length-striding noise and the course angle increment noise.
Step 602, performing particle through-wall judgment on the new particle obtained in step 601 by using a line segment intersection model, connecting the particles of two adjacent steps, and if the line segment intersects with the line segment formed by the wall body and the intersection point is on the line segment, determining that the particle is an invalid particle. And updating the particle weight to complete the updating of the pedestrian positioning.
If the indoor positioning technology is applied to practice, map information is indispensable basic information. Otherwise, the positioning result can only be abstracted into points in the coordinate system, and cannot be converted into geographic information with practical significance. In an indoor environment, the results of positioning are associated with specific locations in specific rooms or hallways to render the technical value of indoor positioning. The invention considers the correction of the pedestrian track and the improvement of the positioning precision by using an indoor map matching technology. The map matching technique, also referred to as a map assist technique in some research, is a pseudo measurement technique that enables positioning correction based on a software technique.
According to practical experience, pedestrians cannot pass through a wall or reach inaccessible areas during walking. For the particles which reach the unreachable region and are considered as invalid particles, the weight of the particles should be set to be zero, so that the problem of wall penetration of the pedestrian track can be avoided through a particle filter algorithm. From the above analysis, the corresponding particle weight update is performed using the following formula:
the formula limits and restricts the walking of the pedestrian through map information, so that the updating of the state quantity is completed, and the updating of the pedestrian positioning is completed. After map matching and particle filtering, the problem of pedestrian track wall penetration is solved, and the unrealistic yaw is corrected, so that the pedestrian track is more in accordance with the real situation.
The embodiment comprises the steps of analyzing gait characteristics of pedestrians in the walking process, extracting key gait events as the step detection conditions, and improving the performance of a step detection algorithm. A sensor data fusion algorithm based on a gradient descent method is improved, magnetometer data are restrictively fused, heading drift suppression is achieved, and filtering divergence caused by magnetic field interference is avoided. And performing parameter estimation of the step-by-step model by using a maximum likelihood estimation method. Through the initial alignment of the system, the problem that the course of the mobile phone is inconsistent with that of the pedestrian is solved, and meanwhile, the error influence on the system positioning caused by different positions of the smart phone placed in the pocket is avoided. The positioning accuracy is improved by constructing a state transition equation and a measurement equation of particle filtering based on pedestrian track calculation.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: modifications of the technical solutions described in the embodiments or equivalent replacements of some or all technical features may be made without departing from the scope of the technical solutions of the embodiments of the present invention.
Claims (5)
1. An autonomous pedestrian positioning method based on a smart phone is characterized by comprising the following processes:
step 1, carrying out error calibration on information acquisition equipment built in a mobile phone;
step 2, acquiring attitude data of the human body in an initial state, and performing attitude calculation and system initial alignment;
step 3, collecting thigh swing angle signal data when the pedestrian normally walks, and carrying out pedestrian step detection;
step 4, fusing data of two sensors, namely a gyroscope and an accelerometer and solving a quaternion by using the obtained pedestrian stride detection result and based on a course estimation algorithm of a gradient descent method;
and 5, utilizing the following step length estimation model based on the accelerometer measurement value to estimate the step length of the pedestrian:
wherein M is1Represents the step length estimation result, k, corresponding to the model one1Representing the parameter, acc, corresponding to the stride modelmIn order to eliminate the resultant acceleration value of the local gravity, N represents the number of acceleration sampling points in a stride period;
step 6, performing map matching based on particle filtering by using the obtained step-span estimation result to realize updating of pedestrian positioning;
step 601, under a pedestrian track calculation system, obtaining a measurement equation of particle filtering by using the following particle filtering state transfer equation, further obtaining pedestrian course angles and step length particles which may appear in the next iteration of the particle filtering, and further realizing updating of pedestrian positioning:
in the state transition equation of particle filtering, the state quantity of the system is
Variation quantity delta psi of course anglekAnd stride length SLkFor system input, the measurement equation of the particle filter can be deduced to be obtained by a course estimation algorithm and a stride length estimation algorithm:
wherein the content of the first and second substances,
and
respectively noise of step length and noise of course angle increment;
step 602, performing particle through-wall judgment on the new particles obtained in step 601 by using a line segment intersection model, connecting the particles of two adjacent steps, and if the line segment is intersected with a line segment formed by a wall body and an intersection point is on the line segment, determining that the particles are invalid particles; and updating the particle weight to complete the updating of the pedestrian positioning.
2. The smartphone-based autonomous pedestrian positioning method of claim 1, wherein the information-collecting device comprises: accelerometers, gyroscopes, and magnetometers.
3. The smartphone-based autonomous pedestrian positioning method according to claim 1, wherein the step 2 of collecting posture data of the human body in an initial state and performing posture calculation and system initial alignment comprises the following processes:
step 201, acquiring posture data of a human body in an initial state;
step 202, determining values of a pitch angle and a roll angle in an Euler angle according to a projection relation of a gravity vector under a navigation system under a sensor coordinate system;
and 203, obtaining a complete initial Euler angle by using the obtained course angle according to the values of the pitch angle and the roll angle in the determined Euler angles, and further realizing the attitude calculation and the initial alignment of the system.
4. The smartphone-based autonomous pedestrian positioning method according to claim 1, wherein step 3 is to collect thigh swing angle signal data of a pedestrian walking normally and perform pedestrian stride detection, and comprises the following steps:
step 301, inputting an angle signal γnDetecting the auxiliary threshold lim corresponding to the thigh angle signal of the nth sampling point, the signal sampling frequency SRpAnd limv；
Step 302, if the angle signal satisfies γn>limp&&γn>γn-1&&γn+1<γnThen record as the first wave peak valueOtherwise, if the condition γ is satisfiedn<limv&&γn<γn-1&&γn+1>γnThen record as the first wave trough value, record the first detected wave peak value or wave trough value as (n)b,γb)，nbDenotes the corresponding sample point, γbRepresenting the corresponding angle signal magnitude; if the angle signal satisfies gamman>limp&&γn>γn-1&&γn+1<γnIf the distance between the sampling point meeting the condition and the sampling point corresponding to the previous wave peak value is greater than 0.3. SR, recording the point as the next wave peak value, otherwise, if the condition gamma is metn<limv&&γn<γn-1&&γn+1>γnRecording as the next wave valley value;
step 303, repeating steps 301 to 302 until all wave peak values and wave trough values are detected and recorded;
step 304, if the sampling point of the peak value is larger than the last valley value, the last peak value is recorded as (n)e,γe)，neRepresents the corresponding sampling point, γeRepresenting the magnitude of the corresponding angle signal, otherwise the last valley value is recorded as (n)e,γe)。
5. The smartphone-based autonomous pedestrian positioning method according to claim 1, wherein in step 4, the obtained pedestrian stride detection result is utilized, a gradient descent method-based course estimation algorithm is used to perform data fusion of two sensors, namely a gyroscope and an accelerometer, and solve quaternion, and the method comprises the following steps:
the objective function in the quaternion estimation process can be expressed as follows:
f(q,UN,US)＝q*oUNoq-US→0 (1)
wherein f (-) is expressed as an error function, UNAnd USRespectively representing vectors under a navigation system and a mobile phone coordinate system, and q represents fourThe element number, q, represents the conjugate quaternion;
solving the optimal attitude quaternion by the following formula:
wherein the content of the first and second substances,
denotes f (q)k,UN,US) At qkThe gradient is positioned, the opposite direction of the position is the direction in which the target function descends most quickly, and lambda represents the size of the step;
the accelerometer reading needs normalization processing, and the accelerometer measurement value in the sensor coordinate system is normalized
Then there should be:
the corresponding jacobian matrix can be found:
the gradient formula of the final objective function can be obtained as follows:
substitution of the above formula
And obtaining the optimal quaternion after multiple iterations, and further finishing course angle estimation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010721022.2A CN111829516B (en) | 2020-07-24 | 2020-07-24 | Autonomous pedestrian positioning method based on smart phone |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010721022.2A CN111829516B (en) | 2020-07-24 | 2020-07-24 | Autonomous pedestrian positioning method based on smart phone |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111829516A true CN111829516A (en) | 2020-10-27 |
| CN111829516B CN111829516B (en) | 2024-04-05 |
Family
ID=72926236
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010721022.2A Active CN111829516B (en) | 2020-07-24 | 2020-07-24 | Autonomous pedestrian positioning method based on smart phone |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111829516B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112304317A (en) * | 2020-10-28 | 2021-02-02 | 中国矿业大学 | Indoor positioning method based on indoor multidimensional geomagnetic features |
| CN112711054A (en) * | 2021-03-24 | 2021-04-27 | 腾讯科技（深圳）有限公司 | Object behavior detection method and device, electronic equipment and storage medium |
| CN112904884A (en) * | 2021-01-28 | 2021-06-04 | 歌尔股份有限公司 | Method and device for tracking trajectory of foot type robot and readable storage medium |
| CN113203416A (en) * | 2021-03-19 | 2021-08-03 | 电子科技大学 | Pedestrian dead reckoning method for swing arm pedestrian |
| CN113790735A (en) * | 2021-08-20 | 2021-12-14 | 北京自动化控制设备研究所 | Pedestrian single-step division method in complex motion state |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103968827A (en) * | 2014-04-09 | 2014-08-06 | 北京信息科技大学 | Wearable human body gait detection self-localization method |
| CN104931049A (en) * | 2015-06-05 | 2015-09-23 | 北京信息科技大学 | Movement classification-based pedestrian self-positioning method |
| CN108957510A (en) * | 2018-07-25 | 2018-12-07 | 南京航空航天大学 | Based on inertia/zero-speed/GPS pedestrian is seamless combined navigation locating method |
| WO2020123988A1 (en) * | 2018-12-13 | 2020-06-18 | Solitonreach, Inc. | System and method for motion based alignment of body parts |
-
2020
- 2020-07-24 CN CN202010721022.2A patent/CN111829516B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103968827A (en) * | 2014-04-09 | 2014-08-06 | 北京信息科技大学 | Wearable human body gait detection self-localization method |
| CN104931049A (en) * | 2015-06-05 | 2015-09-23 | 北京信息科技大学 | Movement classification-based pedestrian self-positioning method |
| CN108957510A (en) * | 2018-07-25 | 2018-12-07 | 南京航空航天大学 | Based on inertia/zero-speed/GPS pedestrian is seamless combined navigation locating method |
| WO2020123988A1 (en) * | 2018-12-13 | 2020-06-18 | Solitonreach, Inc. | System and method for motion based alignment of body parts |
Non-Patent Citations (2)
| Title |
|---|
| 吴金凯;蔡成林;甘才军;孙凯;: "基于手机传感器的行人室内定位算法", 桂林电子科技大学学报, no. 05 * |
| 杨增瑞;段其昌;毛明轩;段盼;黄晓刚;: "基于磁场指纹辅助的手机室内定位系统", 传感技术学报, vol. 29, no. 09 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112304317A (en) * | 2020-10-28 | 2021-02-02 | 中国矿业大学 | Indoor positioning method based on indoor multidimensional geomagnetic features |
| CN112904884A (en) * | 2021-01-28 | 2021-06-04 | 歌尔股份有限公司 | Method and device for tracking trajectory of foot type robot and readable storage medium |
| CN113203416A (en) * | 2021-03-19 | 2021-08-03 | 电子科技大学 | Pedestrian dead reckoning method for swing arm pedestrian |
| CN112711054A (en) * | 2021-03-24 | 2021-04-27 | 腾讯科技（深圳）有限公司 | Object behavior detection method and device, electronic equipment and storage medium |
| CN113790735A (en) * | 2021-08-20 | 2021-12-14 | 北京自动化控制设备研究所 | Pedestrian single-step division method in complex motion state |
| CN113790735B (en) * | 2021-08-20 | 2023-09-12 | 北京自动化控制设备研究所 | Pedestrian single-step dividing method under complex motion state |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111829516B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111829516B (en) | Autonomous pedestrian positioning method based on smart phone | |
| Ban et al. | Indoor positioning method integrating pedestrian Dead Reckoning with magnetic field and WiFi fingerprints | |
| CN110501011B (en) | Determining a location of a mobile device in a geographic area | |
| CA2843747C (en) | Moving direction determination with noisy signals from inertial navigation systems on mobile devices | |
| EP2946167B1 (en) | Method and apparatus for determination of misalignment between device and pedestrian | |
| Zhang et al. | A localization database establishment method based on crowdsourcing inertial sensor data and quality assessment criteria | |
| Seitz et al. | A hidden markov model for urban navigation based on fingerprinting and pedestrian dead reckoning | |
| US9706362B2 (en) | Information determination in a portable electronic device carried by a user | |
| CN106908060A (en) | A kind of high accuracy indoor orientation method based on MEMS inertial sensor | |
| Shi et al. | A robust pedestrian dead reckoning system using low-cost magnetic and inertial sensors | |
| Höflinger et al. | Indoor-localization system using a micro-inertial measurement unit (imu) | |
| Hao et al. | A multi-floor location method based on multi-sensor and WiFi fingerprint fusion | |
| CN113566820B (en) | Fused pedestrian positioning method based on position fingerprint and PDR algorithm | |
| Zhou | Pedestrian dead reckoning on smartphones with varying walking speed | |
| Park et al. | Robust pedestrian dead reckoning for multiple poses in smartphones | |
| KR20150129285A (en) | Information determination in a portable electronic device carried by a user | |
| CN110285809A (en) | A kind of integrated integrated positioning device of indoor and outdoor | |
| Kong et al. | Hybrid indoor positioning method of BLE and PDR based on adaptive feedback EKF with low BLE deployment density | |
| Nguyen et al. | User-friendly heading estimation for arbitrary smartphone orientations | |
| Basso et al. | A smartphone-based indoor localization system for visually impaired people | |
| US10914793B2 (en) | Method and system for magnetometer calibration | |
| Asano et al. | A robust pedestrian dead-reckoning positioning based on pedestrian behavior and sensor validity | |
| CN115406435A (en) | Indoor electronic map construction method and device based on WLAN and MEMS and storage medium | |
| Liu et al. | A calibrated-RSSI/PDR/Map intergrated system based on a novel particle filter for indoor navigation | |
| KR20160004084A (en) | Pedestrian Tracking apparatus and method at indoors |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |