CN116092193A - Pedestrian track reckoning method based on human motion state identification - Google Patents
Pedestrian track reckoning method based on human motion state identification Download PDFInfo
- Publication number
- CN116092193A CN116092193A CN202310113241.6A CN202310113241A CN116092193A CN 116092193 A CN116092193 A CN 116092193A CN 202310113241 A CN202310113241 A CN 202310113241A CN 116092193 A CN116092193 A CN 116092193A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- acceleration
- motion
- motion state
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000001133 acceleration Effects 0.000 claims abstract description 53
- 238000001514 detection method Methods 0.000 claims abstract description 34
- 230000003044 adaptive effect Effects 0.000 claims abstract description 11
- 230000008569 process Effects 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 6
- 230000000284 resting effect Effects 0.000 claims description 2
- 238000005070 sampling Methods 0.000 claims description 2
- 238000005259 measurement Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 description 5
- 230000005021 gait Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000009825 accumulation Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000036544 posture Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
Abstract
The invention discloses a pedestrian dead reckoning method based on pedestrian motion state identification. Aiming at the problem that the traditional pedestrian dead reckoning (Pedestrian Dead Reckoning, PDR) algorithm can only be used in a single state of normal walking and is difficult to meet the actual application requirements, an improved PDR algorithm based on human motion state identification is provided. The algorithm divides the pedestrian movement mode into two states of walking and running, fully considers the relation between the acceleration peak value and the movement state in the pedestrian movement process, and obtains the acceleration peak value under different movement states through experiments, thereby setting a dynamic threshold value and realizing step counting detection and step length estimation under different states. Application of the modified PDR algorithm to pedestrian positioning: the method comprises the steps of utilizing motion data of pedestrians acquired by a micro inertial measurement unit (Inertial Measurement Unit, IMU) during walking, utilizing an improved peak detection method to perform step counting detection and state identification on the pedestrians, estimating step sizes according to the motion states of the pedestrians by adopting an adaptive step size estimation formula, and finally combining calculated heading to obtain the position information of the pedestrians.
Description
Technical Field
The invention belongs to the technical field of inertial navigation, and relates to a pedestrian dead reckoning method based on human motion state identification.
Background
The principle of Pedestrian Dead Reckoning (PDR) is to calculate real-time travel step length and travel direction of a pedestrian through sensor data of an accelerometer, a gyroscope and a magnetometer of an IMU, and calculate the current position in an accumulation mode, so that positioning is realized. With the rapid development of Micro-Electro-Mechanical System (MEMS), IMUs with low cost, small size, low energy consumption and easy integration are increasingly paid attention to, and pedestrian navigation based on IMUs is becoming a research hotspot. The IMU-based positioning system does not need to acquire information of an unknown environment in advance, does not need to deploy other hardware equipment in a positioning area, and has important application value in the field of emergency disaster relief of fire, earthquake and the like.
At present, most of researches mainly focus on reducing the measurement error of an IMU (inertial measurement unit) on the problem of improving the precision of pedestrian positioning navigation, and few researches on the motion state of pedestrians are performed. The principle of the PDR algorithm is that on the premise that the initial position of a pedestrian is known, whether the pedestrian strides over one step is judged through data of an accelerometer and a gyroscope obtained in real time, and then the step length and the course angle of the pedestrian are calculated, so that the current position information of the pedestrian is obtained through accumulation. The conventional PDR algorithm has an unsatisfactory positioning effect in the motion states of pedestrian running, walking, running mixing and the like. Based on the existing research, the method provides a pedestrian dead reckoning method based on human motion state identification. By analyzing corresponding acceleration peaks in the walking state and the running state of the pedestrian, different acceleration thresholds and time windows are set to improve the original peak detection algorithm to perform step counting detection, and meanwhile, for different motion states, a self-adaptive step calculation formula is adopted to reduce the overall positioning error of the PDR algorithm. Experimental results show that the algorithm achieves good positioning effect in the walking state, running state and walking and running state of pedestrians.
CN113239803a, a dead reckoning positioning method based on pedestrian motion state recognition, comprising the steps of: constructing a pedestrian motion state identification and classification model, identifying the pedestrian motion state, and performing step frequency detection, step length estimation, course estimation and dead reckoning. The invention has the beneficial effects that: aiming at the motion state of pedestrians in a two-dimensional space, five types of motion states are collected, and only three-axis acceleration and three-axis gyroscope data are collected to realize pedestrian navigation and positioning, so that the practicability of the dead reckoning method is stronger, and the application and development of the practical application environment are facilitated. The five types of motion state recognition models are used for modeling, and the motion states are recognized more perfectly and accurately by using the models. Gait detection and step frequency detection do not need to change the detection method according to the motion state behaviors, are more universal and accurate, and reduce the coupling with the human motion state identification method to the greatest extent. The dead reckoning can accept more complex motion state types, and has stronger applicability.
The method in the patent needs to obtain enough gait data for training the gait model, also needs to process time domain features of five types of motion states and establish an effective feature matrix, has higher algorithm complexity and higher requirement on the processing capacity of hardware, and is difficult to popularize in practical application; the motion state is classified too much, and erroneous judgment is easy to occur, so that the error becomes large. The algorithm does not need to acquire gait data in advance, the classification of the motion state is based on an acceleration threshold value and a time window threshold value, the time complexity and the space complexity are low, and the hardware cost is small.
Disclosure of Invention
The present invention is directed to solving the above problems of the prior art. A pedestrian dead reckoning method based on human motion state identification is provided. The technical scheme of the invention is as follows:
a pedestrian dead reckoning method based on human motion state identification comprises the following steps:
step a, wearing an IMU on the waist of a pedestrian, acquiring data of an accelerometer and a gyroscope in real time in the motion process of the pedestrian, and preprocessing the data;
step b, classifying and step counting detection are carried out on the motion state of the pedestrian by using the self-adaptive peak detection algorithm by utilizing the collected motion data, and the course is calculated by using a course calculation model;
step c, estimating the actual motion step length of the fitted pedestrian by using the self-adaptive step length according to different motion states of the pedestrian;
and d, acquiring the position information of the pedestrians in the moving process in real time according to step counting detection and step length estimation.
Further, the preprocessing the data in the step a specifically includes:
combining the collected triaxial acceleration data to obtain a combined acceleration module value according to the following steps:
in the formula ,ax 、a y 、a z The three-axis acceleration value is acquired;
preprocessing the original data, eliminating various noises, and keeping the characteristics of the original data as much as possible; preprocessing the combined acceleration data by adopting a moving average filtering method, wherein the window size is 5.
Further, the step b uses the collected motion data to classify the motion state of the pedestrian and detect the step by using a self-adaptive peak detection algorithm, and uses a course calculation model to calculate the course, which specifically comprises the following steps:
the peak detection method is characterized in that the characteristic similar to a sine wave is presented by using an acceleration signal when a human body walks, when a wave crest is detected, the pedestrian is considered to walk one step, and the moment of the wave crest is recorded, wherein the moment is shown as the following formula:
where S represents the set of moments labeled acceleration peaks,
and />
Respectively represent t peak-k and tpeak+k The magnitude of the moment acceleration, wherein k is a condition parameter of the wave crest;
the peak value detection method establishes a threshold value, and only when the peak value acceleration value is larger than the threshold value, the pedestrian can be considered to walk one step; the following formula is shown:
dividing the motion state of pedestrians into three types of walking, resting and running, and providing an adaptive peak detection algorithm according to different acceleration peaks;
the course calculation in PDR mainly depends on the course calculation model, i.e.
wherein ,
and />
The heading at the times t-1 and t are respectively, the initial heading +.>
Initial calibration is performed by magnetometer information; Δt is the interval time of the gyroscope data; Σω is the sum of the angular velocities of the gyroscope outputs at times t-1 to t.
Further, in the step c, according to different motion states of the pedestrian, the step c of estimating the actual motion step of the fitting pedestrian by using the adaptive step specifically includes:
estimating the step size by adopting a mode of fusing a linear step size model and a nonlinear step size model; for the walking state, estimating the step length by adopting a linear step length model in the formula (5); for the running state, a Weinberg nonlinear step-size model which is shown in the formula (8) and is determined by the maximum value and the minimum value of the acceleration is adopted;
S k ＝A+B×LF k +C×LV k (5)
where the parameter A, B, C is a constant, LF k and LVk The step frequency and the acceleration variance of the kth step are respectively calculated as follows:
wherein ,tk and tk+1 The start time and the end time of the kth step, a t The acceleration at the time t is the acceleration at the time t,
for the average acceleration of the kth step, N k A sampling number indicating the acceleration included in the kth step;
wherein K is a constant, a max and amin Representing the maximum and minimum values of the single step acceleration;
in combination with the formula (5) and the formula (8), an adaptive step length estimation algorithm is provided, which is used for meeting the requirement of accurate step counting in two walking modes, namely normal walking and running, and is as follows:
the invention has the advantages and beneficial effects as follows:
the invention discloses a pedestrian dead reckoning method based on human motion state identification, which has the innovation points, advantages and beneficial effects that:
1. the method is used for preprocessing the original data, so that noise interference and false peaks of the original acceleration value, which occur near the peaks, are eliminated, erroneous judgment is avoided, and the accuracy of step counting detection is improved.
2. By analyzing corresponding acceleration peaks in the walking state and the running state of the pedestrian, different acceleration thresholds and time windows are set to improve the original peak detection algorithm to perform step counting detection, and meanwhile, for different motion states, a self-adaptive step calculation formula is adopted to reduce the overall positioning error of the PDR algorithm. Experimental results show that the method achieves good positioning effect in the walking state, running state and walking and running state of pedestrians.
The innovation of the invention mainly comprises: the data preprocessing in the step a adopts moving average filtering to eliminate noise
The false wave peak of the interference and the original acceleration value can appear near the wave peak, thereby avoiding misjudgment and improving the step counting detection
Accuracy; the self-adaptive peak detection algorithm of the step b carries out threshold judgment on the maximum acceleration generated in each step, and realizes the transportation
Classification of dynamic states. Setting threshold according to parameters corresponding to walking state and running state, changing neighborhood window, and realizing different transportation
And (3) self-adaptive judgment of the dynamic state and accurate step counting. The algorithm has good adaptability to the sensor gesture and the pedestrian motion state, and realizes
The step counting accuracy rate of the various sensor postures reaches more than 99%, and the defect that the conventional peak detection off algorithm cannot perform non-correction is overcome
The step counting in a normal state is insufficient; the self-adaptive step length estimation model in the step c is adjusted in real time according to different motion states of pedestrians
And (3) the estimated value of each step is integrated, so that the positioning accuracy is improved.
Drawings
FIG. 1 is a comparison of acceleration data filtering versus back for a preferred embodiment provided by the present invention;
FIG. 2 is a schematic diagram of a PDR algorithm positioning;
FIG. 3 is a graph of walking and running acceleration changes;
FIG. 4 is a block flow diagram of an adaptive peak detection algorithm of the present invention;
FIG. 5 is a flow chart of a pedestrian dead reckoning method based on human motion state identification of the present invention;
FIG. 6 is an experimental roadmap for a validation algorithm;
fig. 7 is a graph showing the reproduction result of the walking track under different algorithms.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and specifically described below with reference to the drawings in the embodiments of the present invention. The described embodiments are only a few embodiments of the present invention.
The technical scheme for solving the technical problems is as follows:
step one: the IMU module is worn on the waist of a pedestrian, data of the accelerometer and the gyroscope are obtained in real time in the motion process of the pedestrian, the data are preprocessed, and the comparison effect before and after the acceleration filtering is shown in the figure 1.
The PDR algorithm positioning principle diagram is shown in figure 2. Setting the pedestrian at the initial time t 0 The starting position of (x) 0 ,y 0 ) Move to the next time t 1 When the step length of the pedestrian is d 1 Heading angle θ 1 Pedestrian t 1 The coordinate information of the time is (x 1 ,y 1 ) Then (x) 0 ,y 0) and (x1 ,y 1 ) The relation between them is shown in the formula (10).
From the above equation, t is after accumulation at a plurality of time points k Coordinate information of pedestrian at moment (x k ,y k ) The formula is satisfied:
step two: and classifying the pedestrian motion state and detecting the step by adopting an adaptive peak detection algorithm. Fig. 3 is a graph showing the change of the acceleration of walking and running, and it can be seen that the acceleration is greatly changed in different exercise states. And according to the characteristics of acceleration in walking and running states, adopting a self-adaptive peak detection algorithm to identify the motion state of the pedestrian and detect the step counting. The flow of the adaptive peak detection algorithm is shown in fig. 4. Firstly, filtering the data, calculating the acceleration peak value in the current state, namely the potential peak value, and then judging whether the current peak value is larger than 1.5g, so that the current motion state of the pedestrian is obtained. And then calculating the time difference between the potential peak value and the previous peak value, and judging whether the time difference meets the corresponding step frequency time window. And finally judging whether the potential peak is the maximum peak value, removing the pseudo peak value, and if the potential peak value is larger than other peak values in the neighborhood, considering the pedestrian to walk one step.
Step three: the step size is estimated by adopting a mode of fusing a linear step size model and a nonlinear step size model, namely, the self-adaptive step size estimation is adopted. For the walking state, the acceleration peak value change is small and the step frequency is low, so that the method is suitable for estimating the step length by adopting a linear step length model of the formula (5). For the running state, the acceleration peak value is large and the step frequency is high, and the Weinberg nonlinear step size model which is shown in the formula (8) and is determined by the maximum value and the minimum value of the acceleration is suitable.
Step four: the traditional PDR algorithm is only suitable for a single motion state of normal walking, has poor positioning effect on pedestrians in running states and the like, and has a large difference between a navigation route and a real track. Therefore, in combination with the self-adaptive peak detection and the self-adaptive step estimation provided above, a pedestrian dead reckoning method based on pedestrian motion state identification is provided and applied to the IMU to realize pedestrian positioning, and a system frame is shown in fig. 5.
The pedestrian dead reckoning method based on the human motion state identification and the traditional PDR algorithm are respectively applied to navigation equipment, experimental verification is carried out according to the route shown in fig. 6, and the results of the different positioning methods are shown in fig. 7. The track errors of the two algorithms are shown in the table 1, so that the track positioning error obtained by the improved PDR algorithm is quite intuitively shown to be smaller, and the algorithm improves the positioning precision.
Table 4 comparison of different algorithm errors
The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.
The above examples should be understood as illustrative only and not limiting the scope of the invention. Various changes and modifications to the present invention may be made by one skilled in the art after reading the teachings herein, and such equivalent changes and modifications are intended to fall within the scope of the invention as defined in the appended claims.
Claims (4)
1. The pedestrian dead reckoning method based on human motion state identification is characterized by comprising the following steps of:
step a, wearing an IMU on the waist of a pedestrian, acquiring data of an accelerometer and a gyroscope in real time in the motion process of the pedestrian, and preprocessing the data;
step b, classifying and step counting detection are carried out on the motion state of the pedestrian by using the self-adaptive peak detection algorithm by utilizing the collected motion data, and the course is calculated by using a course calculation model;
step c, estimating the actual motion step length of the fitted pedestrian by using the self-adaptive step length according to different motion states of the pedestrian;
and d, acquiring the position information of the pedestrians in the moving process in real time according to step counting detection and step length estimation.
2. The pedestrian dead reckoning method based on human motion state recognition according to claim 1, wherein the preprocessing of the data in the step a specifically includes:
combining the collected triaxial acceleration data to obtain a combined acceleration module value according to the following steps:
in the formula ,ax 、a y 、a z The three-axis acceleration value is acquired;
preprocessing the original data, eliminating various noises, and keeping the characteristics of the original data as much as possible; preprocessing the combined acceleration data by adopting a moving average filtering method, wherein the window size is 5.
3. The method for estimating the dead reckoning of the pedestrian based on the identification of the motion state of the human body according to claim 1, wherein the step b uses the collected motion data to classify and step-counting detect the motion state of the pedestrian by using an adaptive peak detection algorithm, and calculates the heading by using a heading estimation model, and the method specifically comprises the following steps:
the peak detection method is characterized in that the characteristic similar to a sine wave is presented by using an acceleration signal when a human body walks, when a wave crest is detected, the pedestrian is considered to walk one step, and the moment of the wave crest is recorded, wherein the moment is shown as the following formula:
where S represents the set of moments labeled acceleration peaks,
and />
Respectively represent t peak-k and tpeak+k The magnitude of the moment acceleration, wherein k is a condition parameter of the wave crest;
the peak value detection method establishes a threshold value, and only when the peak value acceleration value is larger than the threshold value, the pedestrian can be considered to walk one step; the following formula is shown:
dividing the motion state of pedestrians into three types of walking, resting and running, and providing an adaptive peak detection algorithm according to different acceleration peaks;
the course calculation in PDR mainly depends on the course calculation model, i.e.
wherein ,
and />
The heading at the times t-1 and t are respectively, the initial heading +.>
Initial calibration is performed by magnetometer information; deltaT is gyroscope dataIs a time interval of (1); Σω is the sum of the angular velocities of the gyroscope outputs at times t-1 to t.
4. The method for estimating the dead reckoning of the pedestrian based on the human motion state recognition according to claim 3, wherein in the step c, the step of estimating the actual motion step of the pedestrian by using the adaptive step according to the different motion states of the pedestrian comprises the following steps:
estimating the step size by adopting a mode of fusing a linear step size model and a nonlinear step size model; for the walking state, estimating the step length by adopting a linear step length model in the formula (5); for the running state, a Weinberg nonlinear step-size model which is shown in the formula (8) and is determined by the maximum value and the minimum value of the acceleration is adopted;
S k ＝A+B×LF k +C×LV k (5)
where the parameter A, B, C is a constant, LF k and LVk The step frequency and the acceleration variance of the kth step are respectively calculated as follows:
wherein ,tk and tk+1 The start time and the end time of the kth step, a t The acceleration at the time t is the acceleration at the time t,
for the average acceleration of the kth step, N k A sampling number indicating the acceleration included in the kth step;
wherein K is a constant, a max and amin Representing the maximum and minimum values of the single step acceleration;
in combination with the formula (5) and the formula (8), an adaptive step length estimation algorithm is provided, which is used for meeting the requirement of accurate step counting in two walking modes, namely normal walking and running, and is as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310113241.6A CN116092193A (en) | 2023-02-14 | 2023-02-14 | Pedestrian track reckoning method based on human motion state identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310113241.6A CN116092193A (en) | 2023-02-14 | 2023-02-14 | Pedestrian track reckoning method based on human motion state identification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116092193A true CN116092193A (en) | 2023-05-09 |
Family
ID=86208230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310113241.6A Pending CN116092193A (en) | 2023-02-14 | 2023-02-14 | Pedestrian track reckoning method based on human motion state identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116092193A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116793364A (en) * | 2023-06-27 | 2023-09-22 | 汇鲲化鹏(海南)科技有限公司 | Indoor positioning system and method based on dead reckoning |
-
2023
- 2023-02-14 CN CN202310113241.6A patent/CN116092193A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116793364A (en) * | 2023-06-27 | 2023-09-22 | 汇鲲化鹏(海南)科技有限公司 | Indoor positioning system and method based on dead reckoning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Vision-based fall event detection in complex background using attention guided bi-directional LSTM | |
| Mubashir et al. | A survey on fall detection: Principles and approaches | |
| CN111741884A (en) | Method for detecting traffic distress and road rage | |
| KR101165537B1 (en) | User Equipment and method for cogniting user state thereof | |
| US9870535B2 (en) | Method and apparatus for determining probabilistic context awareness of a mobile device user using a single sensor and/or multi-sensor data fusion | |
| US20110190008A1 (en) | Systems, methods, and apparatuses for providing context-based navigation services | |
| WO2021115064A1 (en) | Fitness exercise recognition method based on wearable sensor | |
| CN111523559B (en) | Abnormal behavior detection method based on multi-feature fusion | |
| KR101473957B1 (en) | Apparatus and method for determining insurance premium based on driving pattern recognition | |
| CN108629170A (en) | Personal identification method and corresponding device, mobile terminal | |
| US20210396779A1 (en) | User posture transition detection and classification | |
| Oshin et al. | Energy-efficient real-time human mobility state classification using smartphones | |
| Oshin et al. | ERSP: An energy-efficient real-time smartphone pedometer | |
| KR20150038877A (en) | User interfacing apparatus and method using an event corresponding a user input | |
| Mokhtari et al. | Fall detection in smart home environments using UWB sensors and unsupervised change detection | |
| CN116092193A (en) | Pedestrian track reckoning method based on human motion state identification | |
| CN106971203B (en) | Identity recognition method based on walking characteristic data | |
| Kolobe et al. | A review on fall detection in smart home for elderly and disabled people | |
| CN107688828B (en) | Bus crowding degree estimation method based on mobile phone sensor | |
| CN105975989A (en) | Elbow motion state identification method based on nine-axis motion sensor | |
| EP3382570A1 (en) | Method for characterizing driving events of a vehicle based on an accelerometer sensor | |
| Zinnen et al. | Toward recognition of short and non-repetitive activities from wearable sensors | |
| CN110929766A (en) | Self-adaptive pedestrian mobile phone attitude identification method based on Gaussian mixture clustering algorithm | |
| CN107203259B (en) | Method and apparatus for determining probabilistic content awareness for mobile device users using single and/or multi-sensor data fusion | |
| Martinelli et al. | Daily movement recognition for dead reckoning applications |
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 |