CN113344118A - Bicycle gray fault detection system and detection method - Google Patents
Bicycle gray fault detection system and detection method Download PDFInfo
- Publication number
- CN113344118A CN113344118A CN202110717565.1A CN202110717565A CN113344118A CN 113344118 A CN113344118 A CN 113344118A CN 202110717565 A CN202110717565 A CN 202110717565A CN 113344118 A CN113344118 A CN 113344118A
- Authority
- CN
- China
- Prior art keywords
- acceleration
- bicycle
- value
- data
- vertical
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 40
- 238000005452 bending Methods 0.000 claims abstract description 6
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 5
- 230000001133 acceleration Effects 0.000 claims description 342
- 230000005484 gravity Effects 0.000 claims description 63
- 238000000034 method Methods 0.000 claims description 46
- 230000008569 process Effects 0.000 claims description 33
- 230000009471 action Effects 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000001914 filtration Methods 0.000 claims description 11
- 238000005259 measurement Methods 0.000 claims description 10
- 230000000737 periodic effect Effects 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000007635 classification algorithm Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 2
- 230000001131 transforming effect Effects 0.000 claims description 2
- 239000000284 extract Substances 0.000 abstract description 3
- 230000009467 reduction Effects 0.000 description 6
- 230000001174 ascending effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Electric Propulsion And Braking For Vehicles (AREA)
Abstract
The invention discloses a bicycle gray level fault detection system and a detection method. The detection system mainly comprises an acquisition module, an analysis module and a detection module, wherein the acquisition module acquires bicycle motion data under a riding state of a user and transmits the bicycle motion data to the analysis module, the analysis module extracts constructed features through processing the data, and the detection module judges whether the bicycle has certain type of gray faults or not through a KNN nearest neighbor algorithm. Gray fault types include wheel bending, tire deflation, brake slack, and handlebar skewing, among others. The bicycle motion data are analyzed, so that the bicycle condition can be automatically determined in real time, and the bicycle motion data have certain auxiliary significance and effect in a shared bicycle operation system and other scenes.
Description
Technical Field
The invention relates to the technical field of information service of bicycles, in particular to a system and a method for detecting gray level faults of a bicycle by using a sensor.
Background
Due to frequent use and carelessness in use, some bicycles may be in bad vehicle conditions, and besides faults which can be easily discovered and obviously damaged, gray faults which can be often discovered only when riding exist, and the faults can also cause riding conditions, influence riding experience and even influence driving safety. These gray faults often include wheel bending, tire deflation, brake slack, and handlebar skewing, among others. In the description of the present invention, a gray fault is also referred to as a gray fault.
At present, the gray level fault is mainly found by depending on the self feeling of a user during riding and the following self report, an automatic and accurate detection method is not available, and the vehicle with the fault cannot be effectively marked.
Disclosure of Invention
In view of the above-mentioned deficiencies of the prior art, an object of the present invention is to provide a bicycle grayscale failure detection system, which can automatically detect the status of a bicycle by a sensor that is easy to deploy during the riding process of a user, so as to provide the user with bicycle status information more accurately and conveniently.
In order to achieve the purpose, the invention adopts the following technical scheme:
the first aspect provides a bicycle gray fault detection system, which comprises an acquisition module, an analysis module and a detection module, wherein the acquisition module acquires bicycle motion data of a user in a riding state and transmits the bicycle motion data to the analysis module, the analysis module processes the bicycle motion data to extract a constructed characteristic value, and the detection module judges whether a certain type of gray fault exists in a bicycle or not through a classification algorithm based on the characteristic value.
Wherein the motion data comprises: the bicycle seat back support is provided with an accelerometer, a first acceleration and a first linear acceleration in three axial directions of X, Y and Z, and a second acceleration and a second linear acceleration in three axial directions of X, Y and Z, wherein the first acceleration and the first linear acceleration are collected by the accelerometer, and the second acceleration and the second linear acceleration are collected by an intelligent mobile terminal carried by a user.
Wherein the analysis module comprises: the alignment unit is used for aligning the acceleration data and the linear acceleration data and reserving the acceleration data and the linear acceleration data in the same time period; the coordinate conversion unit is used for converting the coordinate axes of the acceleration data and converting the three axial acceleration data into the vertical direction; the period calculation unit is used for extracting the pedaling action period in the riding process of the user from the intelligent mobile terminal data; and a feature calculation unit for calculating the extracted feature value.
As a preferred embodiment, the coordinate transformation unit transforming the coordinate axes of the acceleration data includes: and determining a gravity acceleration component through the difference value of the linear acceleration and the acceleration, and then converting the acceleration data of the three axial directions through an included angle relation to obtain the vertical acceleration.
As a further preferred embodiment, the determining the gravitational acceleration component by the difference between the linear acceleration and the acceleration, and then converting the acceleration data of the three axial directions into the vertical acceleration by the included angle relationship includes: for the first acceleration and the first linear acceleration, in each axial direction of X, Y and Z three axial directions, the first acceleration is respectively subtracted by the first linear acceleration to obtain the component of the gravity acceleration on each axial direction, the component is used for calculating the measurement value of the gravity acceleration, and the three axial directions are mutually vertical and are calculated by using the pythagorean theorem; calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the first linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration in the first vertical direction;
for the second acceleration and the second linear acceleration, in each axial direction of the X, Y and Z three axial directions, subtracting the second linear acceleration from the second acceleration to obtain the component of the gravity acceleration on each axial direction, calculating the measurement value of the gravity acceleration by using each component, and calculating by using the pythagorean theorem because the three axial directions are mutually vertical; and calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the second linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration of the second vertical direction.
As a preferred embodiment, the feature calculating unit calculates the extracted feature value includes: divide into initial stage, the stage of riding and the brake stage with whole stage of riding, the eigenvalue of initial stage includes: the number of cycles in the starting process and the transverse average acceleration value of the bicycle in the starting process; the characteristic values of the riding stage comprise: the average value of the first vertical acceleration, the data proportion with the value larger than the average first vertical acceleration value, the data proportion with the value larger than the standard deviation of the first vertical acceleration value and the sum of the average first vertical acceleration value, the normalized standard deviation of the first vertical acceleration value, the average number of the mismatching rates of the first vertical acceleration peak values in all periods and the average number of the first vertical acceleration peak values in all periods; the characteristic values of the braking stage comprise: duration of the braking process, sum of time distances from the point of deceleration value exceeding the threshold to the point of maximum deceleration, average acceleration value of the bicycle forward direction during braking.
In a second aspect, a method for detecting grayscale faults of a bicycle is provided, which comprises the following steps:
1) acquiring acceleration data of a bicycle in operation, wherein the acceleration data comprises: the method comprises the steps that first acceleration and first linear acceleration collected by an accelerometer at a rear support of a bicycle seat, and second acceleration and second linear acceleration collected by a user carrying an intelligent mobile terminal;
2) carrying out mean value filtering processing on the acquired data;
3) time stamp alignment is carried out on the accelerometer at the rear support of the bicycle seat and data collected by an intelligent mobile terminal carried by a user, and data in the same time period are reserved;
4) the method comprises the steps that coordinate system conversion is conducted on an accelerometer at the rear support of a bicycle seat and data collected by an intelligent mobile terminal carried by a user, a gravity acceleration component is determined through the difference value of linear acceleration and acceleration, and then acceleration data of three axial directions are converted through an included angle relation to obtain vertical acceleration;
5) the method comprises the steps that data of an intelligent mobile terminal carried by a user are divided periodically, the action period of pedaling of the user is obtained, and abnormal periodic time periods are marked and removed;
6) dividing riding stages of data and periodic data of an accelerometer at a rear support of a bicycle seat, and then performing feature extraction in a segmented manner;
7) and (3) forming a feature vector by using the extracted features, classifying by using a classification algorithm model after training is finished by using a data set formed by the fault vehicle and the normal vehicle, and judging whether a certain type of fault exists.
As a preferred embodiment, the step 4) of determining the gravitational acceleration component by the difference between the linear acceleration and the acceleration, and then converting the acceleration data of the three axial directions into the vertical acceleration by the angle relationship includes:
for the first acceleration and the first linear acceleration, in each axial direction of X, Y and Z three axial directions, the first acceleration is respectively subtracted by the first linear acceleration to obtain the component of the gravity acceleration on each axial direction, the component is used for calculating the measurement value of the gravity acceleration, and the three axial directions are mutually vertical and are calculated by using the pythagorean theorem; calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the first linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration in the first vertical direction;
for the second acceleration and the second linear acceleration, in each axial direction of the X, Y and Z three axial directions, subtracting the second linear acceleration from the second acceleration to obtain the component of the gravity acceleration on each axial direction, calculating the measurement value of the gravity acceleration by using each component, and calculating by using the pythagorean theorem because the three axial directions are mutually vertical; and calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the second linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration of the second vertical direction.
As a preferred embodiment, in step 6), the whole riding phase is an initial phase, a riding phase and a braking phase, wherein the initial phase characteristics include: the number of cycles in the starting process and the transverse average acceleration value of the bicycle in the starting process; the riding stage features include: the average value of the first vertical acceleration, the data proportion of which is greater than the average first vertical acceleration value, the data proportion of which is greater than the standard deviation of the first vertical acceleration value and the sum of the average first vertical acceleration value, the normalized standard deviation of the first vertical acceleration value, the average of the mismatching rates of the first vertical acceleration peak values in all periods, and the average of the number of the first vertical acceleration peak values in all periods; the braking stage features include: duration of the entire braking process, sum of time distances from the point of deceleration exceeding the threshold value to the point of maximum deceleration in the acceleration time series in the forward direction of the entire braking process, average acceleration value of the bicycle in the forward direction of the entire braking process.
As a preferred embodiment, the fault category of the faulty vehicle in step 7) includes: tire leakage, handlebar skewing, brake slack, and wheel bending.
The invention has the following beneficial effects: according to the bicycle gray scale fault detection system and method, whether a bicycle has certain faults or not can be automatically and accurately detected in the process of riding by a user, and the bicycle can be marked, so that the bicycle gray scale fault detection system and method have positive effects on a riding user and a bicycle maintainer in certain scenes.
Drawings
FIG. 1 is a block diagram of a bicycle grayscale failure detection system according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a bicycle grayscale failure detection system according to another embodiment of the present invention;
fig. 3 is a schematic view of an accelerometer configuration of a bicycle grayscale failure detection system according to an embodiment of the invention.
Detailed Description
In order to facilitate understanding of those skilled in the art, the present invention will be further described with reference to the following examples and drawings, which are not intended to limit the present invention.
Referring to fig. 1, in one embodiment, a bicycle grayscale failure detection system includes an acquisition module, an analysis module, and a detection module, wherein the acquisition module acquires bicycle motion data of a user in a riding state and transmits the bicycle motion data to the analysis module, the analysis module extracts structural features by processing the motion data, and the detection module determines whether a certain type of gray failure exists in a bicycle through a KNN nearest neighbor algorithm.
Referring to fig. 2, in one embodiment, the acquisition module includes a sensor unit, which employs an accelerometer, to acquire acceleration and linear acceleration while the bicycle is moving, and a preprocessing unit. Fig. 3 is a schematic diagram illustrating an accelerometer of a bicycle grayscale failure detection system and a user mobile terminal setting position according to an embodiment of the invention. As shown, the accelerometer sensor 10 is disposed at a rigid frame location at or near the back support of the bicycle seat, and the smart mobile terminal 11 (e.g., a smart phone, with an accelerometer pre-programmed therein) is carried in a rider's trouser-side pocket or in a body-worn backpack/satchel, as the invention is not limited in this respect. Bicycle seat back stay department accelerometer collects the first acceleration, the first linear acceleration of bicycle, and the intelligent mobile terminal that the user carried collects the second acceleration, the second linear acceleration of bicycle, and wherein, each acceleration in first acceleration and the second acceleration includes: an X-axis acceleration, a Y-axis acceleration, and a Z-axis acceleration; each of the first linear acceleration and the second linear acceleration includes: x-axis linear acceleration, Y-axis linear acceleration, and Z-axis linear acceleration. These acceleration and linear acceleration data constitute the bicycle motion data.
For the acquired motion data, noise elimination preprocessing is also needed to remove noise interference caused by the sensor and the surrounding environment. The motion data collected by the sensor unit is sent to a preprocessor unit, which in one embodiment performs mean filtering preprocessing on the motion data. The mean value filtering window is set to be 5, and the acceleration calculation mode after filtering is the mean value of 2 data points before and after the acceleration before filtering and the self.
And sending the preprocessed motion data to an analysis module. With continued reference to fig. 2, in one embodiment, the analysis module includes: the device comprises an alignment unit, a coordinate conversion unit, a period calculation unit and a feature calculation unit. The alignment unit is used for aligning data of the accelerometer and the intelligent mobile terminal, wherein the alignment refers to comparing time stamps of the data on two sides and keeping the data in the same time period, namely keeping a first acceleration, a first linear acceleration, a second acceleration and a second linear acceleration in the same time period.
The coordinate conversion unit is used for converting a coordinate system of data collected by an accelerometer at the rear support of the bicycle seat and the intelligent mobile terminal, determining a gravity acceleration component through a difference value of linear acceleration and acceleration, and then converting acceleration data of three axial directions through an included angle relation to obtain acceleration in a vertical direction. For the first acceleration and the first linear acceleration, in each of the three axial directions of X, Y and Z, the first acceleration is subtracted from the first linear acceleration to obtain the component of the gravitational acceleration in each axial direction, and the component is used to calculate the measurement value of the gravitational acceleration. And according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, cosine values of included angles between the gravity direction and the three axes of X, Y and Z can be calculated. And multiplying the accelerations of the first linear acceleration on the X axis, the Y axis and the Z axis by cosine values of included angles between the gravity direction and the X axis, the Y axis and the Z axis respectively, and summing to obtain the acceleration in the first vertical direction. And calculating the second acceleration and the second linear acceleration in the same way to obtain a second vertical direction acceleration.
The period calculation unit extracts the pedaling action period of the user in riding from the intelligent mobile terminal data, and the action period of the user and the wheel rotation period are in a fixed proportion because the related bicycles have no speed change function or do not use the function in riding, so that the wheel rotation information is obtained, and the fault judgment is facilitated. By periodically dividing the data of the intelligent mobile terminal, the action period of pedaling of the user can be obtained. The user pedals the pedal periodically, two processes of descending and ascending exist in one step, and the second vertical acceleration calculated according to the second acceleration and the second linear acceleration can divide the ascending and descending processes through a zero point, so that the action period of the pedal is obtained. There are also situations where the user is riding a bike without pedaling and sliding, where there is no apparent zero point, removing these abnormal periods.
The feature calculation unit calculates the extracted feature values to form feature vectors. After the original data are divided into pedaling action periods, the periods are divided into different riding stages by using the data of the accelerometer, and the whole riding stage is divided into an initial stage, a riding stage and a braking stage. The y-axis of the accelerometer is in the forward direction when deployed and the x-axis is in the horizontal direction perpendicular to the forward direction. The initial stage is judged by the instability of the bicycle lateral acceleration, i.e. the x-axis linear acceleration. And in the braking stage, judgment is carried out by the acceleration in the advancing direction, namely the total value of the linear acceleration of the y axis is a negative value and reaches a certain threshold value. The characteristic values of the initial stage include: the number of cycles during starting, and the average acceleration value of the bicycle in the transverse direction during starting. The characteristic values of the riding stage comprise: the average value of the first vertical acceleration, the data proportion with the value larger than the average first vertical acceleration value, the data proportion with the value larger than the standard deviation of the first vertical acceleration value and the sum of the average first vertical acceleration value, the normalized standard deviation of the first vertical acceleration value, the average number of the mismatching rates of the first vertical acceleration peak values in all periods and the average number of the first vertical acceleration peak values in all periods. Where the mismatch ratio is used to characterize the relationship between the vertical acceleration peaks. Specifically, after the time of the first vertical direction acceleration peak value in a pedaling action period is added with half of the wheel rotation period time, the peak value closest to the time is found, the time passing between the two peak values is calculated, the distance passed by the wheel is obtained through conversion, the circumference of the wheel is subtracted, and then the circumference of the wheel is divided. The characteristic values of the braking stage comprise: duration of the braking process, sum of time distances from the point of deceleration value exceeding the threshold to the point of maximum deceleration, average acceleration value of the bicycle forward direction during braking.
And the detection module classifies the extracted characteristic value vectors by using a KNN nearest neighbor algorithm to judge whether a certain type of fault occurs. The invention can identify four faults of tyre air leakage, handlebar deflection, brake looseness and wheel bending. The starting cycle number extracted at the initial stage and the average acceleration of the bicycle in the transverse direction in the characteristics are helpful for judging whether the handlebar is inclined, the inclined handlebar causes more starting cycle numbers, the balance is difficult to be grasped, and the transverse acceleration is increased. The brake slack is characterized by the duration of the braking process during the braking phase and the average acceleration in the forward direction of the bicycle during braking. The good braking time is short and the acceleration is large. In the acceleration time sequence in the advancing direction during braking, the characteristic of the sum of the time distances from the deceleration point value exceeding the threshold value to the point of the maximum deceleration characterizes the instability of deceleration when the brake is relaxed and the user stops by stepping on the brake. The tire leakage corresponds to a reduction in riding speed, and the average value of the first vertical acceleration after the speed reduction is reduced, but the number of peaks in all cycles is increased due to the reduction of the cushioning performance. Tire skewing then correspondingly occurs with periodic jounce, i.e. a first vertical acceleration peak occurring periodically. The data ratio corresponding to a value greater than the average first vertical acceleration value, the data ratio of a standard deviation greater than the first vertical acceleration value to a sum of the average first vertical acceleration values, the normalized standard deviation of the first vertical acceleration values, the average of the mismatch rates of the first vertical acceleration peaks in all cycles, the average of the numbers of the first vertical acceleration peaks in all cycles.
In another embodiment of the present invention, based on the concept of the bicycle gray scale fault detection system, a bicycle gray scale fault detection method is provided, which includes the following steps:
1) acceleration data during bicycle operation is collected. The acceleration data comprises: the bicycle seat back support comprises a first acceleration and a first linear acceleration collected by an accelerometer at the position of the bicycle seat back support, and a second acceleration and a second linear acceleration collected by a user carrying a smart phone.
2) And carrying out mean value filtering on the acquired data. The mean value filtering window is set to be 5, and the acceleration calculation mode after filtering is the mean value of 2 data points before and after the acceleration before filtering and the self.
3) And aligning the time stamps of the accelerometer at the rear support of the bicycle seat and the data collected by the user carrying the smart phone, and reserving the data in the same time period.
4) The method comprises the steps of converting coordinate systems of an accelerometer at the rear support of a bicycle seat and data collected by a user carrying a smart phone, determining a gravity acceleration component through a difference value of linear acceleration and acceleration, and converting acceleration data of three axial directions through an included angle relation to obtain vertical acceleration. For the first acceleration and the first linear acceleration, in each of the three axial directions of X, Y and Z, the first acceleration is subtracted from the first linear acceleration to obtain the component of the gravitational acceleration in each axial direction, and the component is used to calculate the measurement value of the gravitational acceleration. And according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, cosine values of included angles between the gravity direction and the three axes of X, Y and Z can be calculated. And multiplying the accelerations of the first linear acceleration on the X axis, the Y axis and the Z axis by cosine values of included angles between the gravity direction and the X axis, the Y axis and the Z axis respectively, and summing to obtain the acceleration in the first vertical direction. And calculating the second acceleration and the second linear acceleration in the same way to obtain a second vertical direction acceleration.
5) The method comprises the steps of carrying out periodic division on data of the intelligent mobile terminal carried by a user, obtaining the action period of pedaling of the user, and marking and removing abnormal periodic time periods. The user pedals the pedal periodically, a descending process and an ascending process exist in one-time pedaling, and the ascending process and the descending process can be divided through a zero point according to a second vertical acceleration calculated according to the second acceleration and the second linear acceleration, so that the action period of the pedaling is obtained. There are also situations where the user is riding a bike without pedaling and sliding, where there is no apparent zero point, removing these abnormal periods.
6) And carrying out riding stage division on the data of the accelerometer at the rear support of the bicycle seat and the periodic data, and then carrying out feature extraction in a segmented manner.
6.1) the whole riding phase is an initial phase, a riding phase and a braking phase. After dividing the raw data into pedaling motion cycles, the cycles are divided into different riding phases by using the data of the accelerometer. The y-axis of the accelerometer is in the forward direction when deployed and the x-axis is in the horizontal direction perpendicular to the forward direction. The initial stage is judged by the instability of the bicycle lateral acceleration, i.e. the x-axis linear acceleration. And in the braking stage, judgment is carried out by the acceleration in the advancing direction, namely the total value of the linear acceleration of the y axis is a negative value and reaches a certain threshold value.
6.2) calculating the extracted feature values according to the following feature definitions: the initial stage features include: the number of cycles during starting, and the average acceleration value of the bicycle in the transverse direction during starting. The riding stage features include: the average value of the first vertical acceleration, the data proportion of which is greater than the average first vertical acceleration value, the data proportion of which is greater than the standard deviation of the first vertical acceleration value and the sum of the average first vertical acceleration value, the normalized standard deviation of the first vertical acceleration value, the average of the mismatch rates of the first vertical acceleration peak values in all periods, and the average of the number of the first vertical acceleration peak values in all periods. Where the mismatch ratio is used to characterize the relationship between the vertical acceleration peaks. Specifically, after the time of the first vertical direction acceleration peak in a pedaling action period is added with the time of a half wheel rotation period, the peak closest to the time is found, and the distance traveled by the wheel is obtained by subtracting the circumference of the wheel through calculating the time conversion between the two peaks, and dividing the distance by the circumference of the wheel. The braking stage features include: duration of the entire braking process, sum of time distances from the point of deceleration exceeding the threshold value to the point of maximum deceleration in the acceleration time series in the forward direction of the entire braking process, average acceleration value of the bicycle in the forward direction of the entire braking process.
7) And (3) forming a feature vector for the extracted features, classifying by using a KNN nearest neighbor algorithm model after training is finished by using a data set formed by the fault vehicle and the normal vehicle, and judging whether a certain type of fault exists.
The invention can identify four faults of tyre air leakage, handlebar deflection, brake looseness and wheel bending. The starting cycle number extracted at the initial stage and the average acceleration of the bicycle in the transverse direction in the characteristics are helpful for judging whether the handlebar is inclined, the inclined handlebar causes more starting cycle numbers, the balance is difficult to be grasped, and the transverse acceleration is increased. The brake slack is characterized by the duration of the braking process during the braking phase and the average acceleration in the forward direction of the bicycle during braking. The good braking time is short and the acceleration is large. In the acceleration time sequence in the advancing direction during braking, the characteristic of the sum of the time distances from the deceleration point value exceeding the threshold value to the point of the maximum deceleration characterizes the instability of deceleration when the brake is relaxed and the user stops by stepping on the brake. The tire leakage corresponds to a reduction in riding speed, and the average value of the first vertical acceleration after the speed reduction is reduced, but the number of peaks in all cycles is increased due to the reduction of the cushioning performance. Tire skewing then correspondingly occurs with periodic jounce, i.e. a first vertical acceleration peak occurring periodically. The data ratio corresponding to a value greater than the average first vertical acceleration value, the data ratio of a standard deviation greater than the first vertical acceleration value to a sum of the average first vertical acceleration values, the normalized standard deviation of the first vertical acceleration values, the average of the mismatch rates of the first vertical acceleration peaks in all cycles, the average of the numbers of the first vertical acceleration peaks in all cycles.
The bicycle gray fault detection system and the bicycle gray fault detection method provided by the embodiment of the invention can accurately detect gray fault types, automatically detect whether the vehicle has some abnormal faults or not, report the self condition of the vehicle, provide convenience for maintaining the vehicle and provide guarantee for the riding experience of a user.
Claims (10)
1. A bicycle gray fault detection system is characterized by comprising an acquisition module, an analysis module and a detection module, wherein the acquisition module is used for acquiring bicycle motion data of a user in a riding state; the analysis module is used for analyzing and processing the bicycle motion data and extracting a constructed characteristic value; the detection module is used for judging whether the bicycle has certain type of gray faults or not through a classification algorithm according to the characteristic values.
2. The bicycle grayscale failure detection system of claim 1, wherein the motion data includes: the bicycle seat back support is provided with an accelerometer, a first acceleration and a first linear acceleration in three axial directions of X, Y and Z, and a second acceleration and a second linear acceleration in three axial directions of X, Y and Z, wherein the first acceleration and the first linear acceleration are collected by the accelerometer, and the second acceleration and the second linear acceleration are collected by an intelligent mobile terminal carried by a user.
3. The bicycle grayscale fault detection system of claim 2, wherein the analysis module includes: the alignment unit is used for aligning the acceleration data and the linear acceleration data and reserving the acceleration data and the linear acceleration data in the same time period; the coordinate conversion unit is used for converting the coordinate axes of the acceleration data and converting the three axial acceleration data into the vertical direction; the period calculation unit is used for extracting the pedaling action period in the riding process of the user from the intelligent mobile terminal data; and a feature calculation unit for calculating the extracted feature value.
4. The bicycle grayscale failure detection system of claim 3, wherein the coordinate transformation unit transforming the acceleration data coordinate axis includes: and determining a gravity acceleration component through the difference value of the linear acceleration and the acceleration, and then converting the acceleration data of the three axial directions through an included angle relation to obtain the vertical acceleration.
5. The gray scale bicycle fault detection system of claim 4, wherein the determining the gravitational acceleration component from the difference between the linear acceleration and the acceleration and then converting the acceleration data of the three axial directions into the vertical acceleration by the angle relationship comprises: for the first acceleration and the first linear acceleration, in each axial direction of X, Y and Z three axial directions, the first acceleration is respectively subtracted by the first linear acceleration to obtain the component of the gravity acceleration on each axial direction, the component is used for calculating the measurement value of the gravity acceleration, and the three axial directions are mutually vertical and are calculated by using the pythagorean theorem; calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the first linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration in the first vertical direction;
for the second acceleration and the second linear acceleration, in each axial direction of the X, Y and Z three axial directions, subtracting the second linear acceleration from the second acceleration to obtain the component of the gravity acceleration on each axial direction, calculating the measurement value of the gravity acceleration by using each component, and calculating by using the pythagorean theorem because the three axial directions are mutually vertical; and calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the second linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration of the second vertical direction.
6. The bicycle grayscale failure detection system according to claim 5, wherein the feature calculation unit calculates the extracted feature value includes: divide into initial stage, the stage of riding and the brake stage with whole stage of riding, the eigenvalue of initial stage includes: the number of cycles in the starting process and the transverse average acceleration value of the bicycle in the starting process; the characteristic values of the riding stage comprise: the average value of the first vertical acceleration, the data proportion with the value larger than the average first vertical acceleration value, the data proportion with the value larger than the standard deviation of the first vertical acceleration value and the sum of the average first vertical acceleration value, the normalized standard deviation of the first vertical acceleration value, the average number of the mismatching rates of the first vertical acceleration peak values in all periods and the average number of the first vertical acceleration peak values in all periods; the characteristic values of the braking stage comprise: duration of the braking process, sum of time distances from the point of deceleration value exceeding the threshold to the point of maximum deceleration, average acceleration value of the bicycle forward direction during braking.
7. The bicycle grayscale failure detection system of claim 1, wherein the collection module further comprises a preprocessing unit for performing mean filtering on the collected motion data, and/or
And the detection module classifies the extracted characteristic value vectors by using a KNN nearest neighbor algorithm and judges whether a certain type of fault occurs.
8. A method for detecting gray scale faults of a bicycle is characterized by comprising the following steps:
1) acquiring acceleration data of a bicycle in operation, wherein the acceleration data comprises: the method comprises the steps that first acceleration and first linear acceleration collected by an accelerometer at a rear support of a bicycle seat, and second acceleration and second linear acceleration collected by a user carrying an intelligent mobile terminal;
2) carrying out mean value filtering processing on the acquired data;
3) time stamp alignment is carried out on the accelerometer at the rear support of the bicycle seat and data collected by an intelligent mobile terminal carried by a user, and data in the same time period are reserved;
4) the method comprises the steps that coordinate system conversion is conducted on an accelerometer at the rear support of a bicycle seat and data collected by an intelligent mobile terminal carried by a user, a gravity acceleration component is determined through the difference value of linear acceleration and acceleration, and then acceleration data of three axial directions are converted through an included angle relation to obtain vertical acceleration;
5) the method comprises the steps that data of an intelligent mobile terminal carried by a user are divided periodically, the action period of pedaling of the user is obtained, and abnormal periodic time periods are marked and removed;
6) dividing riding stages of data and periodic data of an accelerometer at a rear support of a bicycle seat, and then performing feature extraction in a segmented manner;
7) and (3) forming a feature vector by using the extracted features, classifying by using a classification algorithm model after training is finished by using a data set formed by the fault vehicle and the normal vehicle, and judging whether a certain type of fault exists.
9. The gray scale fault detection method for bicycles of claim 8, wherein the step 4) of determining the gravity acceleration component by the difference between the linear acceleration and the acceleration and then converting the acceleration data of the three axial directions into the vertical acceleration by the angle relationship comprises:
for the first acceleration and the first linear acceleration, in each axial direction of X, Y and Z three axial directions, the first acceleration is respectively subtracted by the first linear acceleration to obtain the component of the gravity acceleration on each axial direction, the component is used for calculating the measurement value of the gravity acceleration, and the three axial directions are mutually vertical and are calculated by using the pythagorean theorem; calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the first linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration in the first vertical direction;
for the second acceleration and the second linear acceleration, in each axial direction of the X, Y and Z three axial directions, subtracting the second linear acceleration from the second acceleration to obtain the component of the gravity acceleration on each axial direction, calculating the measurement value of the gravity acceleration by using each component, and calculating by using the pythagorean theorem because the three axial directions are mutually vertical; and calculating cosine values of included angles between the gravity direction and the three axes of X, Y and Z according to the measured value of the gravity acceleration and the gravity acceleration component on each axis, multiplying the acceleration of the second linear acceleration on the axes of X, Y and Z by the cosine values of the included angles between the gravity direction and the three axes of X, Y and Z respectively, and summing to obtain the acceleration of the second vertical direction.
10. The gray scale fault detection method for bicycles of claim 9, wherein the whole riding phase in step 6) is an initial phase, a riding phase and a braking phase, wherein the initial phase characteristics comprise: the number of cycles in the starting process and the transverse average acceleration value of the bicycle in the starting process; the riding stage features include: the average value of the first vertical acceleration, the data proportion of which is greater than the average first vertical acceleration value, the data proportion of which is greater than the standard deviation of the first vertical acceleration value and the sum of the average first vertical acceleration value, the normalized standard deviation of the first vertical acceleration value, the average of the mismatching rates of the first vertical acceleration peak values in all periods, and the average of the number of the first vertical acceleration peak values in all periods; the braking stage features include: the duration of the entire braking process, the sum of the time distances from the deceleration point value exceeding the threshold value to the point of maximum deceleration in the acceleration time series in the forward direction in the entire braking process, and the average acceleration value of the bicycle in the forward direction in the entire braking process; and/or
The fault category of the fault vehicle in the step 7) comprises: tire leakage, handlebar skewing, brake slack, and wheel bending.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110717565.1A CN113344118B (en) | 2021-06-28 | 2021-06-28 | Bicycle gray level fault detection system and detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110717565.1A CN113344118B (en) | 2021-06-28 | 2021-06-28 | Bicycle gray level fault detection system and detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113344118A true CN113344118A (en) | 2021-09-03 |
| CN113344118B CN113344118B (en) | 2023-12-26 |
Family
ID=77479128
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110717565.1A Active CN113344118B (en) | 2021-06-28 | 2021-06-28 | Bicycle gray level fault detection system and detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113344118B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113865885A (en) * | 2021-09-26 | 2021-12-31 | 青岛迈金智能科技股份有限公司 | Method and device for detecting bicycle loss |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1765692A (en) * | 2005-11-03 | 2006-05-03 | 李平 | Small-sized electric motion/power assistance bicycle/tricycle and its controller and sensor |
| CN103018058A (en) * | 2012-12-17 | 2013-04-03 | 北京交通大学 | Similarity-based fault isolation method of train suspension system |
| CN108715201A (en) * | 2018-05-24 | 2018-10-30 | 冷承霖 | A kind of bicycle induction brake system and method |
| CN110966978B (en) * | 2018-09-28 | 2021-10-26 | 千寻位置网络有限公司 | Bicycle tire deformation detection method and device and bicycle |
| CN209356189U (en) * | 2019-03-28 | 2019-09-06 | 吉林大学 | A kind of brake data acquisition device, the vehicles and tire for vehicles |
| GB2588237B (en) * | 2019-10-18 | 2023-12-27 | Mclaren Applied Ltd | Joint axis direction estimation |
| US11592810B2 (en) * | 2019-12-16 | 2023-02-28 | Woven Planet North America, Inc. | Systems and methods for injecting faults into an autonomy system |
-
2021
- 2021-06-28 CN CN202110717565.1A patent/CN113344118B/en active Active
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113865885A (en) * | 2021-09-26 | 2021-12-31 | 青岛迈金智能科技股份有限公司 | Method and device for detecting bicycle loss |
| CN113865885B (en) * | 2021-09-26 | 2024-04-26 | 青岛迈金智能科技股份有限公司 | Method and device for detecting bicycle loss |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113344118B (en) | 2023-12-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110751051B (en) | Abnormal driving behavior detection method based on machine vision | |
| CN102473281B (en) | Vehicle vicinity monitoring apparatus | |
| CN109000935B (en) | Method for judging performance of new energy automobile brake system | |
| CN113344118B (en) | Bicycle gray level fault detection system and detection method | |
| CN107512143B (en) | Tire pressure calculation method and device and application to bicycle | |
| KR101473957B1 (en) | Apparatus and method for determining insurance premium based on driving pattern recognition | |
| CN101894252A (en) | Walking movement classification method based on triaxial acceleration transducer signals | |
| CN111721324B (en) | Contact net dropper breakage detection method based on acceleration signals | |
| JP2017073021A (en) | Driving support device | |
| CN111680613A (en) | Method for detecting falling behavior of escalator passengers in real time | |
| CN107358248B (en) | Method for improving falling detection system precision | |
| CN113657265B (en) | Vehicle distance detection method, system, equipment and medium | |
| JP5330143B2 (en) | Moving form discriminating method, moving form discriminating apparatus, and calorie consumption calculating method | |
| CN113221759A (en) | Road scattering identification method and device based on anomaly detection model | |
| CN113222187A (en) | Intelligent monitoring method for brake health degree of shared moped | |
| CN112370048A (en) | Movement posture injury prevention method and system based on joint key points and storage medium | |
| JP4899725B2 (en) | Step counting device | |
| CN113129602A (en) | Vehicle state monitoring method and device, storage medium and electronic equipment | |
| EP3720743A1 (en) | A system for determining an angular speed of an axle of a railway vehicle and corresponding method | |
| CN109190153A (en) | A kind of Calculation Method of Energy Consumption and its system | |
| CN107807056A (en) | A kind of auto parts and components lesion assessment system based on acceleration loading spectrum | |
| CN114084199B (en) | Train stability evaluation method and system based on recursive graph analysis | |
| CN111174885A (en) | Method for classifying and sectionally acquiring signals of vehicle dynamic weighing sensor | |
| CN116242465B (en) | Dynamic vehicle weighing method and system | |
| JP5085166B2 (en) | Tire pressure detector |
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 |