CN107990890B - Multi-sensor tunnel positioning system and positioning method thereof - Google Patents
Multi-sensor tunnel positioning system and positioning method thereof Download PDFInfo
- Publication number
- CN107990890B CN107990890B CN201810046289.9A CN201810046289A CN107990890B CN 107990890 B CN107990890 B CN 107990890B CN 201810046289 A CN201810046289 A CN 201810046289A CN 107990890 B CN107990890 B CN 107990890B
- Authority
- CN
- China
- Prior art keywords
- tunnel
- trend
- robot
- information
- aperture
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 18
- 230000001174 ascending effect Effects 0.000 claims abstract description 23
- 238000012706 support-vector machine Methods 0.000 claims abstract description 21
- 238000012545 processing Methods 0.000 claims abstract description 9
- 238000000605 extraction Methods 0.000 claims abstract description 8
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 7
- 238000010586 diagram Methods 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 5
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000002474 experimental method Methods 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 238000005259 measurement Methods 0.000 claims 2
- 238000003066 decision tree Methods 0.000 description 19
- 230000007547 defect Effects 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000007417 hierarchical cluster analysis Methods 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
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
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
- Manipulator (AREA)
Abstract
The invention relates to a multi-sensor tunnel positioning system and a method thereof. The distance information from the ranging sonar to the hole wall can be obtained by using the ranging sonar; the attitude sensor can be used for obtaining the heading angle information of the robot; depth information of the robot in the water can be obtained by using the depth gauge. The invention relates to a positioning method of a tunnel, which can be divided into 9 categories of ascending, descending, horizontal, advancing, left turning, right turning, aperture shrinkage, aperture expansion, aperture invariable and the like according to the trend of the tunnel. According to the arrangement and combination of the trends, the trends in the holes can be divided into 27 types. And according to the map information of the tunnel, finding out the actual trend category existing in the tunnel from 27 category trends. Through data processing of the sensor, characteristic definition and extraction of the tunnel trend information are carried out, a tunnel trend recognition model based on a support vector machine is constructed, the characteristic position of the robot in the tunnel is recognized, and the positioning function of the robot in the tunnel is realized.
Description
Technical Field
The invention relates to the field of underwater positioning, in particular to an underwater multi-sensor positioning system and a positioning method of a robot in a tunnel.
Background
With the rapid development of automatic control, electronic computers and energy technologies, the application of intelligent underwater vehicles is becoming more and more widespread. The realization of autonomous underwater positioning is a key of the underwater intelligent vehicle to move underwater. In order to guarantee navigation positioning of the underwater intelligent aircraft, the underwater intelligent aircraft is mostly provided with a strapdown inertial navigation system and a Doppler log for dead reckoning.
The strapdown inertial navigation system directly installs an accelerometer and a gyroscope on a carrier, wherein the angular velocity information of the gyroscope sensitive carrier and the specific force information of the accelerometer sensitive carrier are relative to an inertial space; and then converting the proportional information measured by the accelerometer from the carrier coordinate system to the navigation coordinate system through the mathematical platform, so as to calculate navigation parameters. The acceleration information of the strapdown inertial navigation system can be integrated to obtain the speed information, but the speed information is not very accurate and needs to be calibrated by the speed information of the doppler log.
However, when the underwater intelligent robot runs in the tunnel, the function of the Doppler log is disabled due to the limitation of the size of the tunnel, the speed calibration of the Doppler log on the strapdown tubular navigation system is lost, the precision of the whole navigation positioning system is reduced, and therefore the positioning of the robot in the tunnel cannot be realized.
The device combines the actual conditions of the tunnel, has the characteristics of aperture change, tunnel direction change, tunnel gradient change and the like, can measure the characteristic information of the tunnel through a ranging sonar, a depth gauge and an attitude sensor, realizes the positioning of a special point of the robot in the tunnel, and has higher precision relative to the traditional underwater navigation positioning of a strapdown inertial navigation system and a Doppler log based on the ranging sonar, and a multi-sensor tunnel positioning system of the depth gauge and the attitude sensor.
Disclosure of Invention
The invention aims to overcome the defect that the conventional tunnel robot cannot realize accurate positioning in a tunnel, and provides a multi-sensor positioning system and a tunnel positioning method based on a ranging sonar, a depth gauge and an attitude sensor.
The technical scheme of the invention is as follows: the system comprises the steps that distance information from the sonar to the hole wall can be obtained by using the ranging sonar, heading angle information of the robot can be obtained by using the attitude sensor, depth information of the robot in water can be obtained by using the depth gauge, collected data are subjected to data smoothing processing to form a time domain diagram of 3 sensors, tunnel trend signal characteristics are defined according to definition of tunnel trend, characteristic extraction is performed on the tunnel trend signals, modeling recognition is performed through a support vector machine, and the positioning function of the robot at the tunnel characteristic position is realized.
Preferably, the system obtains 27 trend changes in the tunnel according to the arrangement and combination according to 9 basic trends of ascending, descending, advancing, horizontal, left turning, right turning, advancing, aperture shrinkage, aperture expansion and aperture invariance of the tunnel, and then obtains the trend category of the actual tunnel through comparison according to the actual tunnel map information.
Preferably, the system can carry out smoothing processing on the data through a filtering algorithm, so that the processed data is convenient for extracting trend characteristic signals.
Preferably, the system extracts the characteristics of the tunnel trend information through the definition of the tunnel trend classification, the definition of the tunnel trend signal characteristics and the time domain diagram obtained through three sensors, and then trains and tests by using a support vector machine algorithm, and if the effect of the tunnel trend classification accords with the expectation, the system can perform experiments through the defined tunnel trend signal characteristics to realize the positioning of the robot at the special position of the tunnel; and if the classification effect does not meet the expectation, the tunnel trend signal characteristic definition is conducted again.
Preferably, the system can realize the positioning of the robot at a special position in the tunnel through the training and testing of the support vector machine, meanwhile, based on the condition that the robot moves at a uniform speed in the tunnel, the speed information of the robot at each section can be calculated, and the positioning function of the robot in the tunnel at any moment can be realized by combining the built-in time beat information of the robot.
The multi-sensor tunnel positioning method comprises the following steps:
Step 1), measuring distance information from the sonar to the wall surface by using 4 ranging sonars at the head and the tail of the robot, and averaging the distance information of the front and the rear ranging sonars, thereby inhibiting interference to the ranging information caused by movement of the robot in the pipeline.
And 2) acquiring depth information of the robot from the set reference point by using a depth gauge.
And 3) acquiring information of a heading angle of the robot in the tunnel by using an attitude sensor.
Step 4), through the steps 1-3, according to the trend of the tunnel, 9 basic trends of ascending, descending, horizontal, left turning, right turning, advancing, aperture shrinkage, aperture expansion and aperture invariance are carried out according to the trend of the tunnel, and according to the arrangement and combination, 27 trend changes in the tunnel can be obtained. And according to the map information of the tunnel, finding out the actual trend category existing in the tunnel from 27 category trends. Through data processing of the sensor, characteristic definition and extraction of the tunnel trend information are carried out, a tunnel trend recognition model based on a support vector machine is constructed, the characteristic position of the robot in the tunnel is recognized, and the positioning function of the robot in the tunnel is realized.
Preferably, in the step 1), in a working period T, a time domain graph of the measured distances of the two ranging sonars can be drawn according to the average value of the measured distances of the four ranging sonars at the front and the tail of the robot, and if the trend of the curve is a decreasing trend in the working period, it is indicated that the aperture of the tunnel is continuously reduced in the working period; if the trend of the curve is an ascending trend in the working period, the tunnel aperture is continuously enlarged in the path; if the trend of the curve is kept in a horizontal state in the working period, then the curve is in a descending trend, the size of the aperture of the tunnel at the last stage is initially maintained in the path, and then the aperture is continuously reduced; if the trend of the curve is kept in a horizontal state in the working period, then the curve is in an ascending trend, which indicates that the aperture of the tunnel maintains the aperture of the previous stage at first in the path, and then the aperture is continuously enlarged.
Preferably, in the step 2), in a working period T, a time domain diagram of the depth of water can be drawn according to the information obtained by the depth gauge of the robot, and if the trend of the curve is a descending trend in the working period, the tunnel is illustrated as a downhill trend in the path; if the trend of the curve is an ascending trend in the working period, the tunnel is in an ascending trend in the journey.
Preferably, in the step 3), in a working period T, data of a heading angle can be obtained according to an attitude sensor of the robot, a time domain diagram of the heading angle is drawn, and if a trend of a curve is an ascending trend in the working period, it is indicated that in the working period T, the trend of the tunnel is clockwise turning, namely, turning rightwards relative to the original trend; if the trend of the curve is a descending trend in the working period, the tunnel trend is shown to turn anticlockwise in the path, namely, turn leftwards relative to the original trend; if the trend of the curve is horizontal in the working period, the trend of the tunnel is indicated to advance in the journey.
The tunnel trend is divided into 8 categories of ascending, descending, horizontal, left turning, right turning, advancing, aperture shrinking, aperture expanding and the like, and the method based on the support vector machine is used, so that 8 basic trend category definition labels are as follows:
When the trend of the tunnel is that the downhill advancing aperture is unchanged, the label of the tunnel is 269. The remaining 26 cases are analogized.
The waveform characteristic definition of the tunnel trend information mainly comprises two conditions of upper turning and lower turning and upper turning, an upper turning and lower turning value, a lower turning and upper turning value, an upper turning and lower turning difference, an upper turning point curvature, a lower turning point curvature and a lower turning and upper turning difference are defined, wherein 'value' refers to an upper and lower difference value, and 'difference' refers to a left and right difference value. The 8 kinds of basic trends of the tunnel trend are determined by the ratio a of the value to the difference, and the determination of 47 kinds of small kinds of trends formed by the arrangement and combination of the 8 kinds of basic trends is realized by the curvature of the upper inflection point and the curvature of the lower inflection point. .
Extracting characteristic signals:
In the heading angle time domain diagram, when a > b, describing the right turn of the tunnel; when-a > b, the tunnel is described as turning left, and when-b < a < b, the tunnel is described as advancing.
In the depth information time domain graph, when a >0, the tunnel is illustrated as downhill, and when a <0, the tunnel is illustrated as uphill.
In the time domain diagram of the ranging information, when a1>0 and a2>0, the aperture expansion of the tunnel is illustrated; when a1<0 and a2<0, the aperture of the tunnel is reduced.
In the above ranging information time domain diagrams, a1 and a2 respectively represent ratio information of four ranging sonar average values of the head and the tail of the robot, and b is a threshold value, so as to avoid misjudgment of tunnel trend caused by movement of the horizontal plane of the robot.
Through the analysis, three characteristic values of the upper inflection point curvature, the lower inflection point curvature and the ratio a of the value to the difference can be extracted from the sensor time domain graph.
And carrying out data training by using the found characteristic data and the trend category labels defined by the robot, and realizing the identification and positioning of the characteristic trend of the robot in the tunnel by using a multi-class support vector machine based on a decision tree.
According to the time sequence, a time node of each characteristic trend of the robot in the tunnel can be obtained, the distance between the front node and the rear node can be calculated by combining the existing tunnel map, and the speed information in each path can be calculated through the distance and the time difference, so that the position of the robot at each moment in the tunnel can be positioned according to the beat information in the data processing module, and the positioning function is realized.
The beneficial effects of the invention are as follows: the underwater multi-sensor positioning system provided by the invention can effectively realize the positioning function of the intelligent underwater robot in the tunnel, collects and processes the information in the tunnel through the robot, realizes the position matching in the tunnel by utilizing the algorithm of a multi-class support vector machine based on a decision tree, and is a practical tunnel positioning system. Meanwhile, a strapdown inertial navigation system of a traditional underwater navigation system is avoided, and equipment cost is greatly reduced.
Drawings
FIG. 1 is a schematic diagram of the structure of the underwater multi-sensor tunnel locating system of the present invention.
FIG. 2 is a flow chart of tunnel strike signal feature extraction.
Fig. 3 and 4 are defined diagrams of upper inflection point, lower inflection point, upper turn-down turn value, lower turn-up turn value, upper turn-down turn difference, lower turn-up turn difference.
Fig. 5 and 6 are schematic diagrams of decision trees.
Detailed Description
The invention is further described below with reference to the drawings and examples.
Referring to fig. 1, the invention provides a tunnel underwater multi-sensor positioning system, comprising: distance information from the sonar to the hole wall can be obtained by using the distance measuring sonar; the attitude sensor can be used for obtaining the heading angle information of the robot; depth information of the robot in the water can be obtained by using the depth gauge.
The working process of the underwater multi-sensor tunnel locating method of the present invention will be described with reference to fig. 2 to 6.
Extracting a tunnel trend characteristic signal:
referring to fig. 2, distance information from the sonar to the wall surface is measured by using 4 ranging sonars of the bow and tail of the robot, and the distance information of the front and rear four ranging sonars is averaged, so that interference to the ranging information due to movement of the robot in the pipeline is suppressed.
And acquiring depth information of the robot from the set reference point by using a depth gauge.
And acquiring information of a heading angle of the robot in the tunnel by using an attitude sensor.
According to the trend of the tunnel, according to 8 basic trends of ascending, descending, advancing, horizontal, left turning, right turning, advancing, aperture shrinkage, aperture expansion and the like of the tunnel, according to permutation and combination, the trend change in 47 in the tunnel can be obtained. According to the map information of the tunnels, the actual trend categories existing in the tunnels are found out from 47 category trends. Through data processing of the sensor, characteristic definition and extraction of tunnel trend information are carried out, a tunnel trend identification model based on a multi-class support vector machine of a decision tree is constructed, characteristic positions of the robot in the tunnel are identified, and a positioning function of the robot in the tunnel is realized.
In a working period T, a time domain diagram of the measured distances of the two ranging sonars can be drawn according to the average value of the measured distances of the four ranging sonars at the front part and the tail part of the robot, and if the trend of the curve is a descending trend in the working period, the tunnel aperture is continuously reduced in the path; if the trend of the curve is an ascending trend in the working period, the tunnel aperture is continuously enlarged in the path; if the trend of the curve is kept in a horizontal state in the working period, then the curve is in a descending trend, the size of the aperture of the tunnel at the last stage is initially maintained in the path, and then the aperture is continuously reduced; if the trend of the curve is kept in a horizontal state in the working period, then the curve is in an ascending trend, which indicates that the aperture of the tunnel maintains the aperture of the previous stage at first in the path, and then the aperture is continuously enlarged.
In a working period T, a time domain diagram of the depth of water can be drawn according to the information acquired by the depth gauge of the robot, and if the trend of the curve is a descending trend in the working period, the tunnel is a downhill trend in the path; if the trend of the curve is an ascending trend in the working period, the tunnel is in an ascending trend in the journey.
In a working period T, the data of a heading angle can be obtained according to the attitude sensor of the robot, a time domain diagram of the heading angle is drawn, and if the trend of the curve is an ascending trend in the working period, the trend of the tunnel in the working period T is described as clockwise turning, namely rightward turning relative to the original trend; if the trend of the curve is a descending trend in the working period, the tunnel trend is shown to turn anticlockwise in the path, namely, turn leftwards relative to the original trend; if the trend of the curve is horizontal in the working period, the trend of the tunnel is indicated to advance in the journey.
The tunnel trend is divided into 8 categories of ascending, descending, horizontal, left turning, right turning, advancing, aperture shrinking, aperture expanding and the like, and the method based on the support vector machine is used, so that 8 basic trend category definition labels are as follows:
When the trend of the tunnel is going downhill, the label is 16. The remaining 46 cases are analogized.
Referring to fig. 3-4, the waveform characteristic definition of the tunnel trend information mainly comprises two conditions of upper turning and lower turning and upper turning, wherein upper turning and lower turning values are defined, lower turning and upper turning values, upper turning and lower turning differences, upper turning point curvature, lower turning point curvature and lower turning and upper turning differences are defined, wherein 'values' refer to upper and lower difference values, and 'differences' refer to left and right difference values. The 8 kinds of basic trends of the tunnel trend are determined by the ratio a of the value to the difference, and the determination of 47 kinds of small kinds of trends formed by the arrangement and combination of the 8 kinds of basic trends is realized by the curvature of the upper inflection point and the curvature of the lower inflection point.
Extracting characteristic signals:
In the heading angle time domain diagram, when a > b, describing the right turn of the tunnel; when-a > b, the tunnel is described as turning left, and when-b < a < b, the tunnel is described as advancing.
In the depth information time domain graph, when a >0, the tunnel is illustrated as downhill, and when a <0, the tunnel is illustrated as uphill.
In the time domain diagram of the ranging information, when a1>0 and a2>0, the aperture expansion of the tunnel is illustrated; when a1<0 and a2<0, the aperture of the tunnel is reduced.
In the above ranging information time domain diagrams, a1 and a2 respectively represent ratio information of four ranging sonar average values of the head and the tail of the robot, and b is a threshold value, so as to avoid misjudgment of tunnel trend caused by movement of the horizontal plane of the robot.
Through the analysis, three characteristic values of the upper inflection point curvature, the lower inflection point curvature and the ratio a of the value to the difference can be extracted from the sensor time domain graph.
Classifying tunnel trends by using multi-class support vector machine based on decision tree in combination with characteristic values
When used for multi-class pattern recognition, the original problem is typically decomposed into a plurality of two-class recognition problems. Current methods include one-to-many and one-to-one, etc. One-to-many consists of k two classes of SVMs (f 1、…、fk). f j the class label of the j-th class sample is +1 and the class labels of all other samples are-1, so that the output value of the j-th two classes SVM is
The total discriminant function is
f(x)＝argmaxj＝1,…,k gj(x)
The method has the defect that the number of support vectors obtained after training is large, and the recognition speed is influenced.
For this reason, we propose a multi-class support vector machine based on decision trees with the following improvements.
The main idea of the multi-class support vector machine is that hierarchical clustering analysis is firstly carried out on various centers of a training sample set to generate a decision tree, and then (k-1) two classes of SVMs are formed according to the decision tree. When an unknown sample is input, the root node starts, the finer division is carried out step by step along the branches of the decision tree, a certain leaf node is finally reached, and the category represented by the leaf node is the final recognition result. Fig. 5,6 are two possible categories of decision trees.
When k is larger, the more forms of decision trees are selected, and the forms of the decision trees influence the decision effect of the multi-class support vector machine based on the decision trees. If the prior knowledge is provided for various problems to guide the division of the leaf nodes, a SVM decision tree with optimal test speed can be constructed by combining the theory of the decision tree.
If the prior knowledge is not available for various types of problems, a hierarchical clustering algorithm can be adopted to generate a decision tree, and the specific steps are as follows:
(1) For k class samples, k class center sets are calculated:
(2) The first level of division is that each class center is respectively divided into one class;
(3) The second level of division is to combine two centers closest to each other to form a k-1 class;
(4) The hierarchical division is thus performed until the k-1 st level, all centers being merged into class 1.
Thus, a decision binary tree is generated, and the structure of the decision tree is adaptively determined according to the distribution condition of the training sample set and the nearest distance criterion. By comparison, the algorithm of the multi-class support vector machine based on the decision tree has the advantages of less number of support vectors and high test speed.
And carrying out data training by using the found characteristic data and the trend category labels defined by the robot, and realizing the identification and positioning of the characteristic trend of the robot in the tunnel by using a multi-class support vector machine based on a decision tree.
According to the time sequence, a time node of each characteristic trend of the robot in the tunnel can be obtained, the distance between the front node and the rear node can be calculated by combining the existing tunnel map, and the speed information in each path can be calculated through the distance and the time difference, so that the position of the robot at each moment in the tunnel can be positioned according to the beat information in the data processing module, and the positioning function is realized.
Claims (7)
1. A multi-sensor tunnel positioning system is characterized by comprising the steps of obtaining distance information from a sonar to a tunnel wall by using a ranging sonar, obtaining heading angle information of a robot by using a gesture sensor, obtaining depth information of the robot in water by using a depth gauge, performing data smoothing processing on collected data to form a time domain diagram of 3 sensors, calculating speed information of the robot in each section according to ascending, descending, horizontal, left turning, right turning, advancing, aperture shrinkage, aperture expansion and aperture invariance of the tunnel, obtaining 27 trend changes in the tunnel according to arrangement and combination, obtaining trend categories of the actual tunnel by comparing according to actual tunnel map information, performing feature extraction on trend signals of the tunnel, finally performing training and testing of a support vector machine, positioning the robot in a special position in the tunnel, calculating speed information of the robot in each section based on the condition that the robot moves at a constant speed in the tunnel, and combining with built-in time information of the robot to realize the positioning function of the robot in a characteristic position of the tunnel.
2. The multi-sensor tunnel locating system of claim 1, wherein the data can be smoothed by a filtering algorithm such that the processed data facilitates extraction of strike characteristic signals.
3. The multi-sensor tunnel positioning system according to claim 1, wherein the defining of the tunnel trend classification, the defining of the tunnel trend signal characteristics, the extracting of the tunnel trend information characteristics through the time domain diagrams obtained by the three sensors, and the training and testing are carried out by using a support vector machine algorithm, if the tunnel trend classification effect accords with the expectation, the experiment is carried out through the defined tunnel trend signal characteristics, so that the positioning of the robot at the special position of the tunnel is realized; and if the classification effect does not meet the expectation, the tunnel trend signal characteristic definition is conducted again.
4. A multi-sensor tunnel locating method comprising:
Step 1), measuring distance information from the sonar to the wall surface by using 4 ranging sonars at the head and the tail of the robot, and averaging the distance information of the front and the rear ranging sonars, so as to inhibit interference of the movement of the robot in the pipeline on the ranging information;
Step 2), obtaining depth information of a set reference point of the distance of the robot by using a depth gauge;
step 3), acquiring information of a heading angle of the robot in the tunnel by using an attitude sensor;
Step 4), through the steps 1-3, according to the trend of the tunnel, 9 basic trends of ascending, descending, horizontal, left turning, right turning, advancing, aperture shrinkage, aperture expansion and aperture invariance of the tunnel are obtained, and according to arrangement and combination, 27 trend changes exist in the tunnel in total; according to the map information of the tunnel, finding out the actual trend category existing in the tunnel from 27 kinds of classified trends; through data processing of the sensor, characteristic definition and extraction of the tunnel trend information are carried out, a tunnel trend recognition model based on a support vector machine is constructed, the characteristic position of the robot in the tunnel is recognized, and the positioning function of the robot in the tunnel is realized.
5. The method for positioning a multi-sensor tunnel according to claim 4, wherein in the step 1), in a working period T, according to an average value of measurement distances of four ranging sonars at the front and the tail of the robot, a time domain graph of the measurement distances of two ranging sonars is drawn, and if a trend of a curve is a decreasing trend in the working period, it is indicated that the aperture of the tunnel is continuously decreased in the working period; if the trend of the curve is an ascending trend in the working period, the tunnel aperture is continuously enlarged in the path; if the trend of the curve is kept in a horizontal state in the working period, then the curve is in a descending trend, the size of the aperture of the tunnel at the last stage is initially maintained in the path, and then the aperture is continuously reduced; if the trend of the curve is kept in a horizontal state in the working period, then the curve is in an ascending trend, which indicates that the aperture of the tunnel maintains the aperture of the previous stage at first in the path, and then the aperture is continuously enlarged.
6. The method for positioning the multi-sensor tunnel according to claim 4, wherein in the step 2), in a period of working period T, a time domain map of depth of water is drawn according to information obtained by a depth gauge of the robot, and if a trend of a curve is a downward trend in the working period, it is indicated that the tunnel is a downward slope trend in the period of the path; if the trend of the curve is an ascending trend in the working period, the tunnel is in an ascending trend in the journey.
7. The multi-sensor tunnel positioning method according to claim 4, wherein in the step 3), in a period of working period T, according to the data of the heading angle obtained by the attitude sensor of the robot, a time domain diagram of the heading angle is drawn, and if the trend of the curve is an ascending trend in the working period, it is indicated that the trend of the tunnel is clockwise turning, i.e. rightward turning, in the period of the path; if the trend of the curve is a descending trend in the working period, the tunnel trend is shown to turn anticlockwise in the path, namely, turn leftwards relative to the original trend; if the trend of the curve is horizontal in the working period, the trend of the tunnel is indicated to advance in the journey.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810046289.9A CN107990890B (en) | 2018-01-17 | 2018-01-17 | Multi-sensor tunnel positioning system and positioning method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810046289.9A CN107990890B (en) | 2018-01-17 | 2018-01-17 | Multi-sensor tunnel positioning system and positioning method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107990890A CN107990890A (en) | 2018-05-04 |
| CN107990890B true CN107990890B (en) | 2024-04-19 |
Family
ID=62040197
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810046289.9A Active CN107990890B (en) | 2018-01-17 | 2018-01-17 | Multi-sensor tunnel positioning system and positioning method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107990890B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101201626A (en) * | 2007-12-10 | 2008-06-18 | 华中科技大学 | Freedom positioning system for robot |
| CN102116631A (en) * | 2009-12-31 | 2011-07-06 | 北京控制工程研究所 | Method for autonomous determination of gravity direction of moon |
| KR20120103294A (en) * | 2011-03-10 | 2012-09-19 | 주식회사 한화 | Error correction system and method of underwater inertial navigation using movable surface stations |
| CN103047983A (en) * | 2013-01-14 | 2013-04-17 | 哈尔滨工程大学 | Method for navigating underwater robot by matching surface terrain |
| CN105067631A (en) * | 2015-08-04 | 2015-11-18 | 浙江水利水电学院 | Water conveyance tunnel underwater detection method based on dynamic window and detection robot thereof |
-
2018
- 2018-01-17 CN CN201810046289.9A patent/CN107990890B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101201626A (en) * | 2007-12-10 | 2008-06-18 | 华中科技大学 | Freedom positioning system for robot |
| CN102116631A (en) * | 2009-12-31 | 2011-07-06 | 北京控制工程研究所 | Method for autonomous determination of gravity direction of moon |
| KR20120103294A (en) * | 2011-03-10 | 2012-09-19 | 주식회사 한화 | Error correction system and method of underwater inertial navigation using movable surface stations |
| CN103047983A (en) * | 2013-01-14 | 2013-04-17 | 哈尔滨工程大学 | Method for navigating underwater robot by matching surface terrain |
| CN105067631A (en) * | 2015-08-04 | 2015-11-18 | 浙江水利水电学院 | Water conveyance tunnel underwater detection method based on dynamic window and detection robot thereof |
Non-Patent Citations (1)
| Title |
|---|
| 水下机器人技术在密云水库白河泄空隧洞水下探测中的应用;胡明罡;左丰收;邢立丽;;北京水务(第06期);59-62 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107990890A (en) | 2018-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106971194B (en) | Driving intention recognition method based on improved HMM and SVM double-layer algorithm | |
| US11704554B2 (en) | Automated training data extraction method for dynamic models for autonomous driving vehicles | |
| US20080063264A1 (en) | Method for classifying data using an analytic manifold | |
| CN112639502A (en) | Robot pose estimation | |
| Laftchiev et al. | Vehicle localization using in-vehicle pitch data and dynamical models | |
| KR20120064922A (en) | Method for Finger language recognition using EMG and Gyro sensor and Apparatus thereof | |
| Sun et al. | Combining machine learning and dynamic time wrapping for vehicle driving event detection using smartphones | |
| US20190188467A1 (en) | Method for recognizing the driving style of a driver of a land vehicle, and corresponding apparatus | |
| CN109345007B (en) | Advantageous reservoir development area prediction method based on XGboost feature selection | |
| CN110598586A (en) | Target detection method and system | |
| CN111444769B (en) | Laser radar human leg detection method based on multi-scale self-adaptive random forest | |
| WO2020000191A1 (en) | Method for driver identification based on car following modeling | |
| Wang et al. | Matching error of the iterative closest contour point algorithm for terrain-aided navigation | |
| JP6511982B2 (en) | Driving operation discrimination device | |
| CN114140903A (en) | Road type recognition vehicle-mounted device based on decision tree generation rule | |
| CN107990890B (en) | Multi-sensor tunnel positioning system and positioning method thereof | |
| CN112784487B (en) | Flight action recognition method and device | |
| CN108981728B (en) | Intelligent vehicle navigation map building method | |
| CN113076988B (en) | Mobile robot vision SLAM key frame self-adaptive screening method based on neural network | |
| CN111858817B (en) | BilSTM-CRF path inference method for sparse track | |
| CN113532438A (en) | Improved ICCP terrain matching method under large initial positioning error | |
| EP3382570A1 (en) | Method for characterizing driving events of a vehicle based on an accelerometer sensor | |
| CN110378352B (en) | High-precision anti-interference two-dimensional filtering navigation data denoising method in complex underwater environment | |
| Song et al. | Kalman filter based vehicle running data estimation | |
| CN114283397A (en) | Global relocation method, device, equipment and storage medium |
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 |