WO2015192745A1 - 目标即时定位和构建地图的方法与装置 - Google Patents

目标即时定位和构建地图的方法与装置 Download PDF

Info

Publication number
WO2015192745A1
WO2015192745A1 PCT/CN2015/081429 CN2015081429W WO2015192745A1 WO 2015192745 A1 WO2015192745 A1 WO 2015192745A1 CN 2015081429 W CN2015081429 W CN 2015081429W WO 2015192745 A1 WO2015192745 A1 WO 2015192745A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
map
geomagnetic field
range
module
Prior art date
Application number
PCT/CN2015/081429
Other languages
English (en)
French (fr)
Inventor
王新珩
张聪聪
贾尚杰
Original Assignee
无锡知谷网络科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 无锡知谷网络科技有限公司 filed Critical 无锡知谷网络科技有限公司
Priority to JP2016572528A priority Critical patent/JP6400128B2/ja
Priority to KR1020167035078A priority patent/KR101912195B1/ko
Priority to EP15809774.1A priority patent/EP3159658A4/en
Publication of WO2015192745A1 publication Critical patent/WO2015192745A1/zh
Priority to US15/381,307 priority patent/US20170097237A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/04—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by terrestrial means
    • G01C21/08—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by terrestrial means involving use of the magnetic field of the earth
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20—Instruments for performing navigational calculations
    • G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804—Creation or updating of map data
    • G01C21/3807—Creation or updating of map data characterised by the type of data
    • G01C21/383—Indoor data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804—Creation or updating of map data
    • G01C21/3859—Differential updating map data

Definitions

  • the invention relates to the field of indoor positioning and map construction, in particular to a method and a device for combining indoor target positioning and map construction.
  • the problem to be solved by positioning and map construction is that the target uses the sensor data with noise to determine the position of the object in the environment while determining its position.
  • the resolution of the problem of positioning and map construction enables the target to achieve true autonomous navigation.
  • EKF Extended Kalman Filter
  • the problem of simultaneous positioning and map construction has gradually become a research hotspot in the field of mobile robots.
  • the usual practice is to use EKF (Extended Kalman Filter) to simultaneously estimate the position and landmark of the robot.
  • EKF Extended Kalman Filter
  • the method requires an environment to be arranged in advance, for example, it is necessary to arrange a landmark point, thereby consuming human and material resources.
  • the present invention provides a method for target positioning and map construction, which adopts a motion model, corrects the current position by measuring the current geomagnetic field, and performs local map data by using a spatial interpolation algorithm.
  • the update and integration into the global map achieves simultaneous synchronization of target positioning and map construction, while greatly reducing the amount of computation, and eliminating the need to deploy the environment in advance and save resources.
  • a target positioning and map construction method which includes: 1) instantaneously positioning a target moving within a range of activities; 2) instantly updating a geomagnetic field map of a target range of the target according to the positioning of the moving target. 3) Positioning the next position of the target movement according to the updated geomagnetic field map; repeat steps 2) and 3) until the target stops moving.
  • the step 1) may include: initializing a global geomagnetic field map of the target within the range of motion and the range of the activity, the target being provided with a sensor for measuring the distance of the target operation and the angle value and the magnetic field value of the rotation;
  • the geomagnetic field value of the initial position is added to the global geomagnetic field map by spatial interpolation; the control variable of the target operation and the geomagnetic field value of the current position of the target are measured; and the particle filter algorithm is used to estimate the position of the randomly moving target.
  • the step of initializing the geomagnetic field map of the location and the range of activity within the active range may include: abstracting the indoor range of the target activity into grid coordinates, and scaling the size of the grid coordinates to the actual size of the activity room.
  • the initial position of the target within the activity chamber is determined by the target's internal sensors.
  • the measured geomagnetic field value is added to the initial map by spatial interpolation such as Kriging interpolation, and the measured values around the target are weighted to obtain a prediction of the unmeasured position.
  • the meaning of each parameter is: ⁇ i is the unknown weight of the measured value at the i-th position, S 0 is the predicted position, and N is the number of measured values.
  • the weight ⁇ i depends on the fitted model of the spatial relationship between the measured point, the distance of the predicted position, and the measured value around the predicted position.
  • the step of adding the geomagnetic field value to the map using the interpolation algorithm may include: after determining the target position estimate, extracting, in the geomagnetic grid map, the first predetermined range of grid point data around the position target point to interpolate the measured geomagnetic field value And then merging the first predetermined range of grid points obtained after the interpolation into the grid map of the second predetermined range.
  • the present invention also provides a target positioning and map updating device, including a positioning module and a map updating module, wherein the positioning module performs on-the-spot positioning of the movement of the target within the active range; and the map update module updates the target in time according to the positioning of the target.
  • the geomagnetic field map of the active range, the positioning module further performs positioning correction on the next position of the target according to the updated geomagnetic field map until the target stops moving.
  • the positioning module may include: an initialization module configured to initialize the target initial position and the geomagnetic grid map; and a position estimation module configured to estimate the initial position of the target and the position in the motion using the particle filter algorithm and the kinematic model, the map
  • the update module includes an interpolation module configured to interpolate the measured geomagnetic field values of the target range of motion using a spatial interpolation algorithm, and fuse the partial map to the global map.
  • the position estimation module may be implemented based on particle filtering, comprising: a particle filter algorithm module configured to correct a position estimate of the target; and a motion model module configured to analyze the movement of the target within the active range by a predetermined mathematical model.
  • a grid map corresponding to the direction and intensity vectors ⁇ H x , H y , H z > of the magnetic field may be used to form a database, respectively representing the corresponding vector distribution.
  • the features of the map can be increased to facilitate positioning.
  • the constructing map module may include: an interpolation module, after the target position estimation is determined, extracting a predetermined range of grid point data around the position target point in the geomagnetic grid map, interpolating the measured geomagnetic field value, and the fusion module, The inserted predetermined range grid points are merged into a predetermined range of grid maps.
  • the updated magnetic field map is used to further locate the target movement, and the positioning is combined with the map update to accurately and quickly target the target. Position and build a consistent map.
  • the inventor has faced a large number of algorithms through in-depth and extensive research, and selected a particle filter algorithm that corrects the error to update the indoor magnetic field map, and uses the space covariance optimal interpolation method-Kriging algorithm interpolation.
  • the result is highly reliable, and the combination of the two methods can more accurately and quickly locate the robot and construct a continuous and consistent map.
  • the present invention utilizes indoor magnetic field fluctuations to achieve positioning, thereby realizing immediate positioning and Build a map. Therefore, the implementation of the present invention does not require additional placement of reference points required for positioning, thereby reducing cost.
  • FIG. 1 is a schematic diagram of steps of a target positioning method according to an embodiment of the present invention.
  • FIG. 2 is a schematic diagram of an initial position of an initialization target within an active range according to an embodiment of the present invention
  • FIG. 3 is a schematic diagram of an initial map of a geomagnetic field grid of an active range according to an embodiment of the present invention
  • FIG. 4 is a schematic diagram of adding a geomagnetic field value of a target initial position to an initial map by Kriging interpolation according to an embodiment of the present invention
  • FIG. 5 is a schematic diagram of position correction of a moving target according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a geomagnetic field value of a current position of a measurement target according to an embodiment of the present invention
  • FIG. 7 is a schematic diagram of adding a geomagnetic field value to a partial map by using an interpolation method according to an embodiment of the present invention.
  • FIG. 8 is a schematic diagram of merging a partial map into a global map
  • FIG. 9 is a schematic diagram showing the composition of a target real-time positioning and map construction device for a geomagnetic field according to an embodiment of the present invention.
  • FIG. 1 is a schematic diagram showing the steps of a method for real-time target location and map construction of a geomagnetic field in an embodiment of the present invention. As shown in FIG. 1, the steps of the target instant location and map construction method according to an embodiment of the present invention include:
  • Step S101 Perform an on-the-spot measurement on the active area in which the target to be positioned (in this example, an activity room is taken as an example), obtain an indoor plan view, and perform simulation grid map initialization on the plan view.
  • the target to be positioned is combined with the data of the target internal sensor (l i , ⁇ i ), where l i is the distance of the i-th movement, and ⁇ i is the deflection angle of the i-th movement. Determine its position in the simulated initial grid map.
  • 2 is a schematic diagram of an initial position of an initialization target within an active range in an embodiment of the present invention. As shown in Fig. 2, where A is the initial position of the target, the black point next to it is the sampling point of the importance sampling algorithm in the particle filter algorithm for the position space, and the mean of these points represents the estimation of the target position.
  • Step S102 If the target is in the initial position, the particles are evenly distributed around A.
  • the position of A is measured by the target's internal sensor.
  • x k f (x k- 1, v k-1, w k), i.e., by x k-1 and the control input v k-1 can find x k, where, x k is the state at time k, v k-1 is the control input at time k-1 (ie, the distance traveled by the target and the angle of rotation), w k is the correlation error from time k-1 to time k, the sensor itself and its surroundings The environment will make the measurement with a certain error, so it is difficult for the robot to achieve accurate positioning based on these inaccurate control information.
  • the error in the present invention is the sum of random noise and white Gaussian noise.
  • the motion model used in the present invention is:
  • the current position of the target is estimated based on the value of the magnetic field detected by the external sensor carried by the target.
  • the approximate position of the target can be estimated by controlling the input parameters, there are deviations, and such deviations may accumulate. Therefore, in one embodiment, the target position is further corrected so that the measured position coincides with the actual position, and the accumulation of the deviation is eliminated as much as possible.
  • Particle weight is Based on this assumption, the number of particles is 100. Because of the large disadvantage of the particle filter algorithm, the calculation is large and the real-time performance is poor. The number of particles is the key factor determining the size of the calculation. Therefore, in this case, in the case of ensuring the accuracy, reducing the number of particles can reduce the calculation amount and ensure the implementation. Sex. Therefore, the convergence of particle filtering will be very fast and can be achieved in real time;
  • Importance resampling for each particle, update status from To Conditional probability distribution You get a new location for 100 particles. among them, It is the state of the i-th particle at time t-1, and U t is the control input at time t. Particles are sampled from the location space in which the target is located, and the distribution of these particles is a representation of the probability distribution of the location of the target space.
  • Weight calculation The theory of calculating weights is Here,
  • the database used in this example consists of three grid maps of the same size, representing H x , H y , H respectively.
  • the distribution of z In this way, the characteristics of the map will increase, which is conducive to positioning.
  • the calculation of the weight is based on an exponential distribution.
  • the computational complexity of the exponential distribution (generally expressed as time complexity) is O(N), while the computational complexity of the Gaussian distribution is O(N 2 ), where N is the number of particles. It can be seen that the method of calculating the weight based on the exponential distribution in this example can reduce the amount of calculation, speed up the convergence, and improve the positioning accuracy.
  • the exponential distribution function is
  • R.magx is the value of the magnetic field measured in the H x direction of the target at the current position point
  • mapx is the value of the current position on the grid map of the current position
  • Rw is the weight of the particle.
  • the value of the magnetic field on the grid map may be obtained by the interpolation algorithm, or it may be directly added after the target passes here, and R.magy, mapy, R.magz, and mapz are similar.
  • the value of ⁇ can be reasonably adjusted based on the experimental data, which is 2 in this example.
  • a threshold value ⁇ is set, which is set to a ratio of the number of ions according to the empirical value, and can be set to 75% of the total number of particles. If the number of effective particles is less than this threshold, according to the weight of the particles The particle X k is resampled to remove particles with small weights, and the particles with large weights are copied according to the number of particles with small weights removed, and the total number of particles remains unchanged;
  • Estimated target position Calculate the mean value of the spatial distribution of the particle space based on the particles sampled in the previous step. Specifically, the weight of the particle is multiplied by the position coordinate, and then the product is superimposed to estimate the current target position.
  • Step S103 On the indoor simulated grid map obtained in step S101 (shown in FIG. 3), the Kriging interpolation algorithm is used to interpolate the value measured by the target external sensor into the simulated grid map, and the effect of the interpolation is performed.
  • the figure is shown in Figure 4. 4 is a schematic diagram of adding a geomagnetic field value of a target initial position to an initial map by Kriging interpolation according to an embodiment of the present invention, wherein a right portion of the figure is a partial enlargement of the left-side interpolation portion).
  • the geomagnetic field database of the present invention is obtained by measuring a set of real magnetic field values every 0.5 m in the range of 10 m ⁇ 10 m active chamber, and a total of 21 ⁇ 21 groups are measured, and then the 441 groups are selected.
  • the data is spatially interpolated and the result is 81 ⁇ 81 sets of data, due to the stability of the indoor magnetic field fluctuations.
  • this 81 ⁇ 81 group data is the geomagnetic field database used in this example, and the actual value used for the target measurement in the simulation.
  • Step S104 The geomagnetic field value obtained in step S103 is first inserted into the local geomagnetic field.
  • the local geomagnetic field is extracted from the global geomagnetic field.
  • the rule of extraction is: extracting 10 ⁇ 10 grid data centered on the estimated point in the estimated position point grid map; if the estimated point is centered, To a 10 ⁇ 10 local grid, a 10 ⁇ 10 map of the grid corners containing the estimated points is extracted. For example, if the coordinates (x, y) of the estimated point satisfy (abs(y-1) ⁇ 5&&abs(x-1) ⁇ 5), the abscissa point in the global map is 1 to 10, and the ordinate is 1 to 10. Area.
  • the process of merging the local maps (as shown in Figure 7) into the global map is the inverse of the extraction (as shown in Figure 8). This can take into account the correlation between the magnetic field data and the value of the magnetic field around the estimated point, and can also reduce the amount of interpolation.
  • Step S105 Within the range of activities, randomly move the target, collect relevant data obtained by the target internal sensor, and provide a basis for updating the next particle.
  • Step S106 It is detected whether the target stops moving, otherwise it returns to step S102, and the operation is repeated (iteration) until the target stops moving.
  • an embodiment of the present invention further provides a target instant location and a map device, including:
  • the initialization module 201 is configured to initialize the target initial position and the geomagnetic grid map.
  • a position estimation module 202 is configured to estimate a target initial position and a position in motion using a particle filter algorithm and a kinematic model.
  • the map update module 203 is configured to interpolate the geomagnetic field values measured by the sensor by using a spatial interpolation algorithm, and fuse the partial map to the global map.
  • the location estimation module 202 includes:
  • the particle filter algorithm module 2021 includes:
  • Initializing particle distribution module 20211 configured to provide an original reference for movement of subsequent particle sets
  • An importance resampling module 20212 configured to generate a set of particle estimates based on the target movement process
  • the weight calculation module 20213 is configured to assign a weight to the particle, wherein the closer to the target feature, the greater the weight, and vice versa;
  • the normalization weight module 20214 is configured to normalize the weight of the particles so that the sum of the weights of all the particles is equal to one;
  • the effective particle number determining module 20215 is configured to find the number of particles close to the target characteristic
  • a position estimation correction module 20216 is configured to correct an estimate of the target movement position.
  • the particle filter method as described above is used to perform initial particle distribution, importance resampling, weight calculation, normalization weight, operation of calculating the effective particle number, and estimating the current position of the target (robot).
  • the motion model unit 2022 is configured to analyze the movement of the target within the range of motion by a predetermined mathematical model.
  • a nonlinear dynamic model is used, ie
  • the model mainly uses the prior information to predict the observations. After the observation value is obtained, the a posteriori information about the parameter is obtained by the equation, thereby correcting the a priori information. The corrected information is then used as a priori information for prediction. In this way, the a priori information (previous position) of the observation target position is continuously obtained.
  • x t f(x t-1 , v t-1 ) is the same as the motion model in step 102
  • the map update module 203 includes:
  • the interpolation module 2031 is configured to: for example, use the kriging method to weight the surrounding measured values to obtain a prediction of the unmeasured position, and predict from the weighted sum of the data using the following formula: ⁇ i is the position weight of the measured value at the i-th position, S 0 is the predicted position, S i is the random point around the predicted position, Z(S i ) is the value of the magnetic field at the position S i , and N is the number of measured values In this example, it is set to 100.
  • the curve represents the distribution of ⁇ * (h' m ).
  • the semi-variation fitting model used is a circle:
  • Lagrange multiplier solving the equations can be obtained: weight ⁇ i and Lagrange multiplier ⁇ , finally, according to Find the predicted value at S 0 (a 0 , b 0 )
  • the map fusion module 2032 can be configured to extract a local geomagnetic field from a global geomagnetic field.
  • the extracted rule is, for example, extracting 10 ⁇ 10 local grid data centered on the estimated point in the estimated position point global grid map (81 ⁇ 81); if the estimated point is in the peripheral position of the grid map, then extracting A 10 ⁇ 10 map of the grid corners containing the estimated points. Since the local map of the interpolation is completed (as shown in FIG. 7) and the size of the extracted map before the interpolation is completed, the process of merging to the global map (as shown in FIG. 8) is: a partial map extracted from the global map. Cover (or embed) in accordance with the extracted position.

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Automation & Control Theory (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Environmental & Geological Engineering (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Geology (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Navigation (AREA)

Abstract

一种目标即时定位和构建地图的方法和装置，其中该方法包括：1)对在活动范围内移动的目标进行即时定位；2)根据对所述移动目标的定位，即时更新目标所在活动范围的地磁场地图；3)根据所述更新的地磁场地图，对所述目标移动的下一位置进行定位；重复执行步骤2)和3)，直到目标停止移动。该方法可以解决室内目标在没有室内先验磁场地图的条件下，进行即时的目标定位和准确的地磁场地图构建。

Description

目标即时定位和构建地图的方法与装置 技术领域
本发明涉及室内定位及地图构建领域，特别涉及将室内目标的定位和地图构建相结合的方法及装置。
背景技术
同时定位与地图构建所要解决的问题是：目标利用具有噪声的传感器数据在构建出环境中物体的分布情况的同时还要确定自身的位置。另外，定位与地图构建的问题的解决可使目标实现真正的自主导航。在过去的十几年里，同时定位与地图构建的问题逐渐成为移动机器人领域的研究热点。通常的做法是利用EKF(Extended Kalman Filter：扩展卡尔曼滤波器)对机器人位置和地标进行同时估计。但是，当地标位置增多，机器人移动的范围变大时，该方法相应的计算量会变大，实时性变差。此外，该方法需要事先布置环境，例如，需要布置好路标点，因而耗费人力和物力。
发明内容
针对上述现有技术的缺陷，本发明提供了目标定位与地图构建的方法，该方法采用运动模型，通过对当前地磁场的测量，对当前位置进行矫正，和利用空间插值算法对局部地图数据进行更新，并融合到全局地图，达到了目标定位和地图构建的同步进行，同时大大减小了运算量，也不需要事先对环境进行部署，节省资源。
根据本发明的一个方面，提供的目标定位与地图构建方法包括：1)对在活动范围内移动的目标进行即时定位；2)根据对移动目标的定位，即时更新目标所在活动范围的地磁场地图；3)根据更新的地磁场地图，对目标移动的下一位置进行定位；重复执行步骤2)和3)，直到目标停止移动。
在上述方法中，步骤1)可包括：初始化目标在活动范围内的位置和活动范围内的全局地磁场地图，目标设置有测量目标运行的距离和转动的角度值和磁场值的传感器；将目标初始位置的地磁场值通过空间插值法加入初始化全局地磁场地图；测量目标运行的控制变量和目标当前位置的地磁场值；采用粒子滤波算法对随机移动的目标进行位置估计。
初始化目标在活动范围内的位置和活动范围的地磁场地图的步骤可包括：将目标活动的室内范围抽象化为网格坐标，将网格坐标的大小与活动室的实际大小按比例设置。目标在活动室内的初始位置是由目标的内部传感器确定的。
通过空间插值法如克里金插值法将测得的地磁场值加入初始地图，对目标周围的测量值进行加权以得出未测量位置的预测，
Figure PCTCN2015081429-appb-000001
其中各参数的含义是：λi为第i个位置处的测量值的未知权重，S0是预测位置，N是测量值数目。使用克里金方法时，权重λi取决于测量点、预测位置的距离和预测位置周围的测量值之间空间关系的拟合模型。
将地磁场值使用插值算法加入地图的步骤可包括：目标位置估计确定后，在地磁网格地图中抽取位置目标点周围的第一预定范围的网格点数据以将测得的地磁场值插值，然后将插值后得到的第一预定范围的网格点融合到第二预定范围的网格地图中。
另一方面，本发明还提供了目标定位和地图更新装置，包括定位模块和地图更新模块，其中定位模块对目标在活动范围内的移动进行即时定位；地图更新模块根据对目标的定位即时更新目标所在活动范围的地磁场地图，定位模块进一步根据更新的地磁场地图，对目标下一位置进行定位校正，直到目标停止移动。
定位模块可包括：初始化模块，配置为对目标初始位置和地磁场网格地图进行初始化；位置估计模块，配置为使用粒子滤波算法和运动学模型对目标初始位置和运动中的位置进行估计，地图更新模块包括：插值模块,配置为利用空间插值算法将测得的目标活动范围的地磁场值进行插值，和将局部地图融合到全局地图。
位置估计模块可以基于粒子滤波实现，包括：粒子滤波算法模块，配置为对目标的位置估计进行校正；运动模型模块，配置为对目标在活动范围内的移动通过预定数学模型进行分析。
在权重的确定中，可以采用与磁场的方向和强度向量<Hx,Hy,Hz>相应的网格地图构成数据库，分别表示对应的向量分布。由此，可以增加地图的特征，有利于定位。
构建地图模块可包括：插值模块，在目标位置估计确定后，在地磁网格地图中抽取位置目标点周围的预定范围的网格点数据将所测得的地磁场值插值，和融合模块，将插得的预定范围网格点融合到预定范围的网格地图中。
与现有技术相比，由于上述方案采用对目标的即时定位和根据即时定位的结果更新磁场地图，又利用更新的磁场地图对目标移动进一步定位，将定位与地图更新结合来准确快速地对目标进行定位和构建连续一致的地图。
进一步，本发明人通过深入和广泛的研究，面对众多的算法，选择采用校正产生误差的粒子滤波算法对室内地磁场地图的更新，并采用空间协方差最佳插值方法—克里金算法插值，结果可信度高，这两种方法的结合能够更加准确快速地对机器人进行定位和构建连续一致的地图。
在现在的大多数建筑物内，由于钢筋混泥土结构、家具等能影响磁场波动的物件存在，室内磁场波动始终存在，而本发明正是利用室内的磁场波动来实现定位，进而实现即时定位和构建地图。所以，本发明的实现不需要额外布置定位所需的参考点，由此可以减少花费。
附图说明
图1为本发明一实施方式中目标定位方法的步骤示意图；
图2为本发明一实施方式中初始化目标在活动范围内的初始位置的示意图；
图3为本发明一实施方式中活动范围的地磁场网格初始地图的示意图；
图4为本发明一实施方式中将目标初始位置的地磁场值通过克里金插值法加入初始地图的示意图；
图5为本发明一实施方式中对移动的目标进行位置矫正的示意图；
图6为本发明一实施方式中测量目标当前位置的地磁场值示意图；
图7为本发明一实施方式中将地磁场值用插值法加入局部地图的示意图；
图8为将局部地图融合到全局地图的示意图；
图9为本发明一实施方式中地磁场的目标即时定位和地图构建装置的组成示意图。
具体实施方式
下面结合附图对发明作进一步详细的说明。
图1为本发明一种实施方式中的地磁场的即时目标定位和地图构建方法的步骤示意图。如图1所示，本发明一实施方式的目标即时定位和地图构建方法的步骤包括：
步骤S101：对待定位的目标所处的活动区域(本例中以一个活动室为例)进行实地测量，获得室内平面图，并对平面图进行模拟网格地图初始化。对待定位的目标，结合目标内部传感器的数据(li,θi)其中，li为第i次移动的距离，θi为第i次移动的偏向角。确定其在模拟的初始化网格地图中的位置。图2为本发明一种实施方式中初始化目标在活动范围内的初始位置的示意图。如图2所 示，其中A所指位置是目标的初始位置，旁边的黑点是粒子滤波算法中的重要性采样算法对位置空间的采样点，这些采样点的均值表示对目标位置的估计。
步骤S102：若目标处于初始位置，则粒子均匀的分布在A的周围。A的位置是由目标的内部传感器测得的。随着目标的移动，基于运动模型xk＝f(xk-1,vk-1,wk)，即通过xk-1和控制输入量vk-1能求xk，其中，xk为k时刻的状态，vk-1为k-1时刻的控制输入量(即目标运行的距离和转动的角度)，wk为k-1时刻到k时刻的相关误差，传感器本身以及周围的环境会使测量带有一定的误差，所以机器人想要依据这些不确切的控制信息来实现精确地定位是很难的，本发明中的误差为随机噪声和高斯白噪声的累加和。本发明中所使用的运动模型为:
Figure PCTCN2015081429-appb-000002
其中，lt为t-1时刻到t时刻目标移动的距离；θt为t时刻目标相对于t-1时刻目标转过的角度，且以正北方向为基准，按逆时针转；randn(1,2)是产生向量为1×2的随机数，wgn(1,2,0.05)产生0.05dBw的1×2的高斯白噪声。设定的所有粒子会随着运动模型移动(如图5所示)，得到粒子新的位置的采样值。图5中采样点在室内平面图上的移动过程所示，然后再基于目标所携带的外部传感器检测到的磁场值来估计目标的当前位置。虽然通过控制输入参数可以对目标的位置进行大概估计，但存在偏差，并且这种偏差会累积。因此，在一个实施例中，进一步对目标位置进行矫正，使测量的位置和实际的位置一致，尽量消除偏差的累积。
粒子滤波算法的具体实现为：
设置初始粒子分布：粒子的初始分布是在目标(机器人)起始位置周围的均匀分布，产生Num＝100个粒子
Figure PCTCN2015081429-appb-000003
粒子权重为
Figure PCTCN2015081429-appb-000004
基于这种假设，粒子数为100即可。由于粒子滤波算法的一大缺点就是计算量大，实时性差，其中粒子数是决定计算量的大小的关键因素，所以本例中在保证精度的情况下，减少粒子数可以减少计算量，保证实施性。所以粒子滤波的收敛会很快，可以达到实时运行；
重要性重采样：对于每个粒子，更新状态从
Figure PCTCN2015081429-appb-000005
到
Figure PCTCN2015081429-appb-000006
根据条件概率分布
Figure PCTCN2015081429-appb-000007
即可得到100个粒子的新的位置。其中，
Figure PCTCN2015081429-appb-000008
是t-1时刻第i个粒子的状态，Ut是t时刻的控制输入。粒子是从目标所在的位置空间中采样得到的，这些粒子的分布是对目标空间位置的概率分布的表示。
权重计算：计算权重的理论为
Figure PCTCN2015081429-appb-000009
此处，
Figure PCTCN2015081429-appb-000010
为真实分布，而
Figure PCTCN2015081429-appb-000011
为建议性的分布。由于磁场的方向和强度是由向量<Hx,Hy,Hz>表示的，所以本例中采用的数据库是由3个大小相同的网格地图组成，分别表示Hx,Hy,Hz的分布。按此方式，地图的特征会增多，有利于定位。权重的计算是基于指数分布。指数分布的计算复杂度(一般用时间复杂度来表示)是O(N)，而高斯分布的计算复杂度是O(N2)，N表示粒子数。由此可见本例中基于指数分布计算权重的方法可以减少运算量，加快收敛速度，提高定位精度，其指数分布函数为
Figure PCTCN2015081429-appb-000012
其中，R.magx为目标在当前位置点测得的磁场在Hx方向的值，mapx为当前位置网格地图上当前位置的值，R.w为粒子的权重。网格地图上的磁场值可能是由插值算法得到的，也可能目标之前经过这里，是直接加入的值，且R.magy、mapy、R.magz、mapz类似。λ的值可以依据实验所得的数据进行合理的调整，本例中为2。
归一化权重：
Figure PCTCN2015081429-appb-000013
计算有效粒子(如通过权重来确定有效粒子)数
Figure PCTCN2015081429-appb-000014
设定一个阈值δ，如根据经验值，按离子数的比例设置，可设为粒子总数的75％。若有效粒子数小于这个阈值，根据粒子权值
Figure PCTCN2015081429-appb-000015
对粒子Xk进行重采样，去除权值小的粒子，按照去除的权值小的粒子数量复制权值大的粒子，粒子总数保持不变；
估计目标位置：基于上步采样完的粒子，计算粒子空间位置分布的均值，具体是由粒子的权值与位置坐标相乘，然后叠加得到的乘积，估计当前目标位置。
步骤S103：在步骤S101中所获得的室内模拟网格图(如图3所示)上，采用克里金插值算法将目标外部传感器测得的值插值到模拟网格图中，插值后的效果图如图4所示。图4为本发明一种实施方式中将目标初始位置的地磁场值通过克里金插值法加入初始地图的示意图，其中，该图右侧部分是对左侧插值部分的局部放大)。
在仿真实验中，本发明的地磁场数据库的获得是通过在10m×10m的活动室范围内，每隔0.5m测一组真实的磁场值，共测有21×21组，然后对这441组数据进行空间插值运算，结果为81×81组数据，由于室内地磁场波动的稳定 性，这81×81组数据就是本例中采用的地磁场数据库,仿真中用于目标测得的实际值。
步骤S104：在步骤S103中获得的地磁场值首先要***到局部地磁场中。局部地磁场的产生从全局地磁场中抽取，抽取的规则为：在估计的位置点网格地图中抽取以估计点为中心的10×10网格数据实施插值；若估计点为中心，取不到10×10的局部网格，则抽取包含估计点的网格边角10×10地图。如：估计点的坐标(x，y)满足(abs(y-1)<5&&abs(x-1)<5)，则抽取全局地图中横坐标点为1到10的，纵坐标为1到10的区域。插值完成的局部地图(如图7所示)融合到全局地图的过程是抽取的逆过程(如图8所示)。这样可以考虑到磁场数据与估计点周围磁场值的关联性，而且还可以减少插值的运算量。
步骤S105：在活动范围内，随机移动目标，收集目标内部传感器获得的相关数据，为下一步粒子的更新提供依据。
步骤S106：检测目标是否停止运动，否就回到步骤S102，重复操作(迭代)，直到目标停止运动为止。
如图9所示，本发明的一种实施方式中还提供了目标即时定位和构建地图装置，包括：
初始化模块201，配置为对目标初始位置和地磁场网格地图的初始化。
位置估计模块202，配置为使用粒子滤波算法和运动学模型对目标初始位置和运动中的位置进行估计。
地图更新模块203,配置为利用空间插值算法对传感器测得的地磁场值进行插值，并将局部地图融合到全局地图。
其中，在位置估计模块202中，包括：
粒子滤波算法模块2021，包括：
初始化粒子分布模块20211，配置为后面粒子集的移动提供原始的参考；
重要性重采样模块20212，配置为产生基于目标移动过程中的粒子估计集合；
权重计算模块20213，配置为对粒子赋予相关权重大小，其中与目标特性越接近则权重越大，反之越小；
归一化权重模块20214，配置为对粒子的权重进行归一化处理，使所有粒子的权重之和等于一；
有效粒子数确定模块20215，配置为找出与目标特性接近的粒子数；和
位置估计校正模块20216，配置为对目标移动位置的估计进行校正。
通过上述模块的操作，采用如上文所述的粒子滤波方法进行初始粒子分布、重要性重采样、权重计算、归一化权重、计算有效粒子数的操作，估计目标(机器人)的当前位置。
运动模型单元2022，配置为对目标在活动范围内的移动通过预定数学模型进行分析。本例中采用非线性动态模型，即
Figure PCTCN2015081429-appb-000016
该模型主要是利用先验信息对观测值进行预测。在得到观测值后，利用该式得到关于参数的后验信息，从而修正先验信息。再以此修正后的信息作为先验信息进行预测。如此循环进行，不断得到观测值修正目标位置的先验信息(先前位置)。本发明中，xt＝f(xt-1,vt-1)与步骤102中的运动模型相同，而yt＝h(xt,nt)为粒子基于插值产生的地图来修正位置信息。
其中，在地图更新模块203中，包括：
插值模块2031，配置为:例如用克里金法对周围的测量值进行加权以得出未测量位置的预测，使用如下公式由数据加权总和得到预测：
Figure PCTCN2015081429-appb-000017
λi为第i个位置处的测量值的位置权重，S0为预测位置，Si为预测位置周围的随机点，Z(Si)为位置Si处的磁场值，N为测量值数量，本例中设为100。
其中，数据点在网格中用坐标点Si(ai,bi)i＝1,2，...,n表示；待插点坐标用S0(a0,b0)表示，然后求散乱数据点之间的距离
Figure PCTCN2015081429-appb-000018
i＝1,2,...,n；j＝1,2,...,n；将计算所得的距离值按照从小到大排成顺序，然后将距离分成若干组，每组包含一定数量的距离值。将这些距离值组合成距离组(用{h'm}表示)：
Figure PCTCN2015081429-appb-000019
NH表示距离组的个数，根据得出的{h'm}，求出
Figure PCTCN2015081429-appb-000020
(i＝1,...,N(h))分布形状,其中，横坐标是h'm，纵坐标是γ*(h'm)，需要选择合适的模型进行拟合，以产生连续的曲线来表示γ*(h'm)的分布，本例中，采用的半变差拟合模型为圆形：
Figure PCTCN2015081429-appb-000021
其中，c0为块金值，c为基台值，α为变程。通过拟合函数γ*(hm)(其为离散的分布)的分布，可以求出模型参数(如c0，c与α等)，从而得到适合本试验中样本的分布，即变差函数γ*(h)，
然后再用方程组
Figure PCTCN2015081429-appb-000022
其中μ是
Lagrange乘子，求解方程组可以求得：权值λi与Lagrange乘子μ，最后，根据
Figure PCTCN2015081429-appb-000023
求出S0(a0,b0)处的预测值
Figure PCTCN2015081429-appb-000024
地图融合模块2032可配置为：从全局地磁场中抽取局部地磁场。抽取的规则例如为：在估计的位置点全局网格地图(81X 81)中抽取以估计点为中心的10×10局部网格数据实施插值；若估计点在网格地图的周边位置，则抽取包含估计点的网格边角10×10地图。由于插值完成的局部地图(如图7所示)与插值完成之前的抽取地图的大小不变，所以融合到全局地图的过程(如图8所示)为:将从全局地图中抽取的局部地图按照抽取的位置再覆盖(或嵌入)进去。通过上述方式，实质上是创建了一个大小和局部地图一样的小地图(即网格大小与网格上分布的数据一样)，然后经过插值的小地图大小水变，但风格上分布的数据改变了。将插值后的小地图覆盖至全局地图。这样，就可以避免对全局地图插值导致大的计算量。

Claims (14)

  1. 目标即时定位与构建地图的方法，包括：
    1)对在活动范围内移动的目标进行即时定位；
    2)根据对所述移动目标的定位，即时更新目标所在活动范围的地磁场地图；
    3)根据所述更新的地磁场地图，对所述目标移动的下一位置进行定位；
    重复执行步骤2)和3)，直到目标停止移动。
  2. 根据权利要求1所述的方法，其中所述对目标在活动范围内的移动进行即时定位包括：
    初始化目标在活动范围内的位置和活动范围内的全局地磁场地图，所述目标设置有用于测量目标运行的距离和转动的角度值和磁场值的传感器；
    将目标初始位置的地磁场值通过空间插值法加入经过所述初始化的所述全局地磁场地图；
    测量所述目标运行的控制变量和所述目标当前位置的地磁场值；
    采用粒子滤波算法对随机移动的目标进行位置估计。
  3. 根据权利要求2所述的方法，其中所述根据对目标的定位即时更新目标所在活动范围的地磁场地图包括：
    在对目标的位置估计确定后，在所述地磁网格地图中抽取位置目标点周围的预定范围的网格点数据，用所述测得的地磁场值插值，然后将所得到的预定范围网格点融合到预定范围的网格地图中。
  4. 根据权利要求1所述的方法，其中，所述初始化目标在活动范围内的位置和活动范围的地磁场地图包括：
    将目标活动的室内范围抽象化为网格坐标，所述网格坐标的大小与活动室的实际大小按比例设置。
  5. 根据权利要求1-4任一项所述的方法，其中，通过克里金插值法将目标初始位置的地磁场值***初始地磁场地图中，包括：
    对所述目标周围的测量值按如下方式进行加权以得出未测量位置的预测
    Figure PCTCN2015081429-appb-100001
    其中λi为第i个位置处的测量值的未知权重，S0表示预测位置，N为测量值数量，所述权重λi取决于测量点、预测位置的距离和预测位置周围的测量值之间空间关系的拟合模型。
  6. 根据权利要求5所述的方法，其中，所述拟合模型为圆形，其变差函数表达式如下：
    Figure PCTCN2015081429-appb-100002
    其中，C0为块金值，C为基台值，α为变程。
  7. 根据权利要求2所述的方法，其中所述采用粒子滤波算法对随机移动的目标进行定位估计包括：
    测量目标前一时刻到当前时刻移动的距离和偏移的角度，
    设粒子初始为均匀分布，随着目标的移动，所述粒子按与所述目标相同的运动模型移动，基于非线性动态模型
    Figure PCTCN2015081429-appb-100003
    通过粒子的位置均值来估计所述目标的位置。
  8. 根据权利要求7所述的方法，其中所述运动模型为：
    Figure PCTCN2015081429-appb-100004
    其中，lt为t-1时刻到t时刻目标移动的距离；
    θt为t时刻目标相对于t-1时刻目标转过的角度，且以正北方向为基准，按逆时针转；randn(1,2)是产生向量为1×2的随机数，wgn(1,2,0.05)表示产生0.05dBw的1×2的高斯白噪声。
  9. 目标即时定位和地图构建装置，包括定位模块和地图更新模块，其中
    所述定位模块对目标在活动范围内的移动进行即时定位；
    所述地图更新模块根据对目标的定位即时更新目标所在活动范围的地磁场地图，
    所述定位模块进一步根据所述更新的地磁场地图，对目标的下一移动位置进行定位，直到目标停止移动。
  10. 根据权利要求9所述的装置，其中
    所述定位模块包括：
    初始化模块，配置为对目标初始位置和地磁场网格地图进行初始化；
    位置估计模块，配置为使用粒子滤波算法和运动学模型对目标初始位置和运动中的位置进行估计，
    所述地图更新模块包括：
    插值模块,配置为利用空间插值算法将测得的所述目标活动范围的地磁场值插值，和
    地图融合模块，将局部地图融合到全局地图。
  11. 根据权利要求10所述的装置，其中
    所述位置估计模块包括：
    粒子滤波算法模块，配置为对目标的位置估计进行校正；
    运动模型单元，配置为对目标在活动范围内的移动通过预定数学模型进行分析。
  12. 根据权利要求11所述的装置，其中
    所述粒子滤波算法模块包括：
    初始化粒子分布模块，配置为后面粒子集的移动提供原始的参考；
    重要性重采样模块，配置为产生基于目标移动过程中的粒子估计集合；
    权重计算模块，配置为对所述粒子赋予相关权重大小，其中与目标特性越接近则权重越大，反之越小；
    归一化权重模块，配置为对所述粒子的权重进行归一化处理；
    有效粒子数确定模块，配置为找出与目标特性接近的粒子数；和
    位置估计校正模块，配置为对目标移动位置的估计进行校正。
  13. 根据权利要求12所述的装置，其中，所述权重计算模块采用的数据库是基于磁场的方向和强度由向量<Hx,Hy,Hz>表示的三个网格地图数据，各所述向量分别表示Hx,Hy,Hz的分布，所述三个网格地图的网格大小相同。
  14. 根据权利要求13所述的装置，其中所述插值模块配置为在目标位置估计确定后，在地磁网格地图中抽取该位置处所述目标周围的预定范围的网格点数据，对所述测得的地磁场值进行插值，并且还包括：
    融合模块，将所述插值得到的预定范围的网格点融合到预定范围的网格地图中。
PCT/CN2015/081429 2014-06-19 2015-06-15 目标即时定位和构建地图的方法与装置 WO2015192745A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
JP2016572528A JP6400128B2 (ja) 2014-06-19 2015-06-15 目標の実時間測位と地図構築の方法及び装置
KR1020167035078A KR101912195B1 (ko) 2014-06-19 2015-06-15 목표의 실시간 위치 측정과 지도 구축 방법 및 장치
EP15809774.1A EP3159658A4 (en) 2014-06-19 2015-06-15 Method and device for real-time target location and map creation
US15/381,307 US20170097237A1 (en) 2014-06-19 2016-12-16 Method and device for real-time object locating and mapping

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410277471.7A CN104019813B (zh) 2014-06-19 2014-06-19 目标即时定位和构建地图的方法与***
CN201410277471.7 2014-06-19

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/381,307 Continuation US20170097237A1 (en) 2014-06-19 2016-12-16 Method and device for real-time object locating and mapping

Publications (1)

Publication Number Publication Date
WO2015192745A1 true WO2015192745A1 (zh) 2015-12-23

Family

ID=51436695

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/081429 WO2015192745A1 (zh) 2014-06-19 2015-06-15 目标即时定位和构建地图的方法与装置

Country Status (6)

Country Link
US (1) US20170097237A1 (zh)
EP (1) EP3159658A4 (zh)
JP (1) JP6400128B2 (zh)
KR (1) KR101912195B1 (zh)
CN (1) CN104019813B (zh)
WO (1) WO2015192745A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20180061696A (ko) * 2016-11-30 2018-06-08 연세대학교 산학협력단 파티클 간에 지도를 공유하는 이동체의 위치 추정 및 지도 작성 방법 및 장치
CN108469826A (zh) * 2018-04-23 2018-08-31 宁波Gqy视讯股份有限公司 一种基于机器人的地图生成方法及***
CN113701759A (zh) * 2021-08-27 2021-11-26 杭州腓腓科技有限公司 基于可重构超材料的室内同步定位与地图构建方法与***
CN114674307A (zh) * 2022-05-26 2022-06-28 苏州魔视智能科技有限公司 一种重定位方法及电子设备

Families Citing this family (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104019813B (zh) * 2014-06-19 2017-01-25 无锡知谷网络科技有限公司 目标即时定位和构建地图的方法与***
CN105147198A (zh) * 2015-08-10 2015-12-16 深圳先进技术研究院 一种基于扫地机器人的室内地图测绘***及方法
CN105118015A (zh) * 2015-09-21 2015-12-02 无锡知谷网络科技有限公司 用于公共场所的信息提示方法及移动服务终端
US11199850B2 (en) * 2015-10-05 2021-12-14 Pioneer Corporation Estimation device, control method, program and storage medium
CN105676172B (zh) * 2016-01-11 2019-02-22 无锡知谷网络科技有限公司 簇式磁场定位的方法、装置和***
CN105928512A (zh) * 2016-04-26 2016-09-07 杭州欣晟达信息技术有限公司 一种基于地磁场的室内定位方法
CN106289257A (zh) * 2016-07-27 2017-01-04 无锡知谷网络科技有限公司 室内定位方法及定位***
CN106595659A (zh) * 2016-11-03 2017-04-26 南京航空航天大学 城市复杂环境下多无人机视觉slam的地图融合方法
CN106502253B (zh) * 2016-12-22 2019-11-22 深圳乐动机器人有限公司 一种移动装置全自主建图方法及装置
CN106767828A (zh) * 2016-12-29 2017-05-31 南京邮电大学 一种手机室内定位解决方法
JP6673293B2 (ja) * 2017-05-24 2020-03-25 トヨタ自動車株式会社 車両システム
CN107204015B (zh) * 2017-05-27 2021-06-08 中山大学 基于色彩图像和红外图像融合的即时定位与建图***
CN107607107B (zh) * 2017-09-14 2020-07-03 斯坦德机器人（深圳）有限公司 一种基于先验信息的Slam方法和装置
CN107909271A (zh) * 2017-11-15 2018-04-13 国网四川省电力公司经济技术研究院 一种工程勘察数据管理及评估gis***
JP6774445B2 (ja) * 2018-02-05 2020-10-21 本田技研工業株式会社 移動体制御システム、移動体及び移動体制御方法
KR102127359B1 (ko) * 2018-08-22 2020-06-26 한국해양과학기술원 자율수상선을 이용한 수중 자기장 지도 작성 시스템 및 방법
CN111578938B (zh) * 2019-02-19 2022-08-02 珠海格力电器股份有限公司 目标物的定位方法及装置
CN111665470A (zh) * 2019-03-07 2020-09-15 阿里巴巴集团控股有限公司 一种定位方法及装置和机器人
CN110377031B (zh) * 2019-06-28 2022-06-10 炬星科技（深圳）有限公司 运动模型更新方法、装置、电子设备及存储介质
CN110445567B (zh) * 2019-08-06 2021-06-25 中国人民解放军国防科技大学 一种电磁频谱地图的构建方法
CN112566009B (zh) * 2019-09-26 2022-12-27 成都易书桥科技有限公司 一种基于地磁的参与式室内定位***
CN113534826B (zh) * 2020-04-15 2024-02-23 苏州宝时得电动工具有限公司 自移动设备的姿态控制方法、装置及存储介质
CN113566816A (zh) * 2020-04-28 2021-10-29 南宁富桂精密工业有限公司 室内地磁定位方法、服务器及计算机可读存储介质
CN111651547B (zh) * 2020-06-04 2023-07-18 北京四维图新科技股份有限公司 高精度地图数据的获取方法、装置及可读存储介质
CN111879312B (zh) * 2020-07-31 2022-05-17 北京麦钉艾特科技有限公司 一种地磁图在线更新方法
CN112050804B (zh) * 2020-07-31 2022-04-08 东南大学 一种基于地磁梯度的近场磁图构建方法
CN113483747B (zh) * 2021-06-25 2023-03-24 武汉科技大学 基于具有墙角信息的语义地图改进amcl定位方法及机器人
US11845429B2 (en) * 2021-09-30 2023-12-19 GM Global Technology Operations LLC Localizing and updating a map using interpolated lane edge data
US11987251B2 (en) * 2021-11-15 2024-05-21 GM Global Technology Operations LLC Adaptive rationalizer for vehicle perception systems toward robust automated driving control

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101000507A (zh) * 2006-09-29 2007-07-18 浙江大学 移动机器人在未知环境中同时定位与地图构建的方法
US20090093907A1 (en) * 2007-10-05 2009-04-09 Ryoso Masaki Robot System
CN101625572A (zh) * 2009-08-10 2010-01-13 浙江大学 基于改进重采样方法和粒子选取的FastSLAM算法
CN101920498A (zh) * 2009-06-16 2010-12-22 泰怡凯电器(苏州)有限公司 实现室内服务机器人同时定位和地图创建的装置及机器人
CN102445201A (zh) * 2011-09-28 2012-05-09 东北林业大学 用于水下载体的地磁异常特征点匹配导航方法
CN104019813A (zh) * 2014-06-19 2014-09-03 无锡知谷网络科技有限公司 目标即时定位和构建地图的方法与***

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100506097B1 (ko) * 2004-02-04 2005-08-03 삼성전자주식회사 자기장 지도 생성 방법 및 장치와 이를 활용한 이동체의포즈 확인 방법 및 장치
CN101404086B (zh) * 2008-04-30 2012-05-09 浙江大学 基于视频的目标跟踪方法及装置
CN101354252B (zh) * 2008-09-19 2011-04-13 北京航空航天大学 一种基于多尺度估计的地磁辅助导航算法
JP5017392B2 (ja) * 2010-02-24 2012-09-05 クラリオン株式会社 位置推定装置および位置推定方法
CN102256351B (zh) * 2011-05-04 2014-09-17 南京漫城软件科技有限公司 一种基于无线传感网技术的室内精确定位方法
KR101833217B1 (ko) * 2011-12-07 2018-03-05 삼성전자주식회사 자기장 지도 기반 측위 시스템에서 이용되는 이동 단말 및 이를 이용한 위치 추정 방법
JP6160036B2 (ja) * 2012-07-23 2017-07-12 株式会社リコー 移動通信装置、及び位置情報通知方法
CN102829782B (zh) * 2012-09-07 2015-06-17 滨州学院 一种地磁辅助惯性导航方法
CN103838240B (zh) * 2012-11-27 2018-02-27 联想(北京)有限公司 控制方法和电子设备
CN103175529B (zh) * 2013-03-01 2016-01-06 上海美迪索科电子科技有限公司 基于室内磁场特征辅助的行人惯性定位***

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101000507A (zh) * 2006-09-29 2007-07-18 浙江大学 移动机器人在未知环境中同时定位与地图构建的方法
US20090093907A1 (en) * 2007-10-05 2009-04-09 Ryoso Masaki Robot System
CN101920498A (zh) * 2009-06-16 2010-12-22 泰怡凯电器(苏州)有限公司 实现室内服务机器人同时定位和地图创建的装置及机器人
CN101625572A (zh) * 2009-08-10 2010-01-13 浙江大学 基于改进重采样方法和粒子选取的FastSLAM算法
CN102445201A (zh) * 2011-09-28 2012-05-09 东北林业大学 用于水下载体的地磁异常特征点匹配导航方法
CN104019813A (zh) * 2014-06-19 2014-09-03 无锡知谷网络科技有限公司 目标即时定位和构建地图的方法与***

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
LIU, MING ET AL.: "System Modeling of SLAM-based Geomagnetic Aided Inertial Navigation", AVIATION PRECISION MANUFACTURING TECHNOLOGY, vol. 47, no. 6, 30 December 2011 (2011-12-30), pages 13 - 16, XP055245871, ISSN: 1003-5451 *
See also references of EP3159658A4 *
TIAN, MAOJUN ET AL.: "Application of SLAM in Submarine's Geomagnetism Navigation and Positioning Based on EKF", SHIP ELECTRONIC ENGINEERING, vol. 32, no. 8, 30 August 2012 (2012-08-30), pages 52 - 54, XP055245870, ISSN: 1672-9730 *
ZENG, PINSHAN ET AL.: "Research on Simultaneous Localization and Mapping Methods of Autonomous Mobile Robot", ELECTRONIC SCI. & TECH., vol. 26, no. 9, 15 September 2013 (2013-09-15), pages 177 and 178, XP008185470, ISSN: 1007-7820 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20180061696A (ko) * 2016-11-30 2018-06-08 연세대학교 산학협력단 파티클 간에 지도를 공유하는 이동체의 위치 추정 및 지도 작성 방법 및 장치
KR101965296B1 (ko) * 2016-11-30 2019-04-19 연세대학교 산학협력단 파티클 간에 지도를 공유하는 이동체의 위치 추정 및 지도 작성 방법 및 장치
CN108469826A (zh) * 2018-04-23 2018-08-31 宁波Gqy视讯股份有限公司 一种基于机器人的地图生成方法及***
CN113701759A (zh) * 2021-08-27 2021-11-26 杭州腓腓科技有限公司 基于可重构超材料的室内同步定位与地图构建方法与***
CN113701759B (zh) * 2021-08-27 2024-05-03 杭州腓腓科技有限公司 基于可重构超材料的室内同步定位与地图构建方法与***
CN114674307A (zh) * 2022-05-26 2022-06-28 苏州魔视智能科技有限公司 一种重定位方法及电子设备
CN114674307B (zh) * 2022-05-26 2022-09-27 苏州魔视智能科技有限公司 一种重定位方法及电子设备

Also Published As

Publication number Publication date
EP3159658A1 (en) 2017-04-26
KR20170042260A (ko) 2017-04-18
JP6400128B2 (ja) 2018-10-03
EP3159658A4 (en) 2018-01-24
CN104019813B (zh) 2017-01-25
JP2017520841A (ja) 2017-07-27
CN104019813A (zh) 2014-09-03
US20170097237A1 (en) 2017-04-06
KR101912195B1 (ko) 2018-10-26

Similar Documents

Publication Publication Date Title
WO2015192745A1 (zh) 目标即时定位和构建地图的方法与装置
US11295472B2 (en) Positioning method, positioning apparatus, positioning system, storage medium, and method for constructing offline map database
Whelan et al. Deformation-based loop closure for large scale dense RGB-D SLAM
US10288425B2 (en) Generation of map data
JP2018522345A5 (zh)
CN104881029B (zh) 基于一点ransac和fast算法的移动机器人导航方法
Zhang et al. Algorithms analysis of mobile robot SLAM based on Kalman and particle filter
CN102322858B (zh) 用于地磁/捷联惯导组合导航***的地磁匹配导航方法
CN110501010A (zh) 确定移动设备在地理区域中的位置
CN110118556A (zh) 一种基于协方差交叉融合slam的机器人定位方法及装置
KR101960280B1 (ko) 자기장 지도를 수정하는 방법, 자기장 지도를 수정하는 사용자 단말 및 서버
CN106840163A (zh) 一种室内定位方法及***
JP7114686B2 (ja) 拡張現実装置及び位置決め方法
EP3175312A2 (en) Video-assisted landing guidance system and method
CN115307641A (zh) 机器人定位方法、装置、机器人和存储介质
Fissore et al. Towards surveying with a smartphone
Valiente et al. Information-based view initialization in visual SLAM with a single omnidirectional camera
Huang et al. VariFi: Variational Inference for Indoor Pedestrian Localization and Tracking Using IMU and WiFi RSS
Grzonka et al. Look-ahead proposals for robust grid-based slam with rao-blackwellized particle filters
Kupervasser et al. Robust positioning of drones for land use monitoring in strong terrain relief using vision-based navigation
Qian et al. Optical flow-based gait modeling algorithm for pedestrian navigation using smartphone sensors
Geva et al. Estimating camera pose using bundle adjustment and digital terrain model constraints
Liu et al. MagLoc-AR: Magnetic-based localization for visual-free augmented reality in large-scale indoor environments
Reina et al. Iterative path reconstruction for large-scale inertial navigation on smartphones
KR101730714B1 (ko) 지구자기장과 바이큐빅 보간법을 이용한 이동 로봇의 위치 추정 방법 및 장치

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15809774

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2015809774

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2015809774

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2016572528

Country of ref document: JP

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 20167035078

Country of ref document: KR

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE