CN-122017731-A - AR navigation method and system based on GIS and VGIS fusion

CN122017731ACN 122017731 ACN122017731 ACN 122017731ACN-122017731-A

Abstract

The invention discloses an AR navigation method and system based on GIS and VGIS fusion, which relate to the technical field of radio navigation, wherein a ranging residual fingerprint vector reflecting a signal propagation environment is constructed based on a terminal ranging observation value, step intensity of a GIS planning route and VGIS fine-scale route path difference vector is counted, candidate step boundary bands are initially screened, fingerprint vector space distribution is established, fingerprint transition quantity is calculated to screen and confirm key step boundary bands, a route is segmented into stable segments by utilizing the key step boundary bands, independent translation quantity is calculated based on displacement samples, a segment alignment route is generated, transition route coordinates are calculated in the boundary bands, and a final continuous fusion route is generated in a combined mode and AR navigation is performed. The invention adopts a geometric form and fingerprint double-checking mechanism, can accurately distinguish systematic deviation caused by random noise and obstacles, realizes accurate spatial fusion of virtual guidance and a real road surface under a dynamic construction scene, and avoids navigation through-mould and visual drift.

Inventors

WANG SHENG
PAN CONGZHENG
WANG LINTAO
ZOU CHANGHUA
SUN ZHENYU

Assignees

影育(上海)科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260129

Claims (10)

1. The AR navigation method based on the fusion of the GIS and VGIS is characterized by comprising the following steps: calculating a terminal position estimation value based on the reference point coordinate set and the terminal ranging observation value, calculating an observation residual component between the terminal ranging observation value and a theoretical distance determined according to the terminal position estimation value, and constructing a ranging residual fingerprint vector reflecting a signal propagation environment; calculating a path difference vector of the GIS planning path and the VGIS fine-scale path at the same arc length, counting the change characteristics of the path difference vector to obtain step intensity, and extracting candidate step boundary bands based on the distribution characteristics of the step intensity; Establishing a space distribution data set of distance measurement residual fingerprint vectors along the arc length of a route, respectively obtaining distance measurement residual fingerprint vector sets of a left neighborhood and a right neighborhood of a candidate step boundary zone, calculating the statistical difference between the two to obtain a fingerprint transition quantity, and screening the candidate step boundary zone based on the fingerprint transition quantity to obtain a key step boundary zone; Dividing the VGIS fine-scale route into stable segments by utilizing a key step boundary band, determining independent translation amounts of the stable segments based on the terminal position estimated value falling into the stable segments and the coordinate difference value of the VGIS fine-scale route, and correcting the stable segments according to the independent translation amounts to obtain a segment alignment route; And calculating transition route coordinates of the alignment routes of the adjacent segments in the critical step boundary zone, combining to obtain a final continuous fusion route, and performing augmented reality navigation.
2. The AR navigation method based on GIS-VGIS fusion according to claim 1, comprising: determining a terminal position estimated value by a least square optimization algorithm based on the datum point coordinate set and the terminal ranging observed value, wherein the terminal position estimated value is a three-dimensional coordinate of the terminal in a geometric space under the current observation noise level; For each datum point in the datum point coordinate set, calculating a theoretical Euclidean distance between the coordinate of the datum point and a terminal position estimated value, subtracting the theoretical Euclidean distance from an actual measurement terminal ranging observed value corresponding to the datum point, and obtaining a difference value which is an observed residual component corresponding to the datum point, wherein the observed residual component is used for representing the non-line-of-sight shielding degree on a signal transmission path; And arranging the observation residual components corresponding to all the datum points in the datum point coordinate set in sequence according to the preset numbering sequence of the datum points to form a multi-dimensional vector serving as a ranging residual fingerprint vector.
3. The AR navigation method based on GIS-VGIS fusion according to claim 1, comprising: Performing arc length parameterization resampling on the GIS planning route and the VGIS fine standard route, and taking a difference vector obtained by subtracting the space coordinate vector of the GIS planning route from the space coordinate vector of the VGIS fine standard route at the same arc length sampling point as a path difference vector; the step intensity is obtained by counting the change characteristics of the path difference vector, and the method is concretely as follows: defining a left sliding window and a right sliding window which cover a preset number of sampling points along a route by taking the current sampling point as a center; respectively acquiring a path difference vector set falling into a left sliding window and a path difference vector set falling into a right sliding window; Respectively calculating a first median vector of the path difference vector set in the left sliding window and a second median vector of the path difference vector set in the right sliding window; and calculating the Euclidean norm of the difference vector between the second median vector and the first median vector to obtain step strength, wherein the step strength reflects the intensity of discontinuous mutation of the spatial position deviation between the GIS planning route and the VGIS precise marking route at the current position.
4. The AR navigation method based on GIS-VGIS fusion according to claim 3, wherein the specific process of extracting candidate step boundary bands includes: Traversing step intensity sequences distributed along the path arc length, and identifying all local maximum value points; Taking each local maximum point as a central position of the potential mutation, and searching the forward direction and the backward direction of the route by taking the central position as a starting point; stopping searching when the searched step intensity value is attenuated to the background neutral level of the whole road section, and respectively determining the initial arc length and the end arc length corresponding to the searching stop position; and determining a continuous interval from the starting arc length to the ending arc length as a candidate step boundary zone, wherein the candidate step boundary zone is a transition zone which is covered with the position deviation between the GIS planning route and the VGIS fine-scale route in space and is subjected to jitter or jump.
5. The AR navigation method based on GIS-VGIS fusion according to claim 4, comprising: projecting each terminal position estimation value onto VGIS fine-marked routes to obtain corresponding route arc lengths, establishing a one-to-one mapping relation between the calculated ranging residual fingerprint vectors and the route arc lengths, and collecting all the mapping relations into a spatial distribution data set ordered according to the arc lengths; taking a region which extends a first preset length in the opposite direction of the route by taking the starting position of the candidate step boundary zone as an end point as a left neighborhood, taking a region which extends a second preset length in the forward direction of the route by taking the ending position of the candidate step boundary zone as a starting point as a right neighborhood, respectively extracting ranging residual fingerprint vectors falling into the left neighborhood and the right neighborhood from a spatial distribution data set, and obtaining a left neighborhood fingerprint set and a right neighborhood fingerprint set; And taking Euclidean norms of the difference value vectors obtained by subtracting the median vector of the left neighborhood fingerprint set from the median vector of the right neighborhood fingerprint set as fingerprint transition amounts, wherein the fingerprint transition amounts are used for quantifying the mutation degree of signal propagation environments at two sides of the candidate step boundary band.
6. The AR navigation method based on GIS-VGIS fusion according to claim 5, comprising: acquiring fingerprint transition amounts corresponding to all candidate step boundary bands, and performing descending order arrangement according to the numerical values to generate a transition amount ordering sequence; calculating the differential value between adjacent elements in the transition quantity sequencing sequence, constructing a first-order differential sequence of the transition quantity, identifying the position with the largest value change rate in the first-order differential sequence, marking as an elbow point, And judging that the real signal shielding mutation exists in candidate step boundary bands corresponding to all the fingerprint transition amounts positioned at and before the elbow point in the transition amount sequencing sequence, marking the real signal shielding mutation as valid, marking all the candidate step boundary bands marked as valid as key step boundary bands, wherein the key step boundary bands reflect environment demarcation points causing systematic mutation of terminal positioning error distribution characteristics.
7. The AR navigation method based on GIS-VGIS fusion according to claim 6, comprising: Respectively marking a route interval between two adjacent key step boundary bands, a route interval between a route start point and a first key step boundary band and a route interval between a last key step boundary band and a route end point as stable segments with uniform system error characteristics; Obtaining terminal position estimated values of all projection positions falling in the range of the current stable segment, calculating a difference vector obtained by subtracting the corresponding VGIS fine-scale route point coordinate vector from the terminal position estimated value vector, marking the difference vector as a displacement sample, Marking the median vector of all displacement sample vector sets in one stable segment as independent translation quantity, wherein the independent translation quantity represents average systematic positioning error caused by specific environmental shielding or multipath effect in the area where the stable segment is positioned; and adding the arc length coordinate of the VGIS precise marking route to the independent translation quantity of the stable section to obtain the sectional alignment route of the stable section.
8. The AR navigation method based on GIS-VGIS fusion according to claim 7, comprising: For each key step boundary band, determining a left segment alignment route corresponding to a stable segment immediately preceding the key step boundary band and a right segment alignment route corresponding to a stable segment immediately following the key step boundary band; Acquiring the arc length position of a current calculation point in a key step boundary zone, calculating the ratio of the distance between the arc length position and the initial arc length of the key step boundary zone to the total length of the key step boundary zone, determining the ratio as a right weight coefficient, and determining the difference value of subtracting the right weight coefficient from 1 as a left weight coefficient; Weighting the extension coordinates of the left segment alignment route at the current position by using the left weight coefficient, weighting the extension coordinates of the right segment alignment route at the current position by using the right weight coefficient, and adding the weighted results of the two to obtain transition route coordinates; And constructing a virtual guide curve which is continuous in a three-dimensional space, adopting coordinates of a corresponding segment alignment route in each stable segment, adopting transition route coordinates in each key step boundary zone, and rendering continuous navigation guide icons in the augmented reality view according to the virtual guide curve.
9. The AR navigation method based on GIS-VGIS fusion according to claim 3, wherein the specific process of arc length parameterized resampling includes: Projecting a GIS planning route and a VGIS fine marking route to the same engineering coordinate system by utilizing a road network topological relation, and establishing a unified reference frame for space geometric comparison of the GIS planning route and the VGIS fine marking route; calculating the accumulated arc length along the path by taking the common starting point of the GIS planning path and the VGIS precise standard path as a reference, and constructing respective arc length parameter models; And setting uniform sampling intervals, and carrying out equidistant resampling on the two routes by using the arc length parameter model to generate a GIS resampling coordinate set and a VGIS resampling coordinate set which have the same arc length index sequence.
10. An AR navigation system based on GIS and VGIS fusion, for implementing an AR navigation method based on GIS and VGIS fusion according to any one of claims 1 to 9, comprising: The positioning fingerprint module is used for acquiring a reference point coordinate set and a terminal ranging observation value, calculating a terminal position estimation value by using a least square method, calculating an observation residual component and constructing a ranging residual fingerprint vector representing the current signal propagation environment; the candidate boundary module is used for acquiring and aligning a GIS planning route and a VGIS fine-marking route, calculating a path difference vector of the GIS planning route and the VGIS fine-marking route at the same arc length, calculating step intensity by utilizing a sliding window, and extracting a candidate step boundary band set based on the step intensity; the key boundary module is used for executing the space mapping from the terminal position sequence to the route arc length, establishing a space distribution data set, respectively acquiring neighborhood fingerprint sets at two sides of a candidate step boundary zone, calculating fingerprint transition quantity, screening and confirming a key step boundary zone; The segmentation alignment module is used for dividing the VGIS fine-scale route into stable segments through a key step boundary band, counting the difference value between the estimated value of the terminal position falling into each stable segment and the route point to generate a displacement sample, obtaining independent translation quantity and generating a segmentation alignment route; And the splicing navigation module is used for calculating linear weighted transition route coordinates based on the left segment alignment route and the right segment alignment route in the key step boundary zone, generating a final continuous fusion route based on the segment alignment route and the transition route, and carrying out AR navigation.

Description

AR navigation method and system based on GIS and VGIS fusion Technical Field The invention relates to the technical field of radio navigation, in particular to an AR navigation method and system based on GIS and VGIS fusion. Background In complex working environments such as building construction sites, operators wearing the augmented reality intelligent terminal not only need to acquire real-time path navigation and working instructions in a complicated space, but also need to rush to potential safety hazard points or accident sites in emergency when encountering emergency. The construction site has extremely high space complexity and dynamic time variability, steel structure scaffolds are often densely distributed on the site, and metal enclosure which frequently moves along with the progress of construction period, temporarily stopped large mechanical equipment and building materials with unfixed stacking positions exist. The unstructured dynamic environment not only requires the navigation system to have the road network planning capability of a macroscopic level, but also requires the navigation system to accurately spatially fuse the virtual three-dimensional navigation guidance icons with a real road surface so as to ensure the passing efficiency and personal safety of operators when crossing different construction areas or crossing narrow channels. However, existing augmented reality navigation techniques face the core technical difficulty in the above scenario that the environment perception and virtual map data are difficult to match in real time. The evolution speed of the physical environment of the construction site is often faster than the update period of the high-precision map data, so that the geometrical path of the pre-planned geographic information system and the current practical feasible traffic space have mismatch of space-time dimension. In addition, a newly added metal fence or large steel structure facility on a construction site can seriously interfere the propagation of radio range signals, and cause complex non-line-of-sight propagation phenomenon and multipath effect, so that positioning data received by a terminal shows violent jump and drift in non-Gaussian distribution. Existing positioning navigation techniques often have difficulty in effectively distinguishing whether a sudden change in positioning coordinates is false jitter caused by random signal environmental noise or systematic steps caused by real obstruction occlusion when processing such data. The lack of the boundary perception capability directly leads the navigation system to be unable to make correct avoidance or disconnection treatment in time when facing the actual physical obstacle on site, thereby generating visual deviation such as direct penetration of the navigation guide line through a wall or severe shaking at the edge of the obstacle, seriously reducing the reliability of navigation guide and possibly misleading operators to enter a dangerous area. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an AR navigation method and an AR navigation system based on GIS and VGIS fusion, which solve the problems that in a complex construction site environment with dynamic change, the existing navigation method is difficult to effectively distinguish whether mutation of positioning data is caused by random signal noise or real obstacle shielding, so that the augmented reality navigation guidance cannot sense boundaries, wall penetration or false drift occurs, and accurate spatial fusion of virtual information and real pavement cannot be realized. Is a problem of (a). In order to achieve the above purpose, the invention is realized by the following technical scheme: calculating a terminal position estimation value based on the reference point coordinate set and the terminal ranging observation value, calculating an observation residual component between the terminal ranging observation value and a theoretical distance determined according to the terminal position estimation value, and constructing a ranging residual fingerprint vector reflecting a signal propagation environment; calculating a path difference vector of the GIS planning path and the VGIS fine-scale path at the same arc length, counting the change characteristics of the path difference vector to obtain step intensity, and extracting candidate step boundary bands based on the distribution characteristics of the step intensity; Establishing a space distribution data set of distance measurement residual fingerprint vectors along the arc length of a route, respectively obtaining distance measurement residual fingerprint vector sets of a left neighborhood and a right neighborhood of a candidate step boundary zone, calculating the statistical difference between the two to obtain a fingerprint transition quantity, and screening the candidate step boundary zone based on the fingerprint transition quantity to obtain a ke