A multi-condition constraint-based electromagnetic tracking trajectory determination method and system
Patent Information
- Application Number
- CN202610990472.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-03
- Publication Date
- 2026-09-29
AI Technical Summary
半球歧义会导致定位漂移，尤其在接收端Rx靠近对称平面（X≈0或Z≈0）时，基于坐标符号的硬约束会因测量噪声产生高频抖动
[0018]通过融合空间几何初始化筛选、时序连续性约束、指间拓扑校验及姿态同步修正的多条件协同判定机制，实现电磁追踪半球镜像歧义的全程抑制与自主修正，全方位提升手部动作捕捉轨迹输出质量与设备适配性。本发明依托时序约束结合速度限幅校验，基于历史轨迹外推筛选候选解，摒弃坐标绝对值判据，从而消除坐标边界定位高频抖动；该判定流程逐帧持续运行，可实时完成轨迹纠错，摆脱初始化单次判定的局限，具备长时稳定连续追踪能力。通过多接收端指间距离拓扑约束实现多传感器协同校验，以大拇指接收端作为固定锚点识别并修正单点定位异常，强化多通道采集协同一致性。整套判定逻辑仅采用基础算术运算，能够在MCU、FPGA嵌入式处理单元实时运行，同步兼顾动捕定位精度、长时运行稳定性与硬件部署通用性。
Smart Images
Figure CN122837632A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electromagnetic positioning technology, and in particular to a method and system for determining electromagnetic tracking trajectory based on multiple constraints. Background Technology
[0002] In a magnetic field approximated by a single coil as a magnetic dipole, the magnetic field distribution at any point r in space is as follows: Where m is the magnetic moment, It is the direction The unit vector on the surface. Since there exists an inversion point −r such that B(−r)=B(r), the known magnetic field strength and magnetic moment cannot guarantee the uniqueness of the position solution, that is, the phenomenon of "hemispherical ambiguity" occurs. The magnetic field strength of the true position PA=(x,y,z) is exactly the same as that of the mirror position PB=(−x,−y,−z).
[0003] In a hand electromagnetic motion capture system, a transmitting coil is mounted on the back of the hand, and multiple receiving coils are mounted on each fingertip. Hemispherical ambiguity can cause positioning drift, especially when the receiver's Rx is close to the plane of symmetry (X≈0 or Z≈0). Hard constraints based on coordinate signs can produce high-frequency jitter due to measurement noise.
[0004] Existing solutions (such as Polhemus and Magic Leap) rely on auxiliary sensors (IMUs) to perform one-time disambiguation during system startup, which has three drawbacks: 1) It only solves the initial hemisphere determination, and may still flip when it is close to the plane of symmetry during operation; 2) It processes each sensor independently and does not take advantage of the spatial topological constraints between multiple receivers Rx; 3) It has poor stability in the boundary region and is prone to jitter when the coordinates are close to zero, resulting in unstable position output. Summary of the Invention
[0005] To overcome the shortcomings of existing technologies, this invention provides a method and system for determining electromagnetic tracking trajectories based on multiple condition constraints.
[0006] To achieve the above objectives, the present invention adopts the following technical solution:
[0007] An electromagnetic tracking trajectory determination method based on multiple constraints is applied to an electromagnetic tracking system comprising a transmitter Tx and N receivers Rx. The transmitter Tx includes orthogonal triaxial transmitting coils fixed at the origin of the coordinate system, and each receiver Rx includes an orthogonal triaxial receiving coil. The method includes the following steps:
[0008] S1. By obtaining the channel matrix S of each receiver based on the electromagnetic coupling between the transmitter Tx and the receiver Rx, and performing a nonlinear inverse solution on the channel matrix S based on the magnetic dipole model, two candidate position solutions for each receiver Rx are obtained. , ;
[0009] S2. During the initialization phase, solve the two candidate positions of Rx at each receiver in the first K frames. , Perform initialization determination, obtain the initial correct solution for each receiver Rx, and establish a reference distance matrix between each receiver Rx. , where K is a preset positive integer;
[0010] S3. Tracking Phase: Starting from frame K+1, a time continuity constraint is applied to each receiver's Rx. The prediction velocity is estimated based on the historical output position and extrapolated at a uniform speed to obtain the predicted position of the current frame. Select the prediction position from the current frame. The closer candidate solution is selected based on temporal continuity; if the implicit velocity of the selected result exceeds a preset threshold, the prediction position of the current frame is retained. No updates;
[0011] S4. Perform topological detection on the finger-to-finger distance and calculate the current distance matrix of the current frame selection result. According to the current distance matrix Distance matrix with reference ratio For the marker For the abnormal receiver Rx, construct the candidate flip positions for the abnormal receiver Rx, and correct the position if the total distance deviation is smaller after flipping.
[0012] S5. Quaternion synchronization: When the position of the receiver Rx is corrected, the corresponding attitude quaternion is synchronously inverted to eliminate hemispherical ambiguity.
[0013] An electromagnetic tracking system based on multiple constraints includes:
[0014] The transmitting module (transmitter Tx) is installed on the back of the hand as the origin of the coordinate system. It contains orthogonal triaxial transmitting coils to generate a near-field excitation magnetic field.
[0015] The receiving module (i.e., N receiving ends Rx) is installed on each fingertip. Each receiving module contains an orthogonal triaxial receiving coil, which is used to obtain the channel matrix S of the electromagnetic coupling strength between the transmitting module and the receiving module.
[0016] The processing module is connected to the transmitting module and the receiving module respectively, and is used to execute the electromagnetic tracking trajectory determination method based on multiple condition constraints as described in any one of claims 1 to 8, so as to eliminate hemispherical ambiguity.
[0017] The present invention provides a method and system for determining electromagnetic tracking trajectory based on multiple constraints, the advantages of which are as follows:
[0018] By integrating spatial geometric initialization filtering, temporal continuity constraints, finger-to-finger topology verification, and attitude synchronization correction into a multi-condition collaborative judgment mechanism, this invention achieves full-process suppression and autonomous correction of electromagnetic tracking hemispherical mirror ambiguity, comprehensively improving the quality of hand motion capture trajectory output and device adaptability. This invention relies on temporal constraints combined with velocity limiting verification, extrapolating candidate solutions based on historical trajectories, and abandoning absolute coordinate value criteria, thereby eliminating high-frequency jitter in coordinate boundary positioning. This judgment process runs continuously frame-by-frame, enabling real-time trajectory correction, overcoming the limitations of single initialization judgments, and possessing long-term stable continuous tracking capabilities. Multi-sensor collaborative verification is achieved through multi-receiver finger-to-finger distance topology constraints, using the thumb receiver as a fixed anchor point to identify and correct single-point positioning anomalies, strengthening the consistency of multi-channel acquisition. The entire judgment logic uses only basic arithmetic operations, enabling real-time operation on MCU and FPGA embedded processing units, simultaneously considering motion capture positioning accuracy, long-term operational stability, and hardware deployment versatility. Attached Figure Description
[0019] Figure 1 This is a flowchart of the electromagnetic tracking trajectory determination method with multiple constraints in an embodiment of the present invention;
[0020] Figure 2 This is a schematic diagram of the electromagnetic tracking system structure in an embodiment of the present invention;
[0021] Figure 3 This is a schematic diagram illustrating the ambiguity of the magnetic dipole hemisphere in an embodiment of the present invention. Detailed Implementation
[0022] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.
[0023] like Figure 2 As shown, this invention provides an electromagnetic tracking system based on multi-condition constraints, including a transmitting module, N receiving modules, and a processing module. The system employs an electromagnetic tracking architecture with one transmitting end Tx and five receiving ends Rx. Taking hand tracking as an example, N=5. The transmitting module includes orthogonal triaxial transmitting coils, which can independently output alternating excitation magnetic fields along the three spatial axes, constructing a uniform near-field excitation magnetic field environment covering the entire finger's active area. The transmitting end Tx serves as the origin of the coordinate system. The five receiving ends Rx are respectively installed at the tips of the thumb, index finger, middle finger, ring finger, and little finger, each containing an orthogonal triaxial receiving coil, capable of synchronously sensing the coupled voltage signal under the triaxial excitation magnetic field, realizing multi-dimensional magnetic field coupling information acquisition.
[0024] In this embodiment, the coordinate system is defined as a right-handed coordinate system, where the X-axis is along the length of the middle finger, with all fingertips always X > 0; the Y-axis is along the horizontal direction of the back of the hand, adapting to the left and right offset movements of finger opening and closing, and can be positive or negative; the Z-axis is perpendicular to the back of the hand and points towards the palm, with all fingertips always in the positive Z-axis space under normal hand posture, and all fingertips always Z > 0. This spatial geometric constraint characteristic provides stable prior constraints for candidate solution screening in the initialization stage and topology anomaly discrimination in the tracking stage. A near-field excitation magnetic field is generated by an orthogonal triaxial transmitting coil and electromagnetically coupled to an orthogonal triaxial receiving coil.
[0025] Specifically, alternating excitation current is sequentially applied to the orthogonal triaxial coils of the transmitter Tx; during each axis excitation of the transmitter Tx, each receiver Rx synchronously samples the induced voltage of its own orthogonal triaxial receiving coil, and after gain normalization, each receiver Rx obtains a 3×3 channel matrix S, providing raw observation data for subsequent position inverse solution.
[0026] like Figure 1 As shown, this embodiment provides an electromagnetic tracking trajectory determination method based on multiple constraints, applied to an electromagnetic tracking system comprising a transmitter Tx and N receivers Rx. The transmitter Tx comprises an orthogonal triaxial transmitting coil fixed at the origin of the coordinate system, and each receiver Rx comprises an orthogonal triaxial receiving coil. The method includes the following steps S1 to S5:
[0027] S1. By obtaining the channel matrix S of each receiver based on the electromagnetic coupling between the transmitter Tx and the receiver Rx, and performing a nonlinear inverse solution on the channel matrix S based on the magnetic dipole model, two candidate position solutions for each receiver Rx are obtained. , Specifically, refer to Figure 3 According to the magnetic dipole model, the magnetic field strength at any spatial location P=(x,y,z) satisfies a nonlinear mapping relationship with the location. Due to the central symmetry of the magnetic field distribution, i.e., B(P)=B(−P), the same set of channel matrices S corresponds to two symmetric spatial solutions. A nonlinear least squares algorithm is used to solve for the channel matrix S. Specifically, using the channel matrix S as the observed value, the position parameter P is solved to minimize the error between the theoretical coupling matrix and the measured matrix; the algorithm converges to obtain the first candidate position solution. Based on central symmetry, the second candidate location solution is directly derived from... get.
[0028] S2. In the initialization phase, the two candidate position solutions of each receiver Rx in the first K frames are initialized and determined to obtain the initial correct solution of each receiver Rx, where K is a preset positive integer.
[0029] Specifically, during the initialization phase, let the t-th frame be such that t≤K, and the two hemispherical candidate positions of the i-th receiver Rx are as follows: Since each fingertip in the coordinate system with the transmitter Tx as the origin should satisfy X>0 and Z>0 in the initial state, this invention uses an indicator function. Perform hard-constraint scoring: if there are two hemispherical candidate positions , If X>0 and Z>0, the candidate with the larger I value is selected as the initial correct solution. This stage is executed only once during initialization, providing an initial reference position for topology detection in the tracking stage. In hand motion capture scenarios, the movement range of the thumb tip is relatively stable and far from the plane of symmetry. A high-confidence initial correct solution can be obtained through the aforementioned geometric hard constraints. Therefore, in the subsequent tracking stage, the thumb receiver Rx can be used as a stable anchor point for inter-finger distance topology detection, without participating in flip correction, to improve the overall stability of the system.
[0030] S3. Entering the tracking phase, starting from frame K+1, a temporal continuity constraint is applied to each receiver Rx, constructing a temporal constraint mechanism based on the physical continuity characteristics of hand movements. The real-time motion velocity is calculated using the historical steady-state output positions of the previous two frames, and the current frame position is extrapolated and predicted using a uniform motion model to obtain the predicted position of the current frame.
[0031] Specifically, for each receiver Rx, first determine the position of the (t-2)th frame. With the position of frame t-1 For speed of movement Perform calculations and set ,in, , This is the output position for historical frames. This is the time interval between historical sampling frames. Based on this speed... Perform uniform extrapolation to obtain the predicted position of the current frame. Predict the position based on the current frame. Calculate the Euclidean distances from the two candidate positions to the predicted position in the current frame, respectively: , Based on the Euclidean distance, a candidate solution that is closer to the predicted position is selected, and the candidate solution satisfies... ,like ≤ Then select As a result of the choice based on time continuity; if > Then select As a result of the choice based on temporal continuity.
[0032] Meanwhile, to avoid abnormal jumps, this embodiment also uses the temporal continuity to select the implicit velocity of the result relative to the output of the previous frame. With respect to the preset maximum reasonable motion speed threshold Perform a comparison and predict the position of the current frame based on the comparison results. The risk of abnormal jumps is assessed, and the candidate solutions for the predicted position are updated based on the assessment results. The preset maximum reasonable movement speed is set according to the physical movement limit of the receiver Rx. Here is the speed limiting factor, where:
[0033] when At that time, determine the prediction position of the current frame. If there is a risk of abnormal jumps, the predicted position will not be updated;
[0034] when At that time, determine the prediction position of the current frame. If there is no risk of abnormal jumps, then update to the predicted position of the current frame. .
[0035] S4. Based on the initial correct solutions of each receiver Rx, establish pairwise distance reference matrices between N receiver Rx. The current distance matrix of each receiver Rx is calculated based on the candidate solutions of the current frame. Through the current distance matrix Distance matrix with reference ratio The candidate position status of the receiver Rx is marked.
[0036] Specifically, a pairwise distance reference matrix is set. In each motion frame, calculate the current distance matrix based on the selected position. ,set up ,in, Thus, the current distance matrix is calculated. Distance matrix with reference ratio ,Right now The current distance matrix Distance matrix with reference ratio Distance anomaly detection threshold Compare them, and based on the comparison results, determine the first... The state of the first receiver Rx is determined, and the first receiver is adjusted accordingly. Each receiver Rx is marked, where:
[0037] when When, then determine the first If the status of each receiver Rx is normal, no marking is needed.
[0038] when When, then determine the first The status of the first receiver Rx is suspected of being abnormal; at this time, the status of the first receiver Rx is... Each receiver's Rx is marked as potentially abnormal.
[0039] Specifically, for receivers Rx marked as potentially abnormal, a flip candidate position is constructed. Calculate the first one respectively The total distance deviation of each receiver Rx before flipping relative to other receiver Rx and the total distance deviation after flipping are set. ,in For the first flip Candidate positions for each receiver Rx, Flip before Candidate positions for each receiver Rx; total distance deviation after flipping. The total distance deviation before and after the flip is compared. Based on the comparison result, candidate positions of abnormal receiver Rx are determined, and corrections are made based on the determination results.
[0040] like Then the candidate position of the abnormal receiver Rx will be changed from Revised to ;
[0041] like If the current frame data of the receiver's Rx is severely distorted and remains marked as abnormal, the system will discard the magnetic positioning result of that frame and directly use the predicted position. As the final output, this ensures the smoothness of the trajectory.
[0042] In this embodiment, the distance anomaly determination threshold The preferred setting is 1.5. Testing has shown that when... When the value is in the range of 1.2 to 2.0, the system can achieve the best distinction between normal finger bending and hemispherical flipping, which can detect abnormalities in a timely manner and avoid misjudgment caused by measurement noise.
[0043] S5. Quaternion Synchronization: When the position of the receiver Rx is corrected, its corresponding attitude quaternion is synchronously inverted to eliminate hemispherical ambiguity. In one embodiment, when the position output receiver Rx is corrected from P to -P, the corresponding attitude quaternion is synchronously inverted, i.e., from q to −q. Specifically, based on the pose mirroring characteristic of the magnetic dipole model: the pose solution pair (P,q) and (−P,−q) correspond to the same magnetic field distribution, so when the candidate position is flipped, the attitude quaternion is ensured to remain on the mirror solution, thereby eliminating hemispherical ambiguity.
[0044] To verify the effectiveness of the method of the present invention, a prototype electromagnetic motion capture glove system conforming to the content of the present invention was used for testing. One transmitter Tx was installed on the back of the hand, and five triaxial orthogonal receivers Rx were installed on the fingertips of the thumb, index finger, middle finger, ring finger, and little finger, respectively. Each Rx continuously acquired 1000 frames, obtaining a total of 5×1000 frames of data.
[0045] Table 1 lists the statistical results of the spatial distribution of the trajectories of the five receivers' Rx.
[0046] RX section X range (mm) Y range (mm) Z range (mm) Cumulative path (mm) X < 0 frames |X| < 1cm frame count |Z| < 1cm frame count Thumb 22.5~73.5 -7.9~103.5 38.9~124.8 2091 0 0 0 index finger -4.1~137.1 12.3~49.0 2.1~120.8 3402 11 18 9 middle finger -19.4~152.5 -13.5~12.9 9.1~133.6 3943 20 5 3 ring finger -19.8~143.0 -35.3~0.0 33.2~133.0 3919 25 20 0 Little finger -7.5~115.6 -71.6~-1.8 38.6~119.6 3500 29 53 0
[0047] The trajectory of the thumb is generally far from the plane of symmetry. ), by initializing geometric hard constraints ( This can stably disambiguate the problem. However, the other four fingers showed frames with X<0 and boundary frames with |X|<10mm during the bending motion. These frames are areas where traditional disambiguation schemes are prone to high-frequency flipping jitter.
[0048] Table 2 lists a comparison of the actual inter-frame displacement (Euclidean distance between the positions of two consecutive frames) of each receiver's Rx with the order of magnitude of the "position jump caused by hemispherical flip" (hemispherical flip jump is defined as the correct solution). Rather than inversion solution The distance between them, i.e. ).
[0049] Table 2 compares the actual inter-frame displacement with the hemispherical flip jump scale (unit: mm).
[0050] Part Inter-frame shift - mean Inter-frame shift - maximum Flip Jump - Minimum Flip Jump - Mean Jump / displacement ratio (minimum) Thumb 2.09 14.47 185.4 242.8 Approximately 13 times index finger 3.41 35.75 192.7 269.0 Approximately 5 times middle finger 3.95 46.32 172.9 278.2 Approximately 4 times ring finger 3.92 41.56 168.0 263.5 Approximately 4 times Little finger 3.50 32.31 137.1 237.1 Approximately 4 times
[0051] In any frame, the distance between the correct solution and the mirror inversion solution is at least 4 times higher than the actual inter-frame displacement, and the correct hemisphere can be reliably distinguished based on the time continuity constraint.
[0052] Boundary crossing scenario verification
[0053] Taking frames 913-919 of the index finger as an example. The position of frame 913 is (10.60, 13.20, 107.80) mm. The X coordinate decreases frame by frame, and the span of frame 919 is -0.10 mm.
[0054] If the traditional "X > 0" hard constraint is used, the 919th frame will trigger a hard flip due to X < 0, resulting in a non-physical jump of about 200mm. Under the present invention, the 919th frame can still output (-0.10, 14.10, 98.70)mm according to the predicted position, and the position and attitude are smooth and continuous, eliminating the high-frequency flip jitter in the boundary area.
[0055] The above is a description of an electromagnetic tracking trajectory determination method and system based on multi-condition constraints according to the present invention, which is used to help understand the present invention; however, the implementation of the present invention is not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the principle of the present invention shall be considered as equivalent substitutions and are included within the protection scope of the present invention.
Claims
1. A method for determining electromagnetic tracking trajectory based on multiple condition constraints, characterized in that, An electromagnetic tracking system comprising a transmitter Tx and N receivers Rx, wherein the transmitter Tx includes orthogonal triaxial transmitting coils fixed at the origin of the coordinate system, and each receiver Rx includes an orthogonal triaxial receiving coil, the method comprising the following steps: S1. By obtaining the channel matrix S of each receiver based on the electromagnetic coupling between the transmitter Tx and the receiver Rx, and performing a nonlinear inverse solution on the channel matrix S based on the magnetic dipole model, two candidate position solutions for each receiver Rx are obtained. , ; S2. During the initialization phase, solve the two candidate positions of Rx at each receiver in the first K frames. , Perform initialization determination, obtain the initial correct solution for each receiver Rx, and establish a reference distance matrix between each receiver Rx. , where K is a preset positive integer; S3. Tracking Phase: Starting from frame K+1, a time continuity constraint is applied to each receiver's Rx. The prediction velocity is estimated based on the historical output position and extrapolated at a uniform speed to obtain the predicted position of the current frame. Select the prediction position from the current frame. The closer candidate solution is selected based on temporal continuity; if the implicit velocity of the selected result exceeds a preset threshold, the prediction position of the current frame is retained. No updates; S4. Perform topological detection on the finger-to-finger distance and calculate the current distance matrix of the current frame selection result. According to the current distance matrix Distance matrix with reference ratio For the marker For the abnormal receiver Rx, construct the candidate flip positions for the abnormal receiver Rx, and correct the position if the total distance deviation is smaller after flipping. S5. Quaternion synchronization: When the position of the receiver Rx is corrected, the corresponding attitude quaternion is synchronously inverted to eliminate hemispherical ambiguity.
2. The electromagnetic tracking trajectory determination method according to claim 1, characterized in that, In step S1, the nonlinear inverse solution of the channel matrix S based on the magnetic dipole model specifically includes: using a nonlinear least squares algorithm to solve for the position parameters, minimizing the error between the theoretical coupling matrix and the measured channel matrix S; and obtaining the first candidate position solution after convergence. Based on the central symmetry of the magnetic dipole model, the second candidate position solution is directly obtained. .
3. The electromagnetic tracking trajectory determination method according to claim 1, characterized in that, In step S2, the initialization determination is based on a right-handed coordinate system with the transmitter Tx as the origin and the spatial geometric hard constraints of the hand motion capture scene; wherein, the five receivers Rx are respectively installed at the tips of the thumb, index finger, middle finger, ring finger, and little finger; the right-handed coordinate system satisfies: the X-axis is along the direction of the middle finger, the Y-axis is along the horizontal direction of the back of the hand, and the Z-axis is perpendicular to the back of the hand and points towards the palm; the spatial geometric hard constraints include two candidate positions for the receivers Rx. , Using indicator functions Perform constraint scoring and select The higher-scoring solution is used as the initial correct solution; the solution is constructed based on the initial correct position. Pairwise distance reference matrix between each receiver Rx .
4. The electromagnetic tracking trajectory determination method according to claim 1, characterized in that, In step S3, the temporal continuity constraint includes: based on the position of the (t-2)th frame. Position of frame t-1 Calculate the speed of motion ,set up ,in The time interval between sampling frames; based on the motion speed The predicted position of the current frame is obtained by performing uniform extrapolation. ; calculate the solutions for the two candidate positions respectively. , To the current frame prediction position The Euclidean distances are set as follows: ;like ≤ Then select As a result of time continuity selection; otherwise, select .
5. The electromagnetic tracking trajectory determination method according to claim 1, characterized in that, In step S3, the velocity limiting verification includes: calculating the implicit velocity calculated from the temporal continuity selection result relative to the output position of the previous frame. ;like If it is not a physical jump, then it is determined to be a non-physical jump, and the predicted position of the current frame is maintained. No updates; among them, To preset the maximum reasonable speed of movement, This is the speed limiting factor.
6. The electromagnetic tracking trajectory determination method according to claim 1, characterized in that, In step S4, the finger-to-finger distance topology detection includes: calculating the current distance matrix based on the current frame selection result. ; Calculate the ratio When the i-th receiver Rx exists Make > If it is, then it is marked as a suspected anomaly; among them, This is the threshold for determining distance anomalies.
7. The electromagnetic tracking trajectory determination method according to claim 6, characterized in that, In step S4, for the receiver Rx marked as a suspected abnormal receiver, a flip candidate position is constructed. ; Calculate the total distance deviation before flipping respectively and total distance deviation after flipping ;like Then the position of the receiver Rx will be corrected to In hand motion capture scenarios, the thumb receiver Rx is not marked as a suspected anomalous receiver Rx during the tracking phase, serving as a stable anchor point for topology anomaly detection. If multiple consecutive frames of the receiver's Rx are marked as suspected anomalies and no flip correction is triggered, the predicted position is based on the current frame. Fill in the output.
8. The electromagnetic tracking trajectory determination method according to claim 1, characterized in that, In step S5, based on the pose mirroring characteristics of the magnetic dipole model, the pose solution pair (P,q) and (−P,−q) correspond to the same magnetic field distribution. When the candidate position is corrected from P to −P, the pose quaternion q is synchronously inverted to −q, so that the pose solution pair remains on the same set of mirror solutions.
9. An electromagnetic tracking system based on multiple constraints, characterized in that, include: The transmitting module (transmitter Tx) is installed on the back of the hand as the origin of the coordinate system. It contains orthogonal triaxial transmitting coils to generate a near-field excitation magnetic field. The receiving module (i.e., N receiving ends Rx) is installed on each fingertip. Each receiving module contains an orthogonal triaxial receiving coil, which is used to obtain the channel matrix S of the electromagnetic coupling strength between the transmitting module and the receiving module. The processing module is connected to the transmitting module and the receiving module respectively, and is used to execute the electromagnetic tracking trajectory determination method based on multiple condition constraints as described in any one of claims 1 to 8, so as to eliminate hemispherical ambiguity.
10. The system according to claim 9, characterized in that, The system is an electromagnetic hand motion capture glove. The receiving modules N=5, which are respectively installed on the fingertips of the thumb, index finger, middle finger, ring finger and little finger. The processing module completes the establishment of the initial correct solution in the first K frames of the initialization phase.