A method for controlling coupling of optical paths of a fiber micro-nano multi-station based on light field optimization

Abstract

The present application relates to the technical field of micro-electro-mechanical system, specifically to a kind of optical fiber micro-nano multi-station optical path coupling control method based on light field optimization, including micro-electro-mechanical multi-axis displacement system executes translation scanning, collects the light field energy vector of reference light channel and extracts three-dimensional extreme value coordinates;It is mapped as the virtual rotation center of light beam and posture angle deflection scanning is carried out, collects edge light channel signal to construct space posture point set and generates multidimensional posture envelope area;The geometric center of the envelope area is calculated as the target point posture of optical axis alignment is deflected, and the translation and deflection actions are executed in a loop until the spatial and angular displacement are less than the preset convergence threshold value.The present application obtains the candidate coupling pose that satisfies the reference channel candidate extreme value condition and the edge channel threshold condition by the translation search of reference channel, the posture threshold screening of edge channel and the alternative adjustment of translation and deflection, reduces the mutual influence between translation adjustment and deflection adjustment.

Description

Technical Field

[0001] This invention relates to the field of microelectromechanical systems (MEMS) technology, specifically to a fiber optic micro / nano multi-station optical path coupling control method based on optical field optimization. Background Technology

[0002] With the rapid development of optical communication, integrated optoelectronics technology, and silicon-based photonics, optoelectronic chips are evolving towards high integration, miniaturization, and arraying. In the packaging and testing of optoelectronic devices, optical path coupling between optical fibers and optical waveguide devices (such as micro / nano chips) is a crucial step. To meet the high-capacity transmission demands of data centers and other applications, optical devices commonly employ multi-channel array structures. This requires the simultaneous realization of optical path connections at multiple stations (multi-ports) within an extremely small space with mode field sizes ranging from sub-micrometer to nanometer scale. In multi-station optical path coupling at the micro-nano scale of optical fibers, due to the rigid spatial geometric strong correlation between multiple stations, coupled with the extremely sensitive characteristics of the micro-nano optical field to the six degrees of freedom of space (especially small angular deflection), the traditional single-point sequential alignment algorithm produces dynamic misalignment oscillations during optimization, and cannot synchronously plan the global pose that allows all stations to achieve the optimal coupling efficiency in the extremely complex spatial optical field distribution. Specifically, at the micro-nano scale of optical fibers, the beam pattern diameter is extremely small, making the optical path extremely demanding on spatial positioning. Any submicron-level translation or even a fraction of a degree of deflection will cause a sharp attenuation of the optical signal. In multi-station coupling scenarios (such as the docking of fiber arrays and multi-channel chips), multiple coupling end faces are rigidly connected in space, exhibiting strong attitude linkage. When the control system drives a multi-axis precision displacement mechanism to try to adjust and capture the maximum optical power (peak optical field) of station A, it inevitably introduces a slight angular wobble. This slight attitude change, amplified by the spatial distance between the multiple stations, will instantly disrupt the optical field alignment of stations B and C, causing huge insertion loss. Existing control systems, lacking the ability to comprehensively analyze the global spatial optical field distribution gradient, are prone to getting stuck in a dead loop in the local optimum of multiple peak values.

[0003] To address this, a fiber micro / nano multi-station optical path coupling control method based on optical field optimization is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a fiber micro / nano multi-station optical path coupling control method based on optical field optimization, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A fiber micro / nano multi-station optical path coupling control method based on optical field optimization includes:

[0006] The microelectromechanical multi-axis displacement system is controlled to drive the fiber micro-nano multi-station array to perform translational scanning relative to the docking target device, collect the light field energy vector of the beam emitted from the reference optical channel station, and extract the three-dimensional extreme coordinates of the corresponding beam mode field energy peak. The physical rotation center of the microelectromechanical multi-axis displacement system is mapped to the three-dimensional extreme coordinates to generate a virtual rotation center of the beam; the microelectromechanical multi-axis displacement system is driven to perform attitude angle deflection scanning with the virtual rotation center of the beam as the origin; during the scanning, the optical power signal of the edge optical channel station is collected, the attitude angle that meets the preset energy threshold is extracted to construct a spatial attitude point set, and a multi-dimensional attitude envelope region is generated; The geometric center of the multidimensional attitude envelope region is calculated and the corresponding attitude angle combination is extracted as the optical axis alignment target attitude. The microelectromechanical multi-axis displacement system is driven to deflect to the optical axis alignment target attitude. After completion, the virtual rotation center of the beam is released, and the translation scan to deflect to the optical axis alignment target attitude is repeatedly executed until the spatial offset of the extracted three-dimensional extreme coordinates and the angular offset of the optical axis alignment target attitude are both less than the preset convergence threshold.

[0007] Preferably, the translation scanning process involves controlling the microelectromechanical multi-axis displacement system to output minute displacement disturbances along the three orthogonal directions of the X-axis, Y-axis, and Z-axis with an initial spatial step size; simultaneously acquiring the optical power signals of the reference optical channel station before and after the minute displacement disturbances in each orthogonal direction; smoothing the acquired optical power signals using a digital filtering algorithm; performing differential calculations on the smoothed optical power signals to obtain the rate of change of optical power at the current spatial position in the X-axis, Y-axis, and Z-axis directions, and synthesizing and constructing an optical field energy vector representing the spatial distribution direction of the current optical field energy; the X-axis and Y-axis form a beam lateral alignment plane perpendicular to the beam propagation direction, and the Z-axis is parallel to the beam propagation direction as the beam depth alignment axis.

[0008] Preferably, the three-dimensional extreme value coordinate extraction process involves taking the positive direction of the light field energy vector as the current optimization search direction and driving the microelectromechanical multi-axis displacement system to perform adaptive stepping movement along the optimization search direction; The adaptive step movement step size allocation logic is as follows: when the magnitude of the light field energy vector is greater than the first-order judgment threshold, the first spatial step size is extracted as the current step displacement; when the magnitude is less than or equal to the first-order judgment threshold, the second spatial step size is extracted as the current step displacement, and the first spatial step size is greater than the second spatial step size; after each adaptive step movement is completed, a translation scan is re-executed to update the light field energy vector at the current position. The action of moving along the gradient in a loop is executed. When the magnitude of the light field energy vector is less than the extreme value determination threshold, it is determined that the light field energy distribution extreme point has been reached. The spatial position of the microelectromechanical multi-axis displacement system on the X, Y, and Z axes at this time is extracted as the three-dimensional extreme value coordinates of the corresponding beam mode field energy peak.

[0009] Preferably, the virtual rotation center of the beam is the kinematic calculation origin constructed based on the spatial offset vector between the origin of the physical coordinate system of the microelectromechanical multi-axis displacement system and the three-dimensional extreme coordinates; when the attitude angle is deflected with the virtual rotation center of the beam as the origin, the inverse kinematics calculation of the optical path is performed according to the spatial offset vector, and while outputting the deflection commands of pitch angle, yaw angle and roll angle, the linear translation compensation commands along the X-axis, Y-axis and Z-axis are output simultaneously.

[0010] Preferably, the spatial attitude point set construction process involves recording the attitude angle sequence output by the microelectromechanical multi-axis displacement system in real time during the attitude angle deflection scan with the virtual rotation center of the beam as the origin, and simultaneously acquiring the optical power signal corresponding to the edge optical channel station; setting an absolute optical power lower limit as the preset energy threshold; comparing and judging the acquired optical power signal; extracting the attitude angle corresponding to the optical power signal being greater than or equal to the preset energy threshold as the effective attitude angle; and discretizing all effective attitude angles in the attitude space with pitch angle, yaw angle and roll angle as coordinate axes to construct the spatial attitude point set.

[0011] Preferably, the multidimensional attitude envelope generation process involves using an outer contour recognition algorithm to perform boundary calculations on the spatial attitude point set, extracting the outermost boundary feature points that surround all effective attitude angles; connecting the outermost boundary feature points to form a closed multidimensional spatial geometric configuration, and defining the closed multidimensional spatial geometric configuration as the multidimensional attitude envelope region.

[0012] Preferably, the geometric center is the spatial centroid of the multidimensional attitude envelope region within the attitude space constructed by pitch angle, yaw angle, and roll angle; the spatial centroid is obtained by solving the mean of spatial coordinates of the boundary feature points defining the multidimensional attitude envelope region; the attitude angle combination is the angle vector formed by the pitch angle coordinates, yaw angle coordinates, and roll angle coordinates corresponding to the spatial centroid in the attitude space; The microelectromechanical multi-axis displacement system is driven to deflect to the optical axis and align with the target point using a dual closed-loop control strategy. The dual closed-loop control strategy involves integrating MEMS displacement sensors and MEMS attitude sensors on each motion axis of the microelectromechanical multi-axis displacement system. After the control system issues a deflection command, it collects the physical coordinate data and physical deflection angle data output by the MEMS displacement sensor and MEMS attitude sensor, performs a closed-loop comparison with the theoretical coordinates and theoretical angles of the optical axis aligned with the target point, and performs physical displacement error compensation until the actual physical pose fed back by the sensor reaches the optical axis aligned with the target point.

[0013] Preferably, the spatial offset of the three-dimensional extreme coordinates is the spatial Euclidean distance between the three-dimensional extreme coordinates extracted in the current iteration and the three-dimensional extreme coordinates extracted in the previous iteration in two consecutive iterations; the angular offset of the optical axis alignment attitude is the comprehensive angular vector magnitude formed by the angular differences in pitch angle, yaw angle and roll angle between the optical axis alignment attitude extracted in the current iteration and the optical axis alignment attitude extracted in the previous iteration in two consecutive iterations.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs an independent virtual rotation center for the beam by dynamically mapping the physical rotation center of the microelectromechanical multi-axis displacement system to the three-dimensional extreme coordinates of the beam emitted from the reference optical channel. This effectively decouples spatial translation and attitude deflection at the control layer. During subsequent attitude angle optimization, the system continuously rotates around this virtual origin, locking the spatial position of the reference station. This fundamentally eliminates the dynamic misalignment and oscillation problem caused by adjusting the angle of a certain station in traditional alignment algorithms, ensuring the stability of optimization under rigid linkage conditions across multiple stations.

[0015] 2. This invention evaluates the deflection state by selectively acquiring optical power signals from edge optical channel workstations after establishing a virtual rotation center, utilizing the physical leverage amplification effect brought about by the spatial span of the multi-workstation array. The control system transforms the sensitivity disadvantage of multiple channels to slight angular deflection into a perception advantage for attitude optimization, avoiding the risk of decoupling of remote workstations due to angular mismatch, and ensuring the synchronous alignment quality of all channels at the micro-nano mode field scale from the control mechanism perspective.

[0016] 3. This invention constructs a multi-dimensional attitude envelope region by extracting effective attitude angles that meet the energy threshold, and calculates the centroid of this spatial geometric envelope as the final optical axis alignment point with the target point, overcoming the limitation of traditional single-point gradient climbing algorithms that lack global overall analysis. This optimization strategy based on spatial distribution region characteristics has extremely strong robustness and anti-distortion ability, effectively preventing the control system from falling into a dead loop of local multi-peak values ​​when facing end-face defects or irregular mode field distributions, thus reliably and smoothly planning multi-degree-of-freedom poses that meet global coupling requirements. Attached Figure Description

[0017] Figure 1A schematic flowchart of a fiber micro-nano multi-station optical path coupling control method based on optical field optimization is provided for an embodiment of the present invention. Figure 2 This is a schematic diagram of the translation scanning process provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the three-dimensional extreme value coordinate extraction process provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 This invention provides a fiber micro / nano multi-station optical path coupling control method based on optical field optimization, the technical solution of which is as follows: A fiber micro / nano multi-station optical path coupling control method based on optical field optimization includes: The microelectromechanical multi-axis displacement system is controlled to drive the fiber micro-nano multi-station array to perform translational scanning relative to the docking target device, collect the light field energy vector of the beam emitted from the reference optical channel station, and extract the three-dimensional extreme coordinates of the corresponding beam mode field energy peak. The physical rotation center of the microelectromechanical multi-axis displacement system is mapped to the three-dimensional extreme coordinates to generate a virtual rotation center of the beam; the microelectromechanical multi-axis displacement system is driven to perform attitude angle deflection scanning with the virtual rotation center of the beam as the origin; during the scanning, the optical power signal of the edge optical channel station is collected, the attitude angle that meets the preset energy threshold is extracted to construct a spatial attitude point set, and a multi-dimensional attitude envelope region is generated; The geometric center of the multidimensional attitude envelope region is calculated and the corresponding attitude angle combination is extracted as the optical axis alignment target attitude. The microelectromechanical multi-axis displacement system is driven to deflect to the optical axis alignment target attitude. After completion, the virtual rotation center of the beam is released, and the translation scan to deflect to the optical axis alignment target attitude is repeatedly executed until the spatial offset of the extracted three-dimensional extreme coordinates and the angular offset of the optical axis alignment target attitude are both less than the preset convergence threshold.

[0020] Example 1: A microelectromechanical multi-axis displacement system is used to drive a fiber optic micro / nano multi-station array to perform translational scanning relative to the target device. See Figure 2The translation scanning process involves controlling the microelectromechanical multi-axis displacement system to output minute displacement disturbances along the three orthogonal directions of the X-axis, Y-axis, and Z-axis with an initial spatial step size; simultaneously acquiring the optical power signals of the reference optical channel station before and after the minute displacement disturbances in each orthogonal direction; smoothing the acquired optical power signals using a digital filtering algorithm; performing differential calculations on the smoothed optical power signals to obtain the rate of change of optical power at the current spatial position in the X-axis, Y-axis, and Z-axis directions, and synthesizing and constructing an optical field energy vector representing the spatial distribution direction of the current optical field energy; the X-axis and Y-axis form a beam lateral alignment plane perpendicular to the beam propagation direction, and the Z-axis is parallel to the beam propagation direction as the beam depth alignment axis.

[0021] Specifically, a spatial coordinate system is established, and the physical reference plane perpendicular to the beam propagation direction is defined as the beam lateral alignment plane formed by the X-axis and the Y-axis, and the axis parallel to the beam propagation direction is defined as the beam depth alignment axis, i.e., the Z-axis.

[0022] Obtain the fiber mode field diameter value corresponding to the reference optical channel station, and at the same time obtain the minimum mechanical step resolution of the microelectromechanical multi-axis displacement system; compare one-hundredth of the fiber mode field diameter value with the minimum mechanical step resolution, and extract the larger value as the initial spatial step size.

[0023] The microelectromechanical multi-axis displacement system is controlled to output small displacement disturbances independently along the three orthogonal directions of the X-axis, Y-axis, and Z-axis, with the initial spatial step size as the single displacement amount and the current spatial position as the reference center. The execution actions of the small displacement disturbances include positive disturbances that move the initial spatial step size along the positive direction of the coordinate axis and negative disturbances that move the initial spatial step size along the negative direction of the coordinate axis.

[0024] During the execution of the micro-displacement perturbation in each orthogonal direction, the optical power signal of the reference optical channel station is synchronously acquired by the photoelectric detection unit when it is at the reference center position, at the positive perturbation end position, and at the negative perturbation end position; specifically, after the system reaches each of the above spatial positions, it remains briefly still and at that position, it continuously acquires at high frequency at least five optical power signal data, forming a discrete data sequence containing position coordinates and corresponding multiple sets of power values.

[0025] The acquired optical power signal is smoothed by a moving average digital filtering algorithm. Specifically, by setting a sliding window containing five sampling data points, the time series data continuously acquired at each spatial location are summed by arithmetic average. The calculated arithmetic average is used as the final true optical power value at that spatial location. This eliminates extreme value jumps caused by high-frequency ambient light noise and mechanical micro-vibration, and outputs the smoothed optical power signal.

[0026] The smoothed optical power signal is subjected to differential calculation using the center difference method; specifically, the smoothed optical power signal corresponding to the positive disturbance end position on the same coordinate axis is extracted and the smoothed optical power signal corresponding to the negative disturbance end position is extracted, the two are subtracted to obtain the difference, and the difference is divided by twice the initial spatial step distance. Based on the above differential calculation logic, the above subtraction and division by distance operations are performed for the three orthogonal directions respectively, thereby obtaining the rate of change of optical power of the current spatial position in the X-axis direction, the rate of change of optical power in the Y-axis direction, and the rate of change of optical power in the Z-axis direction.

[0027] The obtained rate of change of optical power along the X-axis is taken as the first orthogonal vector component, the rate of change of optical power along the Y-axis is taken as the second orthogonal vector component, and the rate of change of optical power along the Z-axis is taken as the third orthogonal vector component. These three orthogonal vector components are vector-summed in three-dimensional space to synthesize and construct the optical field energy vector representing the current spatial distribution of optical field energy. The direction of the optical field energy vector points to the gradient direction in which the optical power increases the fastest in the current local space. The acquisition and differential calculation of the optical power signal are carried out in the linear domain, and the power unit is set to microwatts. By binding the initial spatial step size to the fiber mode field diameter and mechanical resolution, the displacement perturbation adapts to the optical characteristics and hardware limits, avoiding blind spots in optimization caused by improper fixed step size settings. At each perturbation location, high-frequency acquisition and moving average filtering are combined to effectively filter out stray light and extreme value jumps caused by mechanical micro-vibrations, ensuring the authenticity of the optical power signal. In addition, by using positive and negative bidirectional perturbation combined with the central difference method to synthesize the optical field energy vector, the gradient direction of energy rise can be accurately pointed with low computational complexity, avoiding meaningless mechanical blind scanning.

[0028] Collect the light field energy vector of the beam emitted from the reference optical channel station, and extract the three-dimensional extreme coordinates of the corresponding beam mode field energy peak. See Figure 3The three-dimensional extreme value coordinate extraction process involves taking the positive direction of the light field energy vector as the current optimization search direction and driving the microelectromechanical multi-axis displacement system to perform adaptive stepping movement along the optimization search direction. The adaptive step movement step size allocation logic is as follows: when the magnitude of the light field energy vector is greater than the first-order judgment threshold, the first spatial step size is extracted as the current step displacement; when the magnitude is less than or equal to the first-order judgment threshold, the second spatial step size is extracted as the current step displacement, and the first spatial step size is greater than the second spatial step size; after each adaptive step movement is completed, a translation scan is re-executed to update the light field energy vector at the current position. The action of moving along the gradient in a loop is executed. When the magnitude of the light field energy vector is less than the extreme value determination threshold, it is determined that the light field energy distribution extreme point has been reached. The spatial position of the microelectromechanical multi-axis displacement system on the X, Y, and Z axes at this time is extracted as the three-dimensional extreme value coordinates of the corresponding beam mode field energy peak.

[0029] Specifically, the physical size distribution characteristics of the beam mode field of the current docking target device are read, and the first spatial step size is set to be smaller than the width of the main lobe region of the beam energy, ensuring that a single optimization step does not cross the maximum search span of the entire main lobe of the beam energy. The second spatial step size is set to meet the minimum spatial resolution corresponding to the final target coupling loss tolerance of the optical path. With these physical constraints, the parameters can be configured to adaptively derive specific step size values ​​according to the device characteristics and ensure that the first spatial step size is strictly greater than the second spatial step size. At the same time, the amplitude of the dark current noise floor signal of the optical power meter at the optical path receiver is read, and the extreme value judgment threshold is set to combine the dark current amplitude and the power fluctuation variance of continuous sampling, and calculate the uncertainty conversion coefficient through a mathematical statistical model. The extreme value judgment threshold is set to a safety boundary value greater than the joint noise floor amplitude. At the initial spatial position of the system, the Gaussian energy distribution profile of the beam of the preset reference device is tested, and the slope abrupt change inflection point of the light field energy distribution curve from the steep edge region to the extreme flat region is extracted. The spatial gradient amplitude corresponding to the slope abrupt change inflection point is extracted and set as the first-order judgment threshold.

[0030] Extract the power rate components of the light field energy vector obtained from the current translation scan along the X, Y, and Z axes, respectively, and denot them as follows: , , The magnitude of the current light field energy vector is calculated using the formula. ; calculate separately , , The ratio of the magnitude to the modulus is used to obtain the direction cosine, thereby constructing a normalized three-dimensional direction vector. The positive direction of this normalized three-dimensional direction vector, i.e. the light field energy vector, is used as the current optimization search direction.

[0031] The magnitude of the currently calculated light field energy vector is compared with the first-order judgment threshold. When the magnitude of the light field energy vector is greater than the first-order judgment threshold, it is determined that the current spatial position of the optical path is located in the steep edge region of the light field energy distribution, and the first spatial step size is extracted as the current step displacement. When the magnitude is less than or equal to the first-order judgment threshold, it is determined that the current spatial position of the optical path is close to the extreme flat region of the light field energy distribution, and the second spatial step size is extracted as the current step displacement.

[0032] The extracted current step displacement is multiplied by the normalized three-dimensional direction vector to calculate and output the target displacement drive increments for the X, Y, and Z axes. The controllers of each axis of the microelectromechanical multi-axis displacement system are driven to perform an adaptive step movement along the optimization search direction according to the allocated target displacement drive increments. After each adaptive step movement physical action is completed, the translation scanning process is re-executed to collect and differentially calculate the latest optical power signal, and update the optical field energy vector and its corresponding magnitude at the current spatial position.

[0033] The microelectromechanical multi-axis displacement system is controlled to cyclically execute the above-mentioned movement along the positive gradient direction, and the magnitude of the optical field energy vector after each update is continuously monitored. When the comparison finds that the magnitude of the updated optical field energy vector is less than the extreme value determination threshold, it is determined that the change in the optical field gradient has been submerged in the system's physical noise floor, that is, the optical field energy distribution extreme point has been reached. The stepping movement command output is immediately stopped, and the spatial physical coordinate values ​​of the X, Y, and Z axes recorded in the servo feedback loop of the microelectromechanical multi-axis displacement system at this time are read and stored, and extracted as the three-dimensional extreme value coordinates of the corresponding beam mode field energy peak. In order to eliminate local minima or saddle point interference, after determining that the optical field energy distribution extreme point has been reached, this embodiment further confirms that it is a maximum peak value by comparing the absolute optical power of the point with the three-dimensional neighborhood sampling points.

[0034] By introducing adaptive step size allocation logic based on the magnitude of the light field energy vector, a larger step size is used in the steep edge region of the light field energy distribution, and a smaller step size is automatically switched in the gentle region approaching the pole. Without increasing the additional hardware cost, the algorithm logic effectively avoids out-of-bounds oscillations near the pole during the optimization process. Using the extreme value judgment threshold as the termination condition of the optimization action enables the control algorithm to actively identify and remove the interference of the system's physical noise, preventing disordered high-frequency micro-motions in the microelectromechanical multi-axis displacement system near the extreme point due to the tiny jumps in the instrument readings.

[0035] The physical rotation center of the microelectromechanical multi-axis displacement system is mapped to the three-dimensional extreme coordinates to generate a virtual rotation center of the beam. The virtual rotation center of the beam is the kinematic solution origin constructed based on the spatial offset vector between the origin of the physical coordinate system of the microelectromechanical multi-axis displacement system and the three-dimensional extreme coordinates. When the attitude angle is deflected with the virtual rotation center of the beam as the origin, the inverse kinematics calculation of the optical path is performed according to the spatial offset vector. While outputting the deflection commands of pitch angle, yaw angle and roll angle, the linear translation compensation commands along the X-axis, Y-axis and Z-axis are output simultaneously.

[0036] Specifically, the controller reads the three-dimensional spatial coordinates of the origin of the physical coordinate system when the microelectromechanical multi-axis displacement system is in a mechanical reset zero-point state; simultaneously, it extracts the three-dimensional extreme coordinates recorded in the control program; it directly subtracts the coordinates of the origin of the physical coordinate system from the coordinates of each axis of the three-dimensional extreme coordinates to calculate the three-dimensional directed line segment parameters pointing from the origin of the physical coordinate system to the three-dimensional extreme coordinates, and uses these three-dimensional directed line segment parameters as the spatial offset vector.

[0037] In the controller's underlying coordinate system mapping matrix, the calculated spatial offset vector is used to perform a homogeneous translation transformation on the original physical coordinate system. Before coordinate mapping, the external parameter calibration of the system's physical coordinates is completed in advance: a three-dimensional physical coordinate system is established with the mechanical zero point of the microelectromechanical system base as the absolute origin; the static spatial relative position vector between the geometric center of the fiber optic end face and the physical axis of the microelectromechanical rotary joint is measured and recorded through an offline calibration program. Subsequently, in the controller's underlying coordinate system mapping matrix, the calculated spatial offset vector and the static spatial relative position vector are used to perform a homogeneous translation transformation on the original physical coordinate system. When performing the above deflection scan and kinematic calculation, the rotation matrix is ​​constructed using the extrinsic Euler angle ZYX axis sequence. The deflection actions of the yaw angle, pitch angle and roll angle are sequentially corresponding to the absolute reference frame according to the right-hand rule, ensuring that the rotation calculation has no multiple solutions and eliminating the interference of singular points in the inverse calculation. Through this translation transformation operation, the reference point for the rotation kinematic calculation of the system is forcibly migrated from the origin of the physical coordinate system to the spatial physical position where the three-dimensional extreme coordinates are located. The migrated calculation reference point is established as the kinematic calculation origin, and the virtual rotation center of the beam is established.

[0038] When driving the microelectromechanical multi-axis displacement system to perform attitude angle deflection scanning with the virtual rotation center of the beam as the origin, the pitch angle, yaw angle, and roll angle values ​​required by the current scanning task are obtained; using the obtained pitch angle, yaw angle, and roll angle values, a corresponding three-dimensional spatial rotation matrix is ​​constructed according to the Euler angle rotation transformation rule; the spatial offset vector is multiplied by the three-dimensional spatial rotation matrix to calculate the spatial parasitic displacement vector generated by the origin of the physical coordinate system of the microelectromechanical multi-axis displacement system when following the above attitude angle deflection.

[0039] The calculated components of the spatial parasitic displacement vector on the X, Y, and Z axes are inversely negative to generate linear translation compensation values ​​along the X, Y, and Z axes to compensate for the runout error at the lever end of the robotic arm. The required pitch, yaw, and roll angle values ​​are encapsulated into yaw commands, and the generated linear translation compensation values ​​along the X, Y, and Z axes are also encapsulated into linear translation compensation commands. The controller's multi-axis linkage interpolation algorithm module is invoked, and within the same underlying hardware clock cycle, the yaw commands and linear translation compensation commands are packaged and synchronously sent to each axial servo driver of the microelectromechanical multi-axis displacement system, forcing the physical motors to perform linear translation compensation actions along the X, Y, and Z axes simultaneously while performing attitude angle yaw.

[0040] The underlying hardware clock cycle is set to a strict fixed timestamp (e.g., 1ms); the multi-axis linkage interpolation algorithm module adopts a high-order S-curve trajectory interpolation plan to ensure smooth and continuous speed and acceleration of each servo motor during operation. Under this interpolation strategy, the timing of the rotary axis reaching the specified posture and the linear axis reaching the compensated displacement are forcibly aligned, and the dynamic positioning synchronization error of the measured translational following compensation and posture deflection action is strictly controlled within the mechanical tolerance range (e.g., 10nm). By constructing a virtual rotation center for the beam and combining it with inverse kinematics calculations, the system can accurately calculate the spatial parasitic displacement that accompanies the deflection of the physical axis and implement reverse linear translation compensation. This mechanism effectively counteracts the inherent mechanical lever runout error of the microelectromechanical system during attitude adjustment from the underlying control logic. At the same time, thanks to the synchronous issuance of deflection commands and linear translation compensation commands within the same hardware clock cycle, the time lag between translation and rotation actions is eliminated.

[0041] The microelectromechanical multi-axis displacement system is driven to perform attitude angle deflection scanning with the virtual rotation center of the beam as the origin; during the scanning, the optical power signal of the edge optical channel station is collected, the attitude angle that meets the preset energy threshold is extracted to construct a spatial attitude point set, and a multi-dimensional attitude envelope region is generated. The process of constructing the spatial attitude point set involves recording the attitude angle sequence output by the microelectromechanical multi-axis displacement system in real time during the attitude angle deflection scan with the virtual rotation center of the beam as the origin, and simultaneously acquiring the optical power signal corresponding to the edge optical channel station. An absolute lower limit of optical power is set as the preset energy threshold. The acquired optical power signals are compared and judged. The attitude angles corresponding to the optical power signals that are greater than or equal to the preset energy threshold are extracted as valid attitude angles. All valid attitude angles are then discretized and mapped in the attitude space with pitch angle, yaw angle, and roll angle as coordinate axes to construct the spatial attitude point set.

[0042] The process of generating the multidimensional attitude envelope region involves using an outer contour recognition algorithm to perform boundary calculations on the spatial attitude point set, extracting the outermost boundary feature points that surround all effective attitude angles, connecting the outermost boundary feature points to form a closed multidimensional spatial geometric configuration, and defining the closed multidimensional spatial geometric configuration as the multidimensional attitude envelope region.

[0043] Specifically, before performing attitude angle deflection scanning, the reference emission power value of the light source at the input end of the injected fiber micro-nano multi-station array is first read, and the physical specification parameter table of the current docking target device is called to extract the maximum allowable additional coupling loss value required for the device to achieve normal optical signal transmission; the reference emission power value of the light source is subtracted from the maximum allowable additional coupling loss value to calculate the absolute optical power lower limit that the edge optical channel station can tolerate under extreme alignment deviation; the calculated absolute optical power lower limit is directly written into the comparator and set as the preset energy threshold.

[0044] During the attitude angle deflection scan with the virtual rotation center of the beam as the origin, the hardware clock stamp interrupt mechanism of the microelectromechanical system is invoked. Within the same hardware clock cycle when the servo driver position encoder outputs the current pitch angle, yaw angle, and roll angle angle data, a synchronization trigger command is sent to the optical power acquisition module to read the optical power signal corresponding to the edge optical channel station. The attitude angle sequence and optical power signal are bound together using the same hardware clock stamp to complete the real-time and absolutely synchronized acquisition action. The control logic traverses the bound data stream and compares the acquired optical power signal with the set preset energy threshold. Data groups with optical power signals less than the preset energy threshold are discarded, and the attitude angles corresponding to optical power signals greater than or equal to the preset energy threshold are extracted as valid attitude angles. A three-dimensional orthogonal attitude space coordinate system is established with pitch angle as the first coordinate axis, yaw angle as the second coordinate axis, and roll angle as the third coordinate axis. All extracted valid attitude angles are discretized and mapped in this three-dimensional orthogonal attitude space coordinate system as three-dimensional coordinate points to construct the spatial attitude point set.

[0045] A three-dimensional convex hull algorithm is introduced as the outer contour recognition algorithm for boundary calculation of the spatial attitude point set; the spatial attitude point set is traversed to extract multiple extreme attitude points with the maximum and minimum coordinate values ​​in the directions of pitch angle, yaw angle and roll angle coordinate axes respectively; these extreme attitude points are connected to construct an initial closed polyhedron in a three-dimensional orthogonal attitude space coordinate system; the spatial vertical distances from all remaining valid attitude angle coordinate points in the spatial attitude point set to each outer surface of the initial closed polyhedron are calculated sequentially.

[0046] When it is determined that a valid attitude angle is located outside the outer surface, the valid attitude angle with the largest vertical distance is selected and added as a new vertex. The original local surface is removed and reconnected with the new vertex to generate a new polygonal outer surface. The above distance determination and surface update calculation process is iterated until all valid attitude angles in the spatial attitude point set are contained inside the ever-expanding polyhedron or exactly located on the surface of the polyhedron. At this time, the algorithm converges, and all vertices of the final polyhedron are extracted and defined as the outermost boundary feature points that surround all valid attitude angles. Using the Delaunay triangulation rule, all the extracted outermost boundary feature points are connected to each other with planar triangular patches to form a tight, seamless, and non-deformed closed multidimensional spatial geometry. This closed multidimensional spatial geometry is defined as the multidimensional attitude envelope region that characterizes the tolerance limit of optical path attitude coupling.

[0047] By introducing a hardware clock stamp interrupt mechanism, strict synchronization between multi-axis spatial attitude deflection and edge channel optical power acquisition is achieved, eliminating communication delay misalignment between the control system and the measuring instruments, and ensuring that the constructed spatial attitude point set can truly reflect the distribution boundary of physical optical field energy. On this basis, the three-dimensional convex hull algorithm and Delaunay triangulation rule are used to mathematically transform the selected discrete attitude point set into a closed multi-dimensional attitude envelope region, which objectively and rigorously defines the optical path attitude coupling tolerance limit boundary of the target device in spatial geometry.

[0048] Calculate the geometric center of the multidimensional attitude envelope region and extract the corresponding attitude angle combination as the optical axis alignment target attitude, and drive the microelectromechanical multi-axis displacement system to deflect to the optical axis alignment target attitude. The geometric center is the spatial centroid of the multidimensional attitude envelope region within the attitude space constructed by pitch angle, yaw angle, and roll angle; the spatial centroid is obtained by solving the mean of the spatial coordinates of the boundary feature points defining the multidimensional attitude envelope region; the attitude angle combination is the angle vector formed by the pitch angle coordinates, yaw angle coordinates, and roll angle coordinates corresponding to the spatial centroid in the attitude space; The microelectromechanical multi-axis displacement system is driven to deflect to the optical axis and align with the target point using a dual closed-loop control strategy. The dual closed-loop control strategy involves integrating MEMS displacement sensors and MEMS attitude sensors on each motion axis of the microelectromechanical multi-axis displacement system. After the control system issues a deflection command, it collects the physical coordinate data and physical deflection angle data output by the MEMS displacement sensor and MEMS attitude sensor, performs a closed-loop comparison with the theoretical coordinates and theoretical angles of the optical axis aligned with the target point, and performs physical displacement error compensation until the actual physical pose fed back by the sensor reaches the optical axis aligned with the target point.

[0049] Specifically, all the outermost boundary feature points of the constructed multidimensional pose envelope region are extracted, and the total number of these boundary feature points recorded in the system memory is read and denoted as a numerical value. Iterate through and read the coordinate values ​​of each outermost boundary feature point in the attitude space, and record them as pitch angle components. Yaw angle component and roll angle component ,in From 1 to A positive integer; invoke the spatial discrete point mean calculation logic to calculate the arithmetic mean of all outermost boundary feature points on the three independent coordinate axes, and obtain the center pitch angle. Center yaw angle and center roll angle The spatial centroid, which is the average of the spatial coordinates, is mapped to a spatial location within the attitude space and is also defined as the geometric center of the multidimensional attitude envelope. The pitch angle corresponding to this spatial centroid is... Center yaw angle and roll component A three-dimensional angle vector is constructed, and this angle vector is defined as the attitude of the optical axis aligned with the target point. Considering that the angle span of the tolerance window in the micro-nano optical path coupling is extremely small, and the distribution of the effective attitude point set has local symmetry, this embodiment uses the arithmetic mean of the spatial coordinates of the boundary feature points to approximate the real continuous spatial centroid. On the linear motor motion guide rails of the microelectromechanical multi-axis displacement system that performs X-axis, Y-axis, and Z-axis translation, microelectromechanical system-level (MEMS) displacement sensors are physically attached to collect three-dimensional physical coordinate data in real time. Microelectromechanical system-level (MEMS) attitude sensors are attached to the rotary joints that control pitch, yaw, and roll to collect three-dimensional physical yaw angle data in real time. The control system generates theoretical angle yaw commands based on the extracted optical axis alignment attitude with the target point, and combines it with the previously constructed virtual rotation center of the beam. Through inverse kinematics, it synchronously calculates the theoretical linear translation compensation commands required to maintain the spatial position of the virtual rotation center unchanged, and sends these two sets of commands to the underlying servo drive system. During the deflection motion of the MEMS multi-axis displacement system, the underlying servo drive system employs a cascaded dual closed-loop structure to open the outer loop of attitude control and the inner loop of displacement compensation. In the outer loop of attitude control, the physical deflection angle data output by the MEMS attitude sensor is collected in real time, and the difference is calculated by subtracting it from the theoretical angle of the optical axis aligned with the target point to obtain the dynamic angle deviation. The proportional-integral-derivative (PID) control algorithm is then called to convert the angle deviation into the dynamic drive voltage of the rotary joint motor. Simultaneously, in the inner loop of displacement compensation, the physical coordinate data output by the MEMS displacement sensor is collected in real time, and the difference is calculated by subtracting it from the calculated theoretical coordinates to obtain the parasitic displacement deviation caused by mechanical linkage. The PID control algorithm is then called to convert the parasitic displacement deviation into the dynamic compensation drive voltage of the linear motor to perform high-frequency physical displacement error compensation. On the linear motor motion guide, a capacitive MEMS displacement sensor (e.g., range ±100μm, absolute measurement accuracy better than 1nm) is physically attached to provide the absolute true value data of the three-dimensional physical coordinates under a fixed reference frame; a high-precision MEMS inertial or optical attitude sensor (angle accuracy better than 0.001°) is attached to the rotary joint to collect three-dimensional physical deflection angle data in real time.

[0050] The minimum mechanical step resolution of linear displacement and the minimum angular step resolution of rotary joint, calibrated in the specifications of the microelectromechanical multi-axis displacement system, are read respectively. The minimum mechanical step resolution of linear displacement is set as a preset linear displacement threshold, and the minimum angular step resolution of rotary joint is set as a preset angular threshold. The control system continuously and cyclically executes the above-mentioned outer loop angular comparison and inner loop physical displacement error compensation actions, and monitors the absolute values ​​of the angular deviation and the parasitic displacement deviation in real time. The absolute value of the parasitic displacement deviation is compared with the preset linear displacement threshold, and the absolute value of the angular deviation is compared with the preset angular threshold. If any one of the deviations is greater than or equal to its corresponding preset threshold, the drive system continues to cyclically execute displacement error compensation and angular deflection actions. This continues until it is determined that the absolute values ​​of the two types of deviations are strictly less than their corresponding preset thresholds in multiple consecutive hardware sampling cycles, confirming that the system meets the variable stability termination condition, and the current closed-loop deflection drive task is terminated. By extracting the spatial centroid of the multidimensional attitude envelope region as the alignment target point, the false extremum interference caused by local optical field noise is effectively avoided, ensuring that the optical path is anchored at the center of the attitude tolerance window. At the same time, a dual closed-loop control strategy based on MEMS sensors is adopted, and the minimum mechanical resolution of the device is used as the convergence benchmark, effectively decoupling and compensating for the mechanical parasitic displacement error during spatial deflection from the bottom layer.

[0051] After completion, the virtual rotation center of the beam is released, and the translation scan is repeatedly performed until the optical axis is aligned with the target point. The process continues until the spatial offset of the three-dimensional extreme coordinates extracted twice and the angular offset of the optical axis alignment with the target point are both less than the preset convergence threshold.

[0052] The spatial offset of the three-dimensional extreme coordinates is the spatial Euclidean distance between the three-dimensional extreme coordinates extracted in the current iteration and the three-dimensional extreme coordinates extracted in the previous iteration in two consecutive iterations; the angular offset of the optical axis alignment attitude is the comprehensive angular vector magnitude formed by the angular differences in pitch angle, yaw angle and roll angle between the optical axis alignment attitude extracted in the current iteration and the optical axis alignment attitude extracted in the previous iteration in two consecutive iterations. Specifically, after a single attitude angle deflection scan and alignment action is completed, the control system sends a coordinate system reset command to the underlying servo driver, forcibly clearing each axial component of the spatial offset vector stored in the underlying coordinate system mapping matrix of the controller to zero; by clearing the zeroing operation, the previous spatial translation transformation is canceled, so that the rotational kinematics calculation reference point of the system is returned from the physical position of the spatial extremum and reconnected to the physical coordinate system origin of the microelectromechanical multi-axis displacement system, thus completely removing the virtual rotation center of the beam from the underlying control logic.

[0053] The three-axis physical positioning repeatability value specified in the product specification of the microelectromechanical multi-axis displacement system is read and set as the spatial convergence threshold; the single-step minimum angular resolution value of the servo encoder of the system's rotary joint is read and set as the angular convergence threshold, and the spatial convergence threshold and the angular convergence threshold together constitute the preset convergence threshold determined by the system.

[0054] In the control logic, a global iteration counter is set to record the number of iterations for optimization. Before starting the latest iteration, the historical status register is called to save and solidify the three-dimensional extreme coordinates and the optical axis alignment with the target point attitude extracted in the previous optimization. The complete action flow of translation scanning until the optical axis is aligned with the target point attitude is started again, and the three-dimensional extreme coordinates and the optical axis alignment with the target point attitude extracted in the latest iteration are collected and recorded in real time.

[0055] The spatial offset of the three-dimensional extreme coordinates is calculated by extracting the three-dimensional extreme coordinates of the current acquisition and the three-dimensional extreme coordinates of the previous acquisition, and calculating the difference between the coordinate values ​​of the two on the three orthogonal spatial axes respectively; then, the three coordinate value differences are squared and summed, and the sum is squared to obtain the absolute value of the spatial straight line distance, which is the spatial offset of the three-dimensional extreme coordinates extracted in the two consecutive acquisitions.

[0056] The angular offset of the optical axis alignment with the target point is calculated by extracting the current optical axis alignment with the target point and the previously stored optical axis alignment with the target point. The angular differences between the two in the three independent rotational dimensions of pitch, yaw, and roll are calculated respectively. The angular differences in these three dimensions are squared and summed. The square root of the sum is then calculated to obtain the absolute value of the spatial comprehensive angular deviation, which is the comprehensive angular vector magnitude formed by the angular differences in pitch, yaw, and roll corresponding to the two consecutive extracted optical axis alignment with the target point. The calculated spatial offset value is compared with the obtained spatial convergence threshold, and the calculated comprehensive angle vector magnitude value is compared with the obtained angle convergence threshold. If either of them is greater than or equal to its corresponding convergence threshold, the drive system continues to execute the next translation and deflection iteration. When it is determined that both the spatial offset value and the comprehensive angle vector magnitude value are strictly less than their corresponding preset convergence thresholds, it is confirmed that the beam spatial position and angle attitude have been fully converged at the mechanical level, and the optical path coupling optimization control action is terminated. At the same time, a global maximum allowable iteration count upper limit is set (e.g., 20 times). If convergence is not achieved after exceeding this upper limit, or if an abnormality is detected during the process, such as the optical power dropping below the noise floor or the coordinate command exceeding the mechanical hard limit, the corresponding fault exit will be triggered, and the historical status register will be automatically called to safely roll back each control axis to the target pose of the last effective convergence. By extracting the three-dimensional extreme coordinates of the reference optical channel to construct a virtual rotation center for the beam, the optimization process of spatial position and attitude angle is decoupled from the kinematic level, effectively avoiding the physical parasitic displacement interference that accompanies the attitude adjustment of multi-axis devices. At the same time, considering the characteristics of micro-nano multi-station array devices, this method uses the edge optical channel signal to construct a multi-dimensional attitude envelope region and uses its geometric center as the alignment target point, overcoming the edge channel coupling imbalance defect that is easily caused by traditional single-point alignment, and ensuring the consistency of the global optical power distribution of the array from the spatial geometry level. In addition, by alternately performing translation and deflection actions around this virtual rotation center, and combining spatial and angular offsets as the iteration termination criteria, a closed-loop optimization logic is constructed to prevent the system control from falling into false extremes or disordered oscillations, and to ensure the stable convergence of multi-degree-of-freedom optical path coupling docking.

[0057] Example 2: This embodiment applies a fiber micro-nano multi-station optical path coupling control method based on optical field optimization to the coupling and packaging of multi-channel silicon-based optoelectronic chips and fiber arrays. The microelectromechanical multi-axis displacement system (MEMS) drives a fiber optic micro / nano multi-position array to perform translational scanning relative to the target device, acquiring the optical field energy vector of the beam emitted from the reference optical channel position, and extracting the three-dimensional extreme coordinates of the corresponding beam mode field energy peak. In this specific scenario, the fiber optic micro / nano multi-position array is an 8-channel single-mode fiber array, and the control system selects the fourth channel in the center as the reference optical channel position. The system first controls the MEMS to output small displacement perturbations along the three orthogonal directions of the X, Y, and Z axes with an initial spatial step size. The optical field energy vector is synthesized based on the optical power change rate calculated by smoothing and differentiation, and adaptively steps along the optimization search direction. Finally, the fourth channel is stopped at the optical field energy distribution extreme point, and the spatial position of the MEMS on the X, Y, and Z axes at this time is recorded as the three-dimensional extreme coordinates. The physical rotation center of the MEMS is dynamically mapped to the three-dimensional extreme coordinates of the fourth channel, generating an independent virtual rotation center of the beam. This process effectively decouples spatial translation and attitude deflection at the control layer, allowing the spatial position of the reference station (channel 4) to be locked. The first and eighth channels, with the largest spans at both ends of the fiber array, are designated as the edge optical channel workstations. Due to the rigid spatial geometric correlation between the multiple workstations, the micro-nano optical field is extremely sensitive to the six degrees of freedom in space. Even slight attitude changes, amplified by the spatial distance between the multiple workstations, can disrupt the optical field alignment of the edge workstations. Therefore, during the attitude angle (pitch, yaw, and roll) deflection scan around the virtual origin of the fourth channel, the optical power signals of the first and eighth channels are prioritized for acquisition. The physical leverage amplification effect brought about by the spatial span of the multi-workstation array is utilized to keenly perceive angular deviations. The effective attitude angles when the optical power signals of the first and eighth channels are greater than or equal to a preset energy threshold are discretized and mapped to construct a spatial attitude point set. An outer contour recognition algorithm is then used to extract the outermost boundary feature points, generating a multi-dimensional attitude envelope region. The geometric center of the multidimensional attitude envelope is calculated, and the corresponding attitude angle combination is extracted as the optical axis alignment attitude with the target point. A dual closed-loop control strategy is used to drive the microelectromechanical multi-axis displacement system to deflect to the optical axis alignment attitude with the target point. After completion, the virtual rotation center of the beam is released, and the above actions from translation scanning (re-finding the extremum of the 4th channel) to attitude deflection (finding the overall optimal centroid) are executed again in a loop. The entire optimization process iterates continuously until the spatial offset of the three-dimensional extreme coordinates of the fourth channel extracted twice consecutively and the angular offset of the target point attitude are both less than the preset convergence threshold calibrated by the system. By using an optimization strategy based on the characteristics of the spatial distribution area, the control system is effectively prevented from getting stuck in a dead loop in a complex optical field. A global pose that allows all eight stations to achieve optimal coupling efficiency is planned under a very small mode field size. The global optimal conclusion is based on the following physical premise: Since the 8-channel single-mode fiber array has extremely high spatial rigidity consistency in its manufacturing process, the spacing error between adjacent stations inside it is much smaller than the beam mode field diameter. Therefore, when the extreme value of the reference channel (the fourth channel) and the envelope of the edge channels (the first and eighth channels) both meet the threshold condition, according to the principle of rigid body geometric linkage, it can be regarded that the remaining channels in the middle have synchronously reached the global optimal pose.

[0058] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A fiber optic micro / nano multi-station optical path coupling control method based on optical field optimization, characterized in that, include: The microelectromechanical multi-axis displacement system is controlled to drive the fiber micro-nano multi-station array to perform translational scanning relative to the docking target device, collect the light field energy vector of the beam emitted from the reference optical channel station, and extract the three-dimensional extreme coordinates of the corresponding beam mode field energy peak. The physical rotation center of the microelectromechanical multi-axis displacement system is mapped to the three-dimensional extreme coordinates to generate a virtual rotation center of the beam; the microelectromechanical multi-axis displacement system is driven to perform attitude angle deflection scanning with the virtual rotation center of the beam as the origin; during the scanning, the optical power signal of the edge optical channel station is collected, the attitude angle that meets the preset energy threshold is extracted to construct a spatial attitude point set, and a multi-dimensional attitude envelope region is generated; Calculate the geometric center of the multidimensional attitude envelope region and extract the corresponding attitude angle combination as the optical axis alignment target attitude, and drive the microelectromechanical multi-axis displacement system to deflect to the optical axis alignment target attitude. After completion, the virtual rotation center of the beam is released, and the translation scan is repeatedly performed until the optical axis is aligned with the target point. The process continues until the spatial offset of the three-dimensional extreme coordinates extracted twice and the angular offset of the optical axis alignment with the target point are both less than the preset convergence threshold.

2. The fiber micro / nano multi-station optical path coupling control method based on optical field optimization according to claim 1, characterized in that, The translation scanning process involves controlling the microelectromechanical multi-axis displacement system to output minute displacement disturbances along the three orthogonal directions of the X, Y, and Z axes with an initial spatial step size; simultaneously acquiring the optical power signals of the reference optical channel station before and after the minute displacement disturbances in each orthogonal direction; smoothing the acquired optical power signals using a digital filtering algorithm; performing differential calculations on the smoothed optical power signals to obtain the rate of change of optical power at the current spatial position in the X, Y, and Z axis directions, and synthesizing and constructing an optical field energy vector representing the spatial distribution direction of the current optical field energy; the X and Y axes form a beam lateral alignment plane perpendicular to the beam propagation direction, and the Z axis, parallel to the beam propagation direction, serves as the beam depth alignment axis.

3. The fiber micro / nano multi-station optical path coupling control method based on optical field optimization according to claim 1, characterized in that, The three-dimensional extreme value coordinate extraction process involves taking the positive direction of the light field energy vector as the current optimization search direction and driving the microelectromechanical multi-axis displacement system to perform adaptive stepping movement along the optimization search direction. The adaptive step size allocation logic is as follows: when the magnitude of the light field energy vector is greater than the first-order judgment threshold, the first spatial step size is extracted as the current step displacement; when the magnitude is less than or equal to the first-order judgment threshold, the second spatial step size is extracted as the current step displacement, and the first spatial step size is greater than the second spatial step size. After each adaptive step movement is completed, a translation scan is re-executed to update the light field energy vector at the current position; The action of moving along the gradient in a loop is executed. When the magnitude of the light field energy vector is less than the extreme value determination threshold, it is determined that the light field energy distribution extreme point has been reached. The spatial position of the microelectromechanical multi-axis displacement system on the X, Y, and Z axes at this time is extracted as the three-dimensional extreme value coordinates of the corresponding beam mode field energy peak.

4. The fiber micro / nano multi-station optical path coupling control method based on optical field optimization according to claim 1, characterized in that, The virtual rotation center of the beam is the kinematic solution origin constructed based on the spatial offset vector between the origin of the physical coordinate system of the microelectromechanical multi-axis displacement system and the three-dimensional extreme coordinates. When the attitude angle is deflected with the virtual rotation center of the beam as the origin, the inverse kinematics calculation of the optical path is performed according to the spatial offset vector. While outputting the deflection commands of pitch angle, yaw angle and roll angle, the linear translation compensation commands along the X-axis, Y-axis and Z-axis are output simultaneously.

5. The fiber micro / nano multi-station optical path coupling control method based on optical field optimization according to claim 1, characterized in that, The process of constructing the spatial attitude point set involves recording the attitude angle sequence output by the microelectromechanical multi-axis displacement system in real time during the attitude angle deflection scan with the virtual rotation center of the beam as the origin, and simultaneously acquiring the optical power signal corresponding to the edge optical channel station. An absolute lower limit of optical power is set as the preset energy threshold. The acquired optical power signals are compared and judged. The attitude angles corresponding to the optical power signals that are greater than or equal to the preset energy threshold are extracted as valid attitude angles. All valid attitude angles are then discretized and mapped in the attitude space with pitch angle, yaw angle, and roll angle as coordinate axes to construct the spatial attitude point set.

6. The fiber micro / nano multi-station optical path coupling control method based on optical field optimization according to claim 5, characterized in that, The process of generating the multidimensional attitude envelope region involves using an outer contour recognition algorithm to perform boundary calculations on the spatial attitude point set, extracting the outermost boundary feature points that surround all effective attitude angles, connecting the outermost boundary feature points to form a closed multidimensional spatial geometric configuration, and defining the closed multidimensional spatial geometric configuration as the multidimensional attitude envelope region.

7. The fiber micro / nano multi-station optical path coupling control method based on optical field optimization according to claim 6, characterized in that, The geometric center is the spatial centroid of the multidimensional attitude envelope region within the attitude space constructed by pitch angle, yaw angle, and roll angle; the spatial centroid is obtained by solving the mean of the spatial coordinates of the boundary feature points defining the multidimensional attitude envelope region; the attitude angle combination is the angle vector formed by the pitch angle coordinates, yaw angle coordinates, and roll angle coordinates corresponding to the spatial centroid in the attitude space; The microelectromechanical multi-axis displacement system is driven to deflect to the optical axis and align with the target point using a dual closed-loop control strategy. The dual closed-loop control strategy involves integrating MEMS displacement sensors and MEMS attitude sensors on each motion axis of the microelectromechanical multi-axis displacement system. After the control system issues a deflection command, it collects the physical coordinate data and physical deflection angle data output by the MEMS displacement sensor and MEMS attitude sensor, performs a closed-loop comparison with the theoretical coordinates and theoretical angles of the optical axis aligned with the target point, and performs physical displacement error compensation until the actual physical pose fed back by the sensor reaches the optical axis aligned with the target point.

8. The fiber micro / nano multi-station optical path coupling control method based on optical field optimization according to claim 1, characterized in that, The spatial offset of the three-dimensional extreme coordinates is the spatial Euclidean distance between the three-dimensional extreme coordinates extracted in the current iteration and the three-dimensional extreme coordinates extracted in the previous iteration in two consecutive iterations; the angular offset of the optical axis alignment attitude is the comprehensive angular vector magnitude formed by the angular differences in pitch angle, yaw angle and roll angle between the optical axis alignment attitude extracted in the current iteration and the optical axis alignment attitude extracted in the previous iteration in two consecutive iterations.

Images