A vision-based path planning method for photovoltaic panel intelligent installation robots

Abstract

The present invention provides a vision-based path planning method for a photovoltaic panel intelligent installation robot, which relates to the field of image data processing. The method comprises: using a phase-sensitive camera to collect real-time data on the photovoltaic panel storage and installation area to generate a holographic data stream; constructing a continuous phase field using a bidirectional space-time grid discretization algorithm based on a wave equation solver, combining it with time-reverse light field prediction to generate pre-compensation parameters, fusing them to form a holographic phase field model and constructing a digital elevation model; using a phase gradient tracking algorithm to avoid terrain undulations and obstacles in global optimal path planning, dynamically optimizing path curvature and movement speed; using time-reverse light field prediction to detect phase break risks in real time and generate microsecond-level correction instructions to adjust the robot's trajectory; and verifying phase deviations through differential analysis after installation, triggering time-reverse light field prediction and reinforcement learning to iteratively optimize model parameters, achieving high-precision adaptive path planning and obstacle avoidance control.

Description

Technical Field

[0001] The present invention relates to the field of image data processing, and in particular to a vision-based path planning method for a photovoltaic panel intelligent installation robot. Background Art

[0002] The demand for large-scale photovoltaic panel installation continues to grow with the rapid development of renewable energy. Traditional manual installation is inefficient and costly, making automated installation technology a key industry trend. Existing photovoltaic panel installation robots often rely on preset paths or simple environmental perception, making them difficult to adapt to complex terrain and dynamic obstacles. There is an urgent need for more accurate real-time path planning methods to improve installation efficiency and reliability.

[0003] Among existing technologies, LiDAR SLAM combined with 3D point cloud reconstruction is a common path planning solution for photovoltaic panel installation robots. This approach generates a point cloud map by scanning the environment and plans an obstacle avoidance path. Another solution is based on the visual inertial odometry of an RGB-D camera, which achieves pose estimation and local path optimization through feature matching and depth information fusion. In addition, some studies use genetic algorithms or A* algorithms to search for the global optimal path in rasterized maps, combined with dynamic window methods to achieve real-time obstacle avoidance.

[0004] The shortcomings of existing technologies are: LiDAR solutions are susceptible to interference in strong light environments, point cloud data processing is time-consuming and difficult to capture subtle phase changes; RGB-D cameras are limited by lighting conditions, and depth information is easily distorted in scenes with photovoltaic panel reflections; grid map-based algorithms lack the ability to model high-precision terrain and cannot effectively handle path deviations caused by phase gradients; and it is difficult to meet the high-precision requirements of photovoltaic panel installation. Summary of the Invention

[0005] (1) Technical problems solved

[0006] In response to the shortcomings of the existing technology, the present invention provides a vision-based path planning method for a photovoltaic panel intelligent installation robot to solve the problems of sensors being easily interfered with in strong light environments, resulting in distortion of point cloud data, inaccurate depth information collection in photovoltaic panel reflective scenes, and insufficient ability of rasterized map algorithms to perform high-precision terrain modeling.

[0007] (2) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a vision-based path planning method for a photovoltaic panel intelligent installation robot, comprising: S1: using phase-sensitive cameras deployed on the robot's mechanical arm and mobile chassis to collect light intensity and phase information of the photovoltaic panel storage area and installation area in real time, and embedding the robot's 6-degree-of-freedom posture data and timestamp into the light intensity and phase information to generate a holographic data stream with spatiotemporal labels, wherein the end of the robot's mechanical arm is equipped with an end effector;

[0009] S2: Based on the holographic data stream, a bidirectional space-time grid discretization algorithm is used in the wave equation solver to solve the light field wave equation to generate a continuous phase field, and a light field distribution map is constructed. Then, a dimensionality reduction mapping is performed in combination with the robot motion tangent plane. At the same time, the future phase distortion is deduced through time-reversed light field prediction to generate pre-compensation parameters;

[0010] S3: Fusing the continuous phase field with the pre-compensation parameters to construct a holographic phase field model, converting the phase information into an elevation value through a preset phase difference and incident angle conversion relationship, constructing a digital elevation model of the installation area, and identifying the three-dimensional outline of obstacles based on diffraction ripples and mutation bands in the holographic phase field, and marking the identified obstacle area as a dynamic restricted area;

[0011] S4: Setting the target location for photovoltaic panel installation and the phase gradient tolerance parameter, initializing the path planning algorithm in combination with the holographic phase field model, generating a global optimal path along the optimal phase gradient direction, and then transmitting it to the phase gradient tracking algorithm. The phase gradient tracking algorithm controls the robot to move along the phase gradient direction through a closed-loop feedback mechanism. The phase gradient tolerance parameter is the allowable phase deviation threshold.

[0012] S5: The robot is controlled to move along the direction of the light field energy flow through a phase gradient tracking algorithm, and the path curvature and movement speed are dynamically optimized according to the size and weight of the photovoltaic panels;

[0013] S6: Based on the time-reversed light field prediction results, the phase break risk on the future path is detected in real time, the phase modulation parameters are dynamically adjusted to reshape the light field distribution map, and the robot's motion trajectory is corrected by generating microsecond-level motion trajectory correction instructions;

[0014] S7: After the installation is completed, the actual phase field data is collected through the phase-sensitive camera and differential analysis is performed with the holographic phase field model. If the deviation exceeds the set phase gradient tolerance threshold, the inverse light field is triggered to replan the path and the phase gradient mapping parameters and obstacle contour recognition parameters of the holographic phase field model are iteratively optimized through the reinforcement learning algorithm.

[0015] Preferably, the phase break refers to the phase transition phenomenon caused by propagation interference or structural mutation in the continuous phase field, which is manifested as a mutation and discontinuous area in the light field distribution map; the phase break risk refers to the probability value of phase break occurring in a certain area in the predicted path, reflecting the potential impact of environmental disturbance or path complexity on the continuity of the phase field; the phase break risk value is used to quantify the probability of phase break occurring in a certain grid node in the predicted path; the phase break zone refers to the spatial area where a break may be formed in the continuous phase field in the prediction, and is a connected area formed by the expansion of the phase mutation point; the break zone position refers to the four-dimensional coordinate node in the predicted path, which records the time and space position where the phase break risk exists; the predicted path refers to the spatial trajectory that the robot may execute on the future time axis based on the current state of the robot and the set installation target position, through time-reverse light field prediction, and combined with the wave equation solver to perform four-dimensional deduction of the continuous phase field.

[0016] Preferably, the phase-sensitive camera is deployed at key positions of the robot's mechanical arm and mobile chassis, covering a 360-degree field of view of the photovoltaic panel storage area and the installation area, ensuring that light intensity and phase information are collected without blind spots; the front end of the camera's photosensitive element is integrated with a micro-nano phase modulation layer composed of a piezoelectric ceramic array, and the micro-nano phase modulation layer receives a voltage control signal from the processing unit and dynamically adjusts the phase offset of the incident light to offset ambient light interference; the camera is connected to the robot's IMU / odometer through a nanosecond synchronization bus to obtain the robot's 6-DOF posture data and nanosecond position data in real time. Synchronous acquisition and embedding of second-level timestamps; the collected light intensity, phase offset, and posture data are synchronized with the nanosecond timestamps through an embedded processor to generate a holographic data stream with spatiotemporal tags; the spatiotemporal tags are differentially compressed through an entropy coding algorithm to remove redundant information between adjacent timestamps and spatial positions. The compressed data stream is transmitted to the processing unit of the wave equation solver and phase gradient tracking algorithm in a low-latency manner through a dedicated light field channel to ensure data real-time and integrity; the 6-DOF posture includes X, Y, and Z axis displacements as well as pitch, yaw, and roll angles.

[0017] Preferably, the wave equation solver maps the light intensity and phase information in the holographic data stream to a space-time grid, and each grid node stores the phase field value at a specific time and position; through a bidirectional space-time discretization algorithm, the forward time axis is used to solve the current light field distribution, and the reverse time axis is used to reversely deduce the future phase distortion from the target end point; the principal component characteristics of the phase field are extracted through a principal component analysis algorithm, projected onto the robot motion tangent plane, and a two-dimensional dimensionality reduction mapping is generated; based on the reverse time axis deduction result, the position and compensation amount of the phase fracture zone on the future path are calculated, and pre-compensation parameters including phase modulation intensity, compensation direction and time window are generated; the pre-compensation parameters are generated through GPU parallel computing acceleration and transmitted to the micro-nano phase modulation layer in real time, and the piezoelectric ceramic unit adjusts the piezoelectric control signal according to the parameters, changes the phase offset of the incident light, offsets the predicted phase distortion, and maintains the space-time continuity of the light field. The generation formula of the pre-compensation parameters is: ,in is the piezoelectric ceramic driving voltage; is the phase quantity to be compensated; is the wavelength of light; k is the sensitivity coefficient of the piezoelectric ceramic, and its specific value is 0.5μm / V.

[0018] Preferably, the continuous phase field data and pre-compensation parameters are fused through a weighted superposition algorithm, with the weights determined by the phase gradient stability, to generate a holographic phase field model including terrain height, obstacle outline, and phase gradient direction; based on the linear relationship between phase difference and incident angle, the phase difference between adjacent positions is converted into height difference values to construct a digital elevation model; by detecting the annular diffraction ripples in the phase field, the correlation between its wavelength offset and the surface roughness of the obstacle is analyzed, and the material hardness and geometric dimensions of the obstacle are inferred by combining the refractive index change of the phase mutation band; the obstacle information is marked as a dynamic restricted area, including a coordinate range, shape boundary, and hazard level, and the dynamic restricted area data is input into the path planning algorithm in real time to trigger the rearrangement of path nodes and the update of the obstacle avoidance strategy; the digital elevation model is: ; Where h(x,y) is the terrain height value; is the phase difference between adjacent positions; is the conversion coefficient, with a specific value of 0.1 mm / rad; the hazard level is calculated using a weighted scoring model, and the hazard level R is defined as: , where the weight coefficients of w1, w2, and w3 are 0.5, 0.3, and 0.2 respectively, and material > geometry > dynamics; , , The standardized scores of each dimension ∈ [0,1] are used. The thresholds for the danger level division are: when 0≤R<0.4, it is low danger, the risk can be ignored, and the path planning only needs to be fine-tuned; when 0.4≤R<0.7, it is medium danger, and the obstacle avoidance strategy needs to be triggered to generate a detour path; when R≥0.7, it is high danger, and the movement must be stopped immediately and the global optimal path must be replanned.

[0019] Preferably, the remote monitoring center sets the installation target coordinates and phase gradient tolerance threshold, and the phase gradient tolerance threshold is 0.05rad, which is used to define the upper limit of the allowable phase deviation; the path planning algorithm combines the size and weight of the photovoltaic panel, initializes the search conditions, and generates a global optimal path along the phase gradient descent direction in the holographic phase field model; the path avoids terrain undulations and dynamic restricted areas, and is decomposed into multiple phase gradient nodes through a discretization algorithm, each node containing a target phase value, a movement speed and a tolerance range; the phase gradient tracking algorithm obtains the current phase value of the robot in real time through a closed-loop feedback system, and calculates the difference with the phase value of the target node. If the difference exceeds the tolerance threshold, the tolerance threshold is 10cm, then the angular velocity of the robot arm joint and the movement direction of the chassis are adjusted so that the robot moves along the corrected phase gradient direction until it reaches the target node.

[0020] Preferably, the manipulator dynamics model predefines the mass distribution, inertia matrix and torque limit of each joint, and calculates the load torque of the end effector in combination with the weight of the photovoltaic panel; during the path planning process, the path curvature radius is dynamically adjusted according to the real-time load torque to ensure that the joint torque is always lower than the safety torque threshold of 50Nm; the spacing between adjacent nodes is adjusted in real time based on the phase gradient change rate, and the spacing is reduced to improve control accuracy when the gradient changes drastically, and the spacing is increased to improve motion efficiency when the gradient is gentle; the closed-loop feedback system continuously monitors the manipulator posture and chassis motion state, dynamically optimizes the path curvature and speed parameters, and achieves smooth motion under high load; the formula of the manipulator dynamics model is: , where M is the mass matrix, C is the Coriolis force term, G is the gravity term, r is the joint torque, and q is the joint angle. is the angular acceleration.

[0021] Preferably, the bidirectional space-time grid discretization algorithm discretizes the light field wave equation into a four-dimensional grid, and each grid node stores the phase value of the spatial coordinates and the timestamp; the four-dimensional grid is composed of three-dimensional spatial coordinates (X, Y, Z) and the time axis (T), and includes discretized nodes of the forward time axis and the reverse time axis; based on the dynamic restricted area data and the historical fracture probability, the phase fracture risk value on the future path is calculated in real time; if the risk value exceeds the preset threshold, the pre-compensation mechanism is triggered, and the piezoelectric control signal of the piezoelectric ceramic unit is adjusted through the micro-nano phase modulation layer to inject the compensating phase offset to offset the predicted fracture effect; at the same time, the wave equation solver generates microsecond trajectory correction instructions to control the chassis and the robotic arm to coordinately adjust the movement direction and speed to bypass the fracture area and ensure that the actual trajectory is consistent with the global optimal path. The formula of the bidirectional space-time grid discretization equation is: ;in is the phase value of the space-time grid node (i, j) at time step n; c is the speed of light, which is related to the propagation speed of the light field; Δt and Δx are the time and space discretization steps; it is discretized into a four-dimensional space-time grid as follows: ,in is the light field phase value at the spatial position (i, j, k) in the nth time step, i, j, k are spatial grid indexes corresponding to the x, y, z coordinate axes, n is the time step index, c is the speed of light, Δt is the reverse time step of 0.01s; Δx, Δy, Δz are spatial grid resolutions; i, j, k are spatial grid indexes; n is the time step index; the specific function is to reversely deduce the light field phase distortion distribution from the target end point and predict the future fracture zone position, such as t=tcurrent+0.5s, where t is the time point, generally used to represent the future prediction step; tcurrent is the current time stamp, which serves as the reference point for time reverse deduction; and generates pre-compensation parameters; the phase fracture risk value formula is as follows: ,in represents the probability value of phase rupture occurring at the four-dimensional coordinate point; M is the cluster number of historical rupture events, which is the number of different types of rupture events clustered from historical data; is the weight of the mth type of event, ranging from [0,1], and =1; is the current observed phase value; The phase value predicted by the model; σ m is the Gaussian kernel bandwidth of the mth type event; the dynamic update rule is to record the location, time and phase deviation of the historical fracture event; the current phase field and the model prediction value Compare and calculate the Mahalanobis distance: If Dm<3σm, the current point is considered to belong to the mth type of fracture event. is the covariance matrix of the mth type fracture event characteristics, which is used to reflect the distribution shape; the phase fracture risk value The value range of is [0,1], the higher the value, the greater the probability of phase break; if When the preset threshold value exceeds 0.5, the corresponding obstacle avoidance or pre-compensation mechanism is triggered; ≥0.5, the trajectory correction is triggered.

[0022] Preferably, after the installation is completed, the actual phase field data of the installed photovoltaic panel area is re-collected by a phase-sensitive camera, and a pixel-by-pixel differential comparison is performed with the holographic phase field model to calculate the root mean square value of the phase difference and store it in the historical error data; if the root mean square value exceeds the tolerance threshold of 10 cm, the source of the error is traced back through the time-reverse light field prediction equation to locate the phase gradient mapping error or obstacle recognition deviation in the model; the boundary conditions of the light field wave equation are adjusted to optimize the holographic phase field model, and the boundary conditions are the initial phase distribution of the light field wave equation and the robot motion tangent plane constraint. The robot motion tangent plane is the local coordinate system projection plane of the motion trajectory of the end effector of the manipulator, which is dynamically determined by the robot's 6-degree-of-freedom posture data; the weight coefficient of the phase gradient tracking algorithm is optimized by a reinforcement learning algorithm. The reinforcement learning algorithm iteratively updates the weight coefficient of the phase gradient tracking algorithm based on the historical error data set, adjusts the path search priority and the tolerance allocation strategy, and gradually reduces the phase deviation of subsequent tasks; the rule for updating the weight of the reinforcement learning algorithm is:

[0023]

[0024] in is the state and action value function, s is the current state, a is the current action; is the learning rate, the default is 0.01; is the next action state, For the next possible action, is the discount factor; r is the reward function, which is negatively correlated with the phase deviation.

[0025] Preferably, the piezoelectric ceramic unit of the micro-nano phase modulation layer receives a piezoelectric control signal corresponding to a pre-compensation parameter, and the voltage intensity is linearly related to the phase compensation amount. The optical path length is changed by ceramic deformation to achieve precise adjustment of the phase offset of the incident light. When collecting light intensity and phase data, the camera synchronously receives the 6-degree-of-freedom posture data of the IMU / odometer, and ensures the precise alignment of the space-time tags through the hardware clock synchronization mechanism. The compressed holographic data stream is transmitted through a dedicated channel with a fixed priority to avoid data packet loss or delay, thereby supporting the efficient operation of subsequent algorithms.

[0026] Preferably, the digital elevation model marks terrain undulations by detecting areas with steep increases in phase gradient, calculates altitude values based on the conversion relationship between phase difference and incident angle, and generates a high-precision three-dimensional terrain map; analyzes the wavelength offset of the annular diffraction ripples in the phase field in real time, infers the surface roughness of the obstacle, and determines the material properties through the refractive index change of the phase mutation band; the obstacle information is encapsulated as dynamic restricted area data, including coordinate boundaries, shape parameters and hazard levels, which is transmitted to the path planning algorithm in real time, and the restricted area list is dynamically updated to ensure that the robot's motion path always avoids the latest identified dangerous areas.

[0027] Preferably, the phase gradient tracking algorithm continuously monitors the deviation between the current phase value and the target node through a closed-loop feedback system. If the deviation exceeds the limit, the coordinated adjustment of the robot arm joint angle and the chassis speed is triggered, and the adjustment amount is determined by the deviation ratio. In the post-installation verification stage, if the deviation between the actual phase field and the model continues to exceed the limit, the reinforcement learning algorithm optimizes the weight coefficient of the phase gradient tracking algorithm based on historical error data, and adjusts the gradient weight and tolerance allocation ratio of the path search. Through multiple rounds of iterative training, the model gradually adapts to dynamic changes in the environment, thereby improving the robustness of path planning and installation accuracy.

[0028] Preferably, the path planning method is to use a bidirectional space-time grid discretization algorithm to decompose the light field wave equation into discretized grid nodes of the forward time axis and the reverse time axis; the forward time axis is used to solve the light field phase distribution at the current moment, and the reverse time axis is used to reversely deduce the phase distortion distribution on the future path based on the target end point coordinates, and predict the position, size and distortion of the phase fracture zone; the reverse deduction process is accelerated by GPU parallel computing to generate pre-compensation parameters including phase modulation intensity, compensation direction and time window; the parameters are transmitted in real time to the micro-nano phase modulation layer of the phase-sensitive camera through a dedicated data channel, which is composed of an array of piezoelectric ceramic units, each unit receives a piezoelectric control signal corresponding to the compensation amount, changes the optical path length of the incident light through the deformation of the piezoelectric ceramic, and dynamically adjusts the phase offset; this mechanism realizes real-time cancellation of future phase distortion, ensures the spatiotemporal continuity of light field data acquisition, and avoids the lag problem of traditional single time axis prediction, significantly improving the accuracy and real-time performance of path planning.

[0029] (3) Beneficial effects

[0030] The present invention provides a vision-based path planning method for a photovoltaic panel intelligent installation robot. It has the following beneficial effects:

[0031] 1. Real-time acquisition of light intensity and phase information by a phase-sensitive camera, combined with a bidirectional space-time grid discretization algorithm, generates a high-precision holographic phase field model. Dynamically adjust the path using closed-loop feedback and pre-compensation parameters to improve path planning accuracy and real-time performance in complex environments. Enhance the safety and stability of the installation process through phase gradient tolerance threshold constraints and dynamic restricted area avoidance.

[0032] 2. Based on the time-reverse light field prediction, the system deduces future phase distortion and corrects the motion trajectory in real time to offset the risk of phase breakage. It integrates reinforcement learning algorithms to iteratively optimize model parameters, combines dynamic obstacle contour recognition with terrain height mapping, and improves the system's adaptability and robustness. Through microsecond-level trajectory correction instructions and load torque constraints, it ensures the robot's motion efficiency and equipment safety. DETAILED DESCRIPTION

[0033] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0034] An embodiment of the present invention provides a vision-based path planning method for a photovoltaic panel intelligent installation robot, comprising: a robot mobile chassis and a robotic arm cooperate to enter a photovoltaic panel installation area, and multiple phase-sensitive cameras deployed thereon are synchronously started, covering a 360-degree field of view to collect light intensity and phase information; the micro-nano phase modulation layer of the camera receives a piezoelectric control signal corresponding to a pre-compensation parameter through a piezoelectric ceramic array, dynamically adjusts the phase offset of the incident light, and offsets environmental interference; the camera and the robot IMU / odometer are connected through a nanosecond-level synchronization bus, and 6-degree-of-freedom posture data is collected in real time. The data is strictly aligned with the light intensity, phase offset, and nanosecond-level timestamps through an embedded processor to generate a holographic data stream with spatiotemporal labels; the spatiotemporal labels are differentially compressed through an entropy coding algorithm to remove redundant data between adjacent times and positions, and the compressed data stream is transmitted to a wave equation solver at a fixed frame rate through a dedicated light field channel to ensure high real-time input.

[0035] After receiving the holographic data stream, the wave equation solver uses a bidirectional space-time grid discretization algorithm to decompose the light field wave equation into four-dimensional grid nodes. The forward time axis is used to solve the current light field phase distribution, and the reverse time axis is used to reversely deduce the future phase distortion from the path planning target end point; the principal component characteristics of the phase field are extracted through the principal component analysis algorithm, and projected onto the robot motion tangent plane to form a two-dimensional dimensionality reduction mapping, thereby simplifying subsequent path planning calculations; GPU parallel computing accelerates the phase distortion prediction of the reverse time axis, and generates pre-compensation parameters including the fracture zone position, compensation amount and phase modulation parameters; the parameters are transmitted to the micro-nano phase modulation layer in real time through a dedicated channel, and the piezoelectric ceramic unit adjusts the deformation according to the piezoelectric control signal intensity, accurately corrects the incident light phase offset, offsets the predicted phase fracture interference, and ensures the spatiotemporal continuity of the light field data.

[0036] The continuous phase field data and pre-compensation parameters are fused through a weighted superposition algorithm, with weights dynamically assigned by the phase gradient stability, to generate a holographic phase field model that includes terrain height, obstacle contours, and phase gradient direction. Based on a preset linear conversion relationship between phase difference and incident angle, the phase differences between adjacent areas are mapped to height values, and a digital elevation model of the installation area is constructed pixel by pixel. At the same time, the annular diffraction ripples in the phase field are detected, and the correlation between their wavelength offset and the surface roughness of the obstacle is analyzed. The three-dimensional contours and material properties of the obstacle are inverted by combining the refractive index changes in the phase mutation band. The obstacle information is marked as a dynamic restricted area, including coordinate boundaries, shape parameters, and hazard level, and transmitted to the path planning algorithm in real time to trigger an update of the obstacle avoidance strategy.

[0037] The target coordinates of photovoltaic panel installation and the phase gradient tolerance threshold are set through the remote monitoring center. The path planning algorithm combines the size and weight parameters of the photovoltaic panels to generate a global optimal path along the direction of the steepest descent of the phase gradient in the holographic phase field model. The path must avoid the terrain undulations and dynamic restricted areas in the digital elevation model. The path is decomposed into multiple phase gradient nodes through a discretization algorithm. Each node contains the target phase value, movement speed and allowable deviation range. The phase gradient tracking algorithm obtains the current phase value of the robot in real time through a closed-loop feedback system, and calculates the difference between it and the phase value of the target node. If the deviation exceeds the tolerance threshold, the angular velocity of the robot arm joint and the movement direction of the chassis are adjusted to make the robot move along the corrected phase gradient direction.

[0038] The robotic arm dynamics model predefines the mass, inertia matrix and torque limits of each joint, and calculates the real-time load torque of the end effector based on the weight of the photovoltaic panel. During the path planning process, the path curvature radius is dynamically adjusted according to the load torque to ensure that the torque of each joint is always below the safety threshold. At the same time, the spacing between adjacent nodes is adjusted in real time based on the phase gradient change rate. When the gradient changes drastically, the spacing is reduced to improve control accuracy, and when the gradient is gentle, the spacing is increased to improve motion efficiency. The closed-loop feedback system continuously monitors the robotic arm posture and chassis motion status, dynamically optimizes the path curvature and speed parameters, and ensures smooth movement under high load.

[0039] The light field wave equation is then discretized into a four-dimensional grid through a bidirectional space-time grid discretization algorithm. Based on the dynamic restricted area data and historical fracture probabilities, the phase fracture risk value on the future path is calculated in real time. If the risk value exceeds the preset threshold, the pre-compensation mechanism is triggered. By adjusting the piezoelectric control signal injection compensation parameters of the micro-nano phase modulation layer, the current light field distribution map is reshaped to offset the future fracture risk. At the same time, the wave equation solver generates microsecond-level motion trajectory correction instructions to control the chassis and the robotic arm to coordinately adjust the motion direction and speed to bypass the fracture area and ensure that the actual trajectory is consistent with the global optimal path.

[0040] After the installation is completed, the phase-sensitive camera re-collects the actual phase field data, performs a pixel-by-pixel differential comparison with the holographic phase field model, and calculates the root mean square value of the phase difference; if the deviation exceeds the tolerance threshold, the source of the error is traced through the time-reverse light field prediction equation, the phase gradient mapping error or obstacle recognition deviation is located, and the initial phase distribution of the wave equation and the motion tangent plane constraint parameters are adjusted; the reinforcement learning algorithm is based on the historical error data set, iteratively optimizes the weight coefficient of the phase gradient tracking algorithm, adjusts the gradient weight and tolerance allocation ratio of the path search, and gradually reduces the phase deviation of subsequent tasks.

[0041] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A vision-based photovoltaic panel intelligent installation robot path planning method, characterized in that: include: S1: Phase-sensitive cameras deployed on the robot's mechanical arm and mobile chassis collect light intensity and phase information from the photovoltaic panel storage and installation areas in real time, and embed the robot's 6-DOF posture data and timestamp into the light intensity and phase information to generate a holographic data stream with spatiotemporal tags. The end of the robot's mechanical arm is equipped with an end effector; the front end of the phase-sensitive camera is equipped with a micro-nano phase modulation layer, which dynamically adjusts the phase offset of the incident light by receiving a piezoelectric control signal. S2: Based on the holographic data stream, a bidirectional space-time grid discretization algorithm is used in the wave equation solver to solve the light field wave equation to generate a continuous phase field, and a light field distribution map is constructed. Then, a dimensionality reduction mapping is performed in combination with the robot motion tangent plane. At the same time, the future phase distortion is deduced through time-reversed light field prediction to generate pre-compensation parameters; S3: Fusing the continuous phase field with the pre-compensation parameters to construct a holographic phase field model, converting the phase information into an elevation value through a preset phase difference and incident angle conversion relationship, constructing a digital elevation model of the installation area, and identifying the three-dimensional outline of obstacles based on diffraction ripples and mutation bands in the holographic phase field, and marking the identified obstacle area as a dynamic restricted area; S4: Setting the target location for photovoltaic panel installation and the phase gradient tolerance parameter, initializing the path planning algorithm in combination with the holographic phase field model, generating a global optimal path along the optimal phase gradient direction, and then transmitting it to the phase gradient tracking algorithm. The phase gradient tracking algorithm controls the robot to move along the phase gradient direction through a closed-loop feedback mechanism. The phase gradient tolerance parameter is the allowable phase deviation threshold. S5: The robot is controlled to move along the direction of the light field energy flow through a phase gradient tracking algorithm, and the path curvature and movement speed are dynamically optimized according to the size and weight of the photovoltaic panels; S6: Based on the time-reversed light field prediction results, the phase break risk on the future path is detected in real time, the phase modulation parameters are dynamically adjusted to reshape the light field distribution map, and the robot's motion trajectory is corrected by generating microsecond-level motion trajectory correction instructions; S7: After the installation is completed, the actual phase field data is collected through the phase-sensitive camera and differential analysis is performed with the holographic phase field model. If the deviation exceeds the set phase gradient tolerance threshold, the inverse light field is triggered to replan the path and the phase gradient mapping parameters and obstacle contour recognition parameters of the holographic phase field model are iteratively optimized through the reinforcement learning algorithm.

2. A vision-based photovoltaic panel intelligent installation robot path planning method according to claim 1, characterized in that: The S1 step includes the following operations: S11: Deploy multiple phase-sensitive cameras on the robotic arm and mobile chassis to cover a 360-degree field of view of the photovoltaic panel storage area and installation area. The front end of the photosensitive element of the phase-sensitive camera is integrated with a micro-nano phase modulation layer composed of a piezoelectric ceramic array, which can dynamically adjust the phase offset of the incident light according to the piezoelectric control signal and is connected to the robot IMU / odometer via a nanosecond synchronization bus; S12: Dynamically adjust the phase offset of the incident light according to the pre-compensation parameters through the micro-nano phase modulation layer integrated in the front end of the camera photosensitive element, and simultaneously record the light intensity and phase offset, and embed the 6-DOF pose data and nanosecond timestamp provided by the robot IMU / odometer; S13: Using an entropy coding algorithm to perform differential compression on the spatiotemporal tag to reduce data transmission redundancy, and transmitting the compressed holographic data stream to a processing unit through a dedicated light field data channel; The processing unit is an operating platform for the wave equation solver and the phase gradient tracking algorithm.

3. The vision-based photovoltaic panel intelligent installation robot path planning method according to claim 2, characterized in that: The S2 step includes the following operations: S21: Inputting the light intensity and phase information in the holographic data stream into a wave equation solver, discretizing the light field wave equation using a bidirectional space-time grid discretization algorithm, projecting the high-dimensional phase field data onto a two-dimensional motion tangent plane using a principal component analysis algorithm to complete dimensionality reduction mapping, and projecting the discretized data onto the robot motion tangent plane; S22: Based on the target endpoint of the path planning, the phase distortion distribution is deduced through time-reversed light field prediction to generate pre-compensation parameters including the fracture zone position, compensation amount, and phase modulation parameters. The solution process is accelerated through GPU parallel computing. The phase modulation parameters are used to dynamically adjust the phase offset of the incident light through the micro-nano phase modulation layer. S23: Input the pre-compensation parameter into the micro-nano phase modulation layer of the phase-sensitive camera, adjust the current phase offset according to the pre-compensation parameter through the piezoelectric control signal, and offset the interference of future light field phase break on current data acquisition through time-reverse light field prediction propagation.

4. The vision-based photovoltaic panel intelligent installation robot path planning method according to claim 1, characterized in that: The S3 step includes the following operations: S31: fusing the continuous phase field with the pre-compensation parameters through a weighted superposition algorithm to generate a holographic phase field model containing phase gradient, obstacle profile, and terrain height information, and extracting the phase gradient direction as a motion reference for the phase gradient tracking algorithm; S32: Marking terrain undulations based on the phase gradient steep increase area, solving the height value based on the preset phase difference and incident angle conversion relationship, and constructing a digital elevation model; S33: Detect the annular diffraction ripples and phase mutation bands in the phase field, invert the three-dimensional outline and material properties of the obstacle by analyzing the wavelength offset of the annular diffraction ripples and the refractive index change characteristics of the phase mutation band, mark the obstacle area as a dynamic restricted area, update the path avoidance strategy in real time, and encapsulate the restricted area coordinates, shape parameters and hazard level data contained in the dynamic restricted area into structured data.

5. The vision-based photovoltaic panel intelligent installation robot path planning method according to claim 1, characterized in that: The S4 step includes the following operations: S41: The installation target coordinates and phase gradient tolerance threshold are set through the remote monitoring center, and the photovoltaic panel size and weight parameters are transmitted to the path planning algorithm to dynamically constrain the path curvature and movement speed; S42: generating a global optimal path along the direction of steepest descent of the phase gradient in the holographic phase field model, avoiding terrain undulations and dynamic restricted areas; S43: Decomposing the global optimal path into discrete phase gradient nodes, each node including a target phase value, a movement speed and an allowable deviation range.

6. The vision-based photovoltaic panel intelligent installation robot path planning method according to claim 1, characterized in that: The S5 step includes the following operations: S51: controlling the robot to move along the phase gradient direction by using the phase gradient tracking algorithm, the robot arm end effector comparing the phase gradient tolerance of the current phase value with the target node phase value in real time, and adjusting the robot arm posture by using the closed-loop feedback mechanism of the phase gradient tracking algorithm; S52: adjusting the path curvature radius according to the weight of the photovoltaic panel and the dynamic model of the robotic arm so that the load torque of the robotic arm does not exceed a preset safety torque threshold; the dynamic model of the robotic arm is a predefined mathematical model including the mass, inertia, joint torque constraints and kinematic parameters of the robotic arm, and is used to calculate the motion trajectory and load torque of the robotic arm end effector; S53: In a motion scenario, the phase gradient tracking algorithm monitors the phase gradient change rate in real time and dynamically adjusts the spacing of the phase gradient nodes according to a preset ratio to improve control accuracy.

7. The vision-based photovoltaic panel intelligent installation robot path planning method according to claim 1, characterized in that: The S6 step includes the following operations: S61: discretizing the light field wave equation into a four-dimensional grid including a forward time axis and a reverse time axis using the bidirectional space-time grid discretization algorithm, and calculating the phase break probability on the future path in real time based on the dynamic forbidden zone; S62: If the phase break probability exceeds the break probability threshold, the light field distribution spectrum is reshaped to offset future phase breaks by adjusting the piezoelectric control signal injection pre-compensation parameters of the micro-nano phase modulation layer; S63: Generate a motion trajectory correction instruction through the wave equation solver to control the robot to collaboratively adjust the motion trajectory to avoid obstacles so that it is consistent with the global optimal path.

8. The vision-based photovoltaic panel intelligent installation robot path planning method according to claim 1, characterized in that: The S7 step includes the following operations: S71: Collecting actual phase field data of the installed photovoltaic panel area through a phase-sensitive camera, calculating the root mean square value of the phase difference between the actual phase field data and the holographic phase field model, and storing the data in the historical error data; S72: If the phase difference exceeds the phase gradient tolerance threshold, deduce the error source through the time inversion light field prediction, and adjust the boundary conditions of the light field wave equation to correct the phase gradient mapping relationship and obstacle contour recognition parameters of the holographic phase field model; The weight coefficients of the phase gradient tracking algorithm are iteratively updated through a reinforcement learning algorithm.

Images