A method for controlling a pile driver of an excavator based on multi-sensor fusion

Abstract

The application discloses a kind of based on multi-sensor fusion's excavator pile driver pipe pile control method, comprising the following steps: collecting environment and pipe pile data, pre-processing;Heterogeneous sensor is jointly calibrated in space-time;The positional deviation between pile driver and design stake is calculated;Based on the adaptive weighted fusion algorithm of sensor confidence, sensor data is handled, the real-time perpendicularity and perpendicularity deviation of pipe pile are calculated;Adjust the pipe pile construction of pile driver mast pose;Monitoring perpendicularity change rate, pipe pile temperature distribution and surrounding environment state, when monitoring value is abnormal, corresponding alarm and control instruction are triggered;After the construction of single pipe pile is completed, calibration error is calculated, and relevant parameters are corrected.The application solves the problems of insufficient construction precision, poor environmental adaptability, low automation level and inability to maintain long-term continuous operation precision of pipe pile construction method in complex construction scenes such as fish-light complementation.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction equipment and automated control technology, specifically to a control method for excavator pile drivers and pipe piles based on multi-sensor fusion. Background Technology

[0002] In solar-aquaculture hybrid photovoltaic power generation projects, the construction of prestressed concrete pipe pile foundations must take into account complex terrains such as water areas and wetlands. The verticality and positioning accuracy of the pipe piles directly determine the stability of the photovoltaic support installation and the power generation efficiency. Traditional construction relies on manual observation or simple inclinometers, which has problems such as verticality errors greater than 1%, large positioning deviations, high labor intensity, and low efficiency.

[0003] To improve automation levels, some existing technologies employ single or limited sensors for assistance. For example, a single lidar sensor is used to scan the profile of the pipe pile to calculate verticality, or a positioning system is combined for pile positioning guidance. However, these solutions exhibit significant shortcomings under the unique and complex operating conditions of "fishery-solar complementary" systems.

[0004] With poor environmental adaptability, lidar suffers from a sharp decline in point cloud quality or even failure under conditions of rain, fog, dust, and strong water surface reflection interference; positioning signals may also exhibit multipath errors in open water areas.

[0005] Lacking reliable redundancy and fault tolerance mechanisms, the system often relies on a single sensor, and once it fails, the entire system becomes inoperable. In addition, existing technologies mostly lack closed-loop calibration mechanisms, making accuracy prone to degradation after continuous construction, and data fusion often uses Kalman filtering, failing to consider the dynamic confidence changes of sensor data in fishery-solar complementary scenarios.

[0006] Therefore, there is an urgent need for a precise control method that is adapted to this specific scenario, integrates the advantages of multiple sensors, and has closed-loop optimization capabilities. Summary of the Invention

[0007] The purpose of this invention is to provide a control method for excavator pile drivers based on multi-sensor fusion, in order to solve the problems of insufficient construction accuracy, poor environmental adaptability, low level of automation, and inability to maintain accuracy during long-term continuous operation in complex construction scenarios such as fishery-solar complementary projects.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: A control method for excavator pile drivers based on multi-sensor fusion includes the following steps: S1: Collect environmental and pile data through a multi-sensor fusion system mounted on an excavator pile driver, and preprocess the data collected by each sensor; S2: Perform spatiotemporal joint calibration on the heterogeneous sensors in the multi-sensor fusion system to complete the initialization of the positioning system; S3: The real-time pose information of the pile driver is obtained based on the multi-sensor fusion system, and the positional deviation between the pile driver and the designed pile position is calculated in combination with the positioning system; an adaptive weighted fusion algorithm based on sensor confidence is used to process data from different sensors and calculate the real-time verticality and verticality deviation of the pipe pile. S4: Compare the position deviation and verticality deviation with their respective preset thresholds, generate control commands based on the comparison results, and drive the actuator to adjust the position of the pile driver mast to carry out pipe pile construction; S5: During the construction of pipe piles, the verticality change rate, pipe pile temperature distribution and surrounding environmental conditions are monitored in real time at a set frequency, and corresponding alarms and control commands are triggered when the monitored values ​​are abnormal. S6: After the construction of a single pipe pile is completed, calculate the calibration error based on the construction results, correct the calibration parameters of the heterogeneous sensor and the relevant parameters of the adaptive weighted fusion algorithm, and repeat steps S3-S5.

[0009] To optimize the above technical solution, the specific limitations also include: The heterogeneous sensors include lidar, millimeter-wave radar, temperature sensors, sonar detectors, infrared thermal imaging sensors, and satellite positioning modules.

[0010] Specifically, in step S1, the preprocessing of the data collected by each sensor includes: using Gaussian filtering to remove noise from the point cloud data collected by the lidar, using median filtering to eliminate jitter from the distance data collected by the millimeter-wave radar, performing outlier removal processing on the temperature data collected by the temperature sensor, and processing the data collected by the sonar detector using the sliding window averaging method.

[0011] Specifically, in step S2, the joint calibration includes: establishing the external parameter relationship between the lidar and the millimeter-wave radar using the hand-eye calibration method; establishing the mapping relationship between pixel coordinates and three-dimensional coordinates between the infrared thermal imaging sensor and the lidar using the PNP algorithm; and establishing the coordinate transformation relationship between the sonar detector and the satellite positioning module through the field coordinate calibration method to correct the signal refraction error in the aquatic environment.

[0012] Further, in step S3, the process of acquiring the real-time pose information of the pile driver based on the multi-sensor fusion system and calculating the positional deviation between the pile driver and the designed pile position using the positioning system specifically involves: acquiring the real-time geographic coordinates of the pile driver based on the initialized positioning system and the multi-sensor fusion system; using the Gauss-Kruger projection algorithm to convert the geographic coordinates of the pile driver into rectangular coordinates of the construction plane that use the same coordinate system as the designed pile position; and calculating the positional deviation between the pile driver and the designed pile position in this coordinate system.

[0013] Furthermore, in step S3, the adaptive weighted fusion algorithm is a weighted fusion algorithm based on sensor confidence, and the specific calculation formula is as follows:

[0014] in, w i Let i be the weight of the i-th sensor. c i Let be the confidence level of the i-th sensor, and n be the number of sensors; The confidence levels of the sensors include the confidence levels of lidar, millimeter-wave radar, infrared thermal imaging sensors, and sonar detectors.

[0015] Specifically, the formula for calculating the confidence level of the sensor is as follows:

[0016]

[0017]

[0018]

[0019] in, The confidence level of the lidar. N valid The number of effective contour points, N total This represents the total number of scan points; For the confidence level of millimeter-wave radar, d To detect distance, k The attenuation coefficient in the fishing-light scenario; The confidence level of the infrared thermal imaging sensor. The standard deviation of temperature in the pipe pile area. , This represents the extreme temperature value of the pipe pile. The confidence level of the sonar detector. The variance of the sonar measurement is used.

[0020] Preferably, in step S4, the threshold includes a position deviation threshold and a verticality deviation threshold; the position deviation threshold is dynamically adjusted according to the slope of the construction area; the verticality deviation threshold includes a first threshold for triggering an early warning and a second threshold for triggering an emergency adjustment, wherein the first threshold is less than the second threshold.

[0021] Furthermore, the formula for calculating the position deviation threshold is as follows:

[0022] in, This represents the positional deviation threshold in the x-direction of the coordinate system. θ represents the positional deviation threshold in the y-direction of the coordinate system, and θ is the slope angle of the construction area.

[0023] Furthermore, in step S4, the actuator is driven using a PID control algorithm, and its control formula is:

[0024] in, The degree of drive control of the actuator, This is a deviation signal. The proportional gain of the controller is optimized for the fishery-solar hybrid scenario. The integral coefficient of the controller optimized for the fishery-solar hybrid scenario. These are the differential coefficients of the controller optimized for the fishery-solar hybrid scenario.

[0025] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a multi-sensor fusion-based control method for excavator pile drivers' pipe piles. By constructing a heterogeneous redundant fusion sensing system composed of lidar, millimeter-wave radar, temperature sensors, sonar detectors, infrared thermal imaging sensors, and satellite positioning modules, it solves the reliability problem of traditional single-sensor solutions in complex environments with integrated solar and fish farming, achieving high environmental adaptability and all-weather operation. Furthermore, by employing an adaptive weighted fusion algorithm based on real-time sensor confidence, the system can dynamically allocate fusion weights according to the data quality of each sensor. Even when some sensors are interfered with, the core verticality measurement can still maintain high accuracy relying on high-confidence sensors, improving the system's robustness and verticality measurement accuracy.

[0026] Secondly, by introducing a dual-threshold triggering and hierarchical control mechanism for position deviation and verticality deviation, this invention achieves a smooth transition from monitoring and early warning to different levels of intervention, avoiding frequent over-adjustment and oscillation during the construction process. While ensuring the accuracy of planar positioning and verticality, it also improves the stability and safety of construction.

[0027] This invention performs closed-loop online calibration based on back-calculation of results after each pipe pile is constructed, correcting sensor extrinsic parameters and fusion weights in real time. This effectively compensates for sensor drift and calibration errors, ensuring the system's accuracy and stability under long-term continuous operation. Furthermore, this invention extends the control system from geometric quantity control to full-process status monitoring and safety early warning by real-time monitoring of verticality change rate, pipe pile temperature distribution, and surrounding environmental conditions during construction. This proactive intervention in abnormal situations improves system safety and enhances project predictability. Attached Figure Description

[0028] Figure 1 : A flowchart illustrating the control method for excavator pile drivers based on multi-sensor fusion of the present invention.

[0029] Figure 2 The timing diagram of the excavator pile driver pipe pile control method based on multi-sensor fusion of the present invention. Detailed Implementation

[0030] The present invention will be further described in detail below through specific embodiments, but it should not be construed as limiting the scope of the subject matter of the present invention to the following embodiments. All technologies implemented based on the above content of the present invention fall within the scope of the present invention.

[0031] In one embodiment, this invention proposes a control method for excavator pile drivers based on multi-sensor fusion, the flowchart of which is shown below. Figure 1 As shown, the entire method includes the following steps: S1: Collect environmental and pile data through a multi-sensor fusion system mounted on an excavator pile driver, and preprocess the data collected by each sensor; S2: Perform spatiotemporal joint calibration of heterogeneous sensors in the multi-sensor fusion system to complete the initialization of the positioning system; S3: The real-time pose information of the pile driver is obtained based on the multi-sensor fusion system, and the positional deviation between the pile driver and the designed pile position is calculated in combination with the positioning system; an adaptive weighted fusion algorithm based on sensor confidence is used to process data from different sensors and calculate the real-time verticality and verticality deviation of the pipe pile. S4: Compare the position deviation and verticality deviation with their respective preset thresholds, generate control commands based on the comparison results, and drive the actuator to adjust the position of the pile driver mast to carry out pipe pile construction; S5: During the construction of pipe piles, the verticality change rate, pipe pile temperature distribution and surrounding environmental conditions are monitored in real time at a set frequency, and corresponding alarms and control commands are triggered when the monitored values ​​are abnormal. S6: After the construction of a single pipe pile is completed, calculate the calibration error based on the construction results, correct the calibration parameters of the heterogeneous sensor and the relevant parameters of the adaptive weighted fusion algorithm, and repeat steps S3-S5.

[0032] The heterogeneous sensors include lidar, millimeter-wave radar, temperature sensor, sonar detector, infrared thermal imaging sensor, and satellite positioning module. Among them, the lidar is a 16-line automotive-grade lidar with a scanning frequency of 10Hz~20Hz and a horizontal field of view ≥90°, used to fit the axis of the pipe pile; the millimeter-wave radar is a 77GHz automotive-grade radar with a detection range of 0.5~10m; and the satellite positioning module is a Beidou RTK with a positioning accuracy of ±3cm.

[0033] In step S1, the preprocessing of the data collected by each sensor specifically includes: using Gaussian filtering to remove noise from the point cloud data collected by the lidar to retain the feature points of the pipe pile outline; using median filtering to eliminate jitter from the distance data collected by the millimeter-wave radar to retain effective distance information; performing outlier removal processing on the temperature data collected by the temperature sensor to retain effective temperature compensation data; processing the data collected by the sonar detector using the sliding window averaging method to improve the positioning stability in the water environment; and extracting the pipe pile area from the infrared thermal imaging data through threshold segmentation to eliminate background interference.

[0034] In step S2, the joint calibration method specifically includes: establishing the external parameter relationship between the lidar and the millimeter-wave radar using the hand-eye calibration method; establishing the mapping relationship between pixel coordinates and three-dimensional coordinates between the infrared thermal imaging sensor and the lidar using the Perspective-n-Point algorithm; and establishing the coordinate transformation relationship between the sonar detector and the satellite positioning module through the field coordinate calibration method to correct the signal refraction error in the aquatic environment.

[0035] The real-time geographic coordinates of the pile driver are obtained based on the initialized Beidou RTK system and multi-sensor fusion system. The geographic coordinates of the pile driver are converted into rectangular coordinates of the construction plane that use the same coordinate system as the designed pile position. The positional deviation between the pile driver and the designed pile position is calculated in this coordinate system. Let the real-time geographic coordinates of the pile driver be (X, Y, Z), and the coordinates of the designed pile location be (X0, Y0, Z0). Then, the position deviation is ( X, Y, The formula for calculating Z is: X0 -Y0 -Z0 Among them, the geographic coordinate transformation adopts the Gauss-Kruger projection algorithm, and the transformation error is controlled within 2cm to adapt to the large-scale construction needs of the fishery-solar complementary project. The elevation value Z is transformed from the geodetic height measured by the Beidou RTK system to the construction elevation system through elevation fitting, and Z0 is the design elevation.

[0036] An adaptive weighted fusion algorithm based on sensor confidence is used to process data from lidar, millimeter-wave radar, infrared thermal imaging sensors, and sonar detectors to calculate the real-time verticality and deviation of the pipe pile. The sensor weight calculation formula is as follows:

[0037] in, w i Let i be the weight of the i-th sensor. c i Let be the confidence level of the i-th sensor, and n be the number of sensors; Specifically, the formula for calculating sensor confidence is as follows:

[0038]

[0039]

[0040]

[0041] in, The confidence level of the lidar. N valid The number of effective contour points, N total This represents the total number of scan points; For the confidence level of millimeter-wave radar, d To detect distance, k The attenuation coefficient is used in the fishing-solar scenario. k =0.1m -1 ; The confidence level of the infrared thermal imaging sensor. σ temp The standard deviation of temperature in the pipe pile area. , This represents the extreme temperature value of the pipe pile. The confidence level of the sonar detector. δ sonar To measure the variance of sonar; When the number of effective contour points is greater than 200 per frame and the calculated confidence level of the LiDAR is not less than 0.9, the LiDAR data quality is considered excellent, and it is given a higher fusion weight, ranging from 0.7 to 0.9. When the number of effective contour points is less than 50 per frame and the calculated confidence level of the LiDAR is no greater than 0.5, it is determined that the reliability of the LiDAR data has decreased. The system reduces the weight of the LiDAR to 0.2~0.4 and correspondingly increases the weight of other complementary sensors to maintain the overall perception accuracy of the fusion system.

[0042] LiDAR scans the side of the pipe pile to acquire point cloud data, and the pipe pile axis is fitted using the RANSAC algorithm to calculate the current verticality α. Simultaneously, millimeter-wave radar acquires distance data from the top of the pipe pile to assist in correcting the axis fitting result. The formula for calculating the verticality deviation is:

[0043] in, α represents the verticality deviation, where α is the current verticality of the pipe pile.

[0044] The positional deviation and verticality deviation are compared with their respective preset thresholds, including the positional deviation threshold and the verticality deviation threshold. The positional deviation threshold is dynamically adjusted according to the slope of the construction area, and the calculation formula for the positional deviation threshold is as follows:

[0045] in, This represents the positional deviation threshold in the x-direction of the coordinate system. θ represents the positional deviation threshold in the y-direction of the coordinate system, and θ is the slope angle of the construction area.

[0046] The verticality deviation threshold includes a first threshold for triggering an early warning and a second threshold for triggering emergency adjustments. The first threshold is smaller than the second threshold. Both thresholds are derived from the structural mechanics and bearing safety factor of the pipe pile and are obtained through an engineering mechanics model. First threshold for When the verticality deviation reaches this value, the system issues a warning and prepares to intervene for adjustment; Second threshold for This threshold is a safety threshold and requires immediate intervention.

[0047] Based on the comparison results, control commands are generated to drive the actuator to adjust the mast position of the pile driver for pipe pile construction. The specific adjustment mechanism executes a three-level adjustment strategy: Level 1: and and The control decision is to maintain the state; the hydraulic servo system remains locked and does not intervene, allowing the piling operation to continue. Level 2: or or The control decision is to issue a smooth adjustment command. The hydraulic servo control unit adjusts the mast position or angle smoothly at a low speed of 0.1 cm / s, gradually correcting the deviation. The piling operation can be carried out simultaneously. Level 3: or or The control decision triggers an emergency adjustment command, immediately suspending the piling operation. The hydraulic servo control unit prioritizes adjusting the mast position at a speed of 0.3 cm / s to quickly eliminate the deviation. Once the deviation returns to the first-level state, the piling operation automatically resumes.

[0048] The actuator is driven by a PID control algorithm, and its control formula is as follows:

[0049] in, The degree of drive control of the actuator, This is a deviation signal. The proportional gain of the controller is optimized for the fishery-solar hybrid scenario. The integral coefficient of the controller optimized for the fishery-solar hybrid scenario. These are the differential coefficients of the controller optimized for the fishery-solar hybrid scenario.

[0050] During the pipe pile driving process, multiple status parameters are monitored in real time at a set frequency, and corresponding instructions are triggered when abnormalities occur. The verticality change rate is monitored, and if it exceeds the set value, the pile driving hammer force is automatically reduced. The temperature distribution of the pipe pile is monitored by infrared thermal imaging, and a high temperature warning is triggered if the local temperature is abnormal. The condition around the pile is monitored by sonar, and a dredging prompt is triggered if the siltation is too thick.

[0051] After the construction of a single pipe pile is completed, the calibration error of the sensor system is calculated based on the final construction results: the system controls the lidar to scan the pipe pile, and the actual verticality of the pipe pile is obtained through point cloud fitting. Calculate calibration error ; Based on the calibration error, the calibration parameters of the heterogeneous sensor and the relevant parameters of the adaptive weighted fusion algorithm are corrected in reverse: Correction of the extrinsic parameter matrix of lidar:

[0052] in, These are the correction values ​​for the lidar extrinsic parameter matrix. For calibration error, Let be the extrinsic parameter matrix of the lidar, representing the transformation matrix between the sensor coordinate system and the vehicle coordinate system; Corrected extrinsic parameter matrix This is used to improve the accuracy of coordinate system transformation in subsequent measurements.

[0053] The basic weighting coefficients of millimeter-wave radar in the adaptive weighted fusion algorithm are adjusted. The formula for calculating the correction amount of the millimeter-wave radar weighting coefficients is as follows:

[0054] in, Δw This is the correction amount for the weighting coefficient of millimeter-wave radar. This is for calibration error.

[0055] After calibration, the next pipe pile is constructed, and the key calibration data is stored in a historical database. This database is used for long-term trend analysis, providing data support for the weight allocation strategy of the adaptive weighted fusion algorithm and the online optimization of dual-threshold control parameters in subsequent construction, giving the system a certain degree of self-learning capability.

[0056] In addition, the system automatically records the following data: design pile location coordinates, real-time geographic coordinates of the pile driver, verticality deviation curve, sensor data fusion logs, adjustment command records, and environmental parameters. It generates a single pile construction quality report, including a pass rate determination, and exports the data through the human-computer interaction module to support project acceptance and quality traceability.

[0057] To further understand the technical solution of the present invention, a detailed description is provided in conjunction with specific embodiments: The selected machine is a 20-ton excavator pile driver, with the following specific configuration: LiDAR: RoboSense RS-HeliX16 line, installed on the left side of the mast 3m above the ground, with the scanning direction perpendicular to the side of the pipe pile; Millimeter-wave radar: Continental ARS54877GHz radar, mounted on the top right side of the mast at a 20° angle to the vertical direction; Beidou RTK: Huace T5Pro, installed on the unobstructed top of the cab, with a receiving frequency of 1Hz; Infrared thermal imaging sensor: Hikvision DS-2TD2136-6 / PA, installed in the middle of the mast, with a field of view of 30°; Temperature sensor: PT100, mounted on the lidar housing, measuring range -20℃~80℃, accuracy ±0.5℃; Sonar detector: Zhuoying ZYS-01, installed at the bottom of the fuselage, with a detection depth of 0.5m~5m; Data fusion processing unit: NVIDIA Jetson AGX Xavier, running an adaptive weighted fusion algorithm with a data processing latency of <50ms; Hydraulic servo control unit: Bosch Rexroth 4WREE6 electro-hydraulic proportional valve, which controls the mast's adjustment cylinders in four directions (left, right, forward, and backward), with a response time of <100ms; Human-computer interaction module: 10.1-inch touch screen, displaying real-time parameters and alarm information.

[0058] The application scenario is a solar-fishery complementary photovoltaic power generation project with a water area of ​​300 mu (approximately 20 hectares) and a total of 500 prestressed pipe piles (Φ500mm). Specific implementation steps are as follows: Figure 2 As shown: Before construction, a Beidou reference station was set up 800m away from the construction area, and joint calibration of various heterogeneous sensors was completed, with a calibration error of <0.5°. The operator inputs the design pile coordinates (X0=352876.231m, Y0=4836219.527m, Z0=125.63m) via the touch screen. After system initialization, the initial position deviation is calculated (ΔX=+0.82m, ΔY=-0.65m). Once the pile driver is moved to the vicinity of the construction site, the Beidou RTK system updates the coordinates in real time, and the hydraulic system automatically adjusts the position of the machine. After 15 seconds, the deviation is reduced to (ΔX=+2.3cm, ΔY=-1.8cm), meeting the positioning threshold requirements. The pipe pile is hoisted, the lidar begins scanning, and the point cloud processing module fits the pipe pile axis every 0.1 seconds. The initial verticality α = 88.5° (verticality deviation Δα = -1.5°), triggering an emergency adjustment. The data fusion processing unit assigns weights through a dynamic weighting algorithm: 0.3 for lidar, 0.6 for millimeter-wave radar, and 0.1 for sonar detector. It then sends an adjustment command, and the hydraulic servo control unit drives the cylinder to extend and retract. After 3.2 seconds, the verticality is adjusted to α=89.8° (verticality deviation Δα=-0.2°). When the pile driving operation is started, the vibratory hammer frequency is 1500 times / minute. During the operation, if there is a short-term dust storm, the effective number of laser point cloud points drops to 30 points / frame. The system automatically increases the weight of the millimeter-wave radar to 0.7 to maintain stable accuracy. Piling was driven to the design elevation (Z=118.36m), and the system recorded the final data: ΔX=+2.1cm, ΔY=-1.8cm, Δα=+0.2°, all of which meet the design requirements; Once single-pile construction is completed, closed-loop calibration is initiated, the lidar extrinsic parameter matrix is ​​corrected, construction data is saved, and a prompt is given to proceed with construction at the next pile location.

[0059] Furthermore, to verify the advantages of the multi-sensor fusion-based excavator pile driver control method for pipe piles provided by this invention, 100 pipe piles were selected at the construction site of a solar-fishery complementary photovoltaic power generation project. The method provided by this invention was compared with existing technologies and construction methods. The test results are shown in the table below: Table 1 Comparison of test results of the present invention and existing construction and distribution methods.

[0060] Test results show that the method of this invention has certain advantages over existing technologies. In terms of key accuracy indicators, the positioning deviation and verticality deviation of this invention are significantly better than traditional manual methods and the other two existing construction techniques. This demonstrates the core value of multi-sensor fusion and adaptive algorithms in improving construction accuracy. Furthermore, after constructing 100 pipe piles consecutively, the accuracy decay of this invention is far lower than other methods, indicating that the closed-loop calibration mechanism effectively guarantees the long-term stability of pipe pile construction. In addition, regarding adaptability to actual working conditions, this invention can operate without failure even in adverse weather conditions such as rain and dust, while the comparative methods show failure or lag. Simultaneously, this invention shortens the single-pile construction time to 12 minutes and increases the pass rate to 100%, achieving a comprehensive improvement in both efficiency and quality. This invention not only solves the core pain points of low accuracy and reliance on manual labor in traditional methods, but also establishes significant advantages in environmental adaptability, long-term stability, and overall cost-effectiveness, providing a reliable technical path for high-precision pile foundation construction in complex scenarios.

[0061] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent substitutions, and improvements made by those skilled in the art to the above embodiments without departing from the scope of the technical solution of the present invention, based on the technical essence of the present invention, shall still fall within the protection scope of the technical solution of the present invention.

Claims

1. A control method for excavator pile drivers based on multi-sensor fusion, characterized in that, Includes the following steps: S1: Collect environmental and pile data through a multi-sensor fusion system mounted on an excavator pile driver, and preprocess the data collected by each sensor; S2: Perform spatiotemporal joint calibration on the heterogeneous sensors in the multi-sensor fusion system to complete the initialization of the positioning system; S3: The real-time pose information of the pile driver is obtained based on the multi-sensor fusion system, and the positional deviation between the pile driver and the designed pile position is calculated in combination with the positioning system; an adaptive weighted fusion algorithm based on sensor confidence is used to process data from different sensors and calculate the real-time verticality and verticality deviation of the pipe pile. S4: Compare the position deviation and verticality deviation with their respective preset thresholds, generate control commands based on the comparison results, and drive the actuator to adjust the position of the pile driver mast to carry out pipe pile construction; S5: During the construction of pipe piles, the verticality change rate, pipe pile temperature distribution and surrounding environmental conditions are monitored in real time at a set frequency, and corresponding alarms and control commands are triggered when the monitored values ​​are abnormal. S6: After the construction of a single pipe pile is completed, calculate the calibration error based on the construction results, correct the calibration parameters of the heterogeneous sensor and the relevant parameters of the adaptive weighted fusion algorithm, and repeat steps S3-S5.

2. The multi-sensor fusion based control method for pile driving of excavator pile driver as claimed in claim 1 wherein: The heterogeneous sensors include lidar, millimeter-wave radar, temperature sensors, sonar detectors, infrared thermal imaging sensors, and satellite positioning modules.

3. The multi-sensor fusion based control method for pile driving of excavator pile driver as claimed in claim 2, wherein: In step S1, the preprocessing of the data collected by each sensor specifically includes: using Gaussian filtering to remove noise from the point cloud data collected by the lidar, using median filtering to eliminate jitter from the distance data collected by the millimeter-wave radar, performing outlier removal processing on the temperature data collected by the temperature sensor, and processing the data collected by the sonar detector using the sliding window averaging method.

4. The method for controlling excavator pile drivers based on multi-sensor fusion according to claim 2, characterized in that: In step S2, the joint calibration specifically includes: establishing the external parameter relationship between the lidar and the millimeter-wave radar using the hand-eye calibration method; establishing the mapping relationship from pixel coordinates to three-dimensional coordinates between the infrared thermal imaging sensor and the lidar using the PNP algorithm; and establishing the coordinate transformation relationship between the sonar detector and the satellite positioning module through the field coordinate calibration method to correct the signal refraction error in the aquatic environment.

5. The multi-sensor fusion based control method for pile driving of excavator pile driver as claimed in claim 1, wherein: In step S3, the process of acquiring the real-time pose information of the pile driver based on the multi-sensor fusion system and calculating the positional deviation between the pile driver and the designed pile position using the positioning system is as follows: Based on the initialized positioning system and the multi-sensor fusion system, the real-time geographic coordinates of the pile driver are acquired. The Gauss-Kruger projection algorithm is used to convert the geographic coordinates of the pile driver into rectangular coordinates of the construction plane that use the same coordinate system as the designed pile position. The positional deviation between the pile driver and the designed pile position is calculated in this coordinate system.

6. The multi-sensor fusion based control method for pile driving of excavator pile driver as claimed in claim 2, wherein: In step S3, the adaptive weighted fusion algorithm is a weighted fusion algorithm based on sensor confidence, and the specific calculation formula is as follows: wherein, w i wi is the weight of the i-th sensor, c i ci is the confidence of the i-th sensor, n is the number of sensors; The confidence levels of the sensors include the confidence levels of lidar, millimeter-wave radar, infrared thermal imaging sensors, and sonar detectors.

7. The multi-sensor fusion based control method of a pile driver of an excavator as claimed in claim 6, wherein: The specific formula for calculating the confidence level of the sensor is as follows: in, The confidence level of the lidar. N valid The number of effective contour points, N total This represents the total number of scan points; For the confidence level of millimeter-wave radar, d To detect distance, k The attenuation coefficient in the fishing-light scenario; The confidence level of the infrared thermal imaging sensor. The standard deviation of temperature in the pipe pile area. , This represents the extreme temperature value of the pipe pile. The confidence level of the sonar detector. The variance of the sonar measurement is used.

8. The method for controlling excavator pile drivers based on multi-sensor fusion according to claim 1, characterized in that: In step S4, the threshold includes a position deviation threshold and a verticality deviation threshold; the position deviation threshold is dynamically adjusted according to the slope of the construction area; the verticality deviation threshold includes a first threshold for triggering an early warning and a second threshold for triggering an emergency adjustment, wherein the first threshold is less than the second threshold.

9. The multi-sensor fusion based control method of a pile driver of an excavator as claimed in claim 8, wherein: The formula for calculating the position deviation threshold is: wherein is a position deviation threshold value in the x direction in the coordinate system, is a position deviation threshold value in the y direction in the coordinate system, and θ is a construction area slope angle.

10. The multi-sensor fusion based control method for pile driving of excavator pile driver as claimed in claim 1, wherein: In step S4, the actuator is driven using a PID control algorithm, and its control formula is as follows: in, The degree of drive control of the actuator, This is a deviation signal. The proportional gain of the controller is optimized for the fishery-solar hybrid scenario. The integral coefficient of the controller optimized for the fishery-solar hybrid scenario. These are the differential coefficients of the controller optimized for the fishery-solar hybrid scenario.

Images