Indoor narrow space large-scale equipment self-adaptive installation method based on multi-source information fusion

Abstract

The invention relates to the technical field of large-scale equipment installation, and provides an indoor narrow space large-scale equipment adaptive installation method based on multi-source information fusion, comprising the following steps: providing an integrated laser guiding system and a visual identification system, the visual identification system obtaining real-time image data of equipment through an image sensor; the visual identification system compares the position of the equipment with a target position to generate spatial position deviation information of the equipment, the laser guiding system and the visual identification system are fused to generate and fuse feedback data, and current position data and angle data of the equipment are calculated in real time. According to the method for installing the large equipment in place in the indoor narrow space, the space position deviation information of the equipment can be obtained in real time through the integrated laser guiding system and the visual recognition system, the current position and angle of the equipment are accurately calculated through feedback data, and therefore accurate butt joint of the equipment is achieved. Compared with the prior art, errors caused by manual adjustment are avoided.

Description

Technical Field

[0001] This invention relates to the field of large equipment installation technology, specifically to an adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion. Background Technology

[0002] In modern engineering construction and industrial manufacturing, the installation and positioning of equipment in confined spaces has always been a technical challenge. With increasingly compact building and factory workshop spaces, how to handle, position, and install large equipment within limited spaces has become a crucial issue in equipment installation. Currently, mainstream technologies for installing large equipment in confined spaces include hoisting, rail transport, and thrust devices. Hoisting is the most common technique, typically using cranes or hoisting equipment to lift the equipment to a designated location. However, in confined spaces, hoisting methods are often limited by space constraints and difficult to implement. Meanwhile, rail transport technology uses rails and trolley systems to push equipment to a predetermined position. While effective under certain conditions, this method requires precise rail placement, and for larger equipment, stability must be ensured during transport; otherwise, tilting or shifting can easily occur, leading to equipment damage or installation failure. Thrust devices use mechanical thrust to precisely position the equipment, suitable for handling smaller equipment, but less applicable to ultra-large equipment.

[0003] The core deficiency in existing technologies lies in the precise installation and positioning of equipment. In confined spaces, due to limited operating space, existing technologies often struggle to guarantee accurate docking and placement of equipment. Traditional hoisting, rail transport, and thrust device technologies struggle to achieve precise control of the equipment during installation. This deficiency leads to lengthy installation times, low efficiency, and potential deviations during installation, causing difficulties in subsequent commissioning or functional failure, and in severe cases, potentially resulting in equipment damage.

[0004] Therefore, there is an urgent need for a method that can achieve high-precision, adaptive, and fully controllable installation of large equipment in confined spaces, in order to improve installation efficiency, ensure equipment safety, and reduce the uncertainty caused by human intervention. Summary of the Invention

[0005] This invention aims to solve the problems of low precision, poor efficiency, and insufficient safety when installing large equipment in narrow indoor spaces. It provides an adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion. Through laser guidance, visual recognition, and adaptive adjustment algorithms, it automatically optimizes the equipment position, path, and speed to ensure accurate docking and generate an installation completion report.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] An adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion, characterized by including:

[0008] S100. Real-time spatial pose perception steps: A real-time pose perception field for the device in a narrow space is constructed by fusing a visual recognition system and a laser guidance system. The visual recognition system obtains the spatial position deviation information of the device through image feature matching, and the laser guidance system performs precise ranging by emitting and receiving reflected laser signals. After the data of the two are fused, high-precision spatial position coordinates and attitude angle data of the device are output.

[0009] S200. First-level position servo adjustment steps: Based on the spatial position deviation information output by S100, an adaptive adjustment algorithm is used to generate the first control command to automatically adjust the position, angle and movement path of the device, driving the device to initially approach the target position;

[0010] S300. Installation Interface Status Monitoring Steps: During the equipment movement and initial positioning process, the contact pressure between the equipment and the installation base interface, as well as the temperature changes of key components of the equipment, are monitored in real time by pressure sensors and temperature sensors installed on the equipment to obtain pressure data and temperature data; when the pressure data exceeds the preset safety threshold or the temperature data is abnormal, an alarm is generated and a second control command is triggered to adjust the movement mode of the equipment or stop the installation.

[0011] S400. Third-level multi-objective optimization and fine-tuning step: Construct a second adaptive controller with spatial error, motion state, and interface state as common optimization objectives. The input of this controller is the real-time position and angle data continuously output from S100 and the pressure and temperature data continuously output from step S300, which are input into the adaptive control algorithm. The adaptive control algorithm dynamically generates a third control command based on the fused multi-source feedback data, and performs closed-loop optimization and adjustment of the equipment's speed, movement path, and assembly sequence to achieve precise and stable movement of the equipment to the target position in a narrow space while ensuring installation safety.

[0012] S500. Installation Quality Closed-Loop Verification Steps: After the equipment is moved to the theoretical target position according to the S400 steps, the final real-time position data, pressure data, and temperature data in S100 and S300 are correlated and analyzed with the external environment data to automatically detect and verify the installation results. When the verification results meet the preset tolerance standards, the installation is deemed successful and an installation completion report is automatically generated.

[0013] Furthermore, in S200 and S400, the adaptive adjustment algorithm and the adaptive control algorithm employ an optimization objective function and achieve adjustment through the following steps:

[0014] S41. Based on the S100 data, calculate the spatial error E between the current position of the device and the target position, using the following formula:

[0015]

[0016] Among them, (x) d y d , z d (x) represents the three-dimensional coordinates of the device's current position. t y t , z t () represents the three-dimensional coordinates of the target location;

[0017] S42. Minimize the optimization objective min[αE] 2 +βv 2 +γθ 2 Generate control commands.

[0018] Where α is the weighting coefficient of spatial error E, β is the weighting coefficient of equipment speed v, and γ is the weighting coefficient of the relative angle θ between the equipment and the target position, and α,β,γ>0.

[0019] S43. Based on the real-time data of S100 and S300, dynamically update the weight coefficients α, β, γ.

[0020] Furthermore, the dynamic update strategy for the weight coefficients α, β, γ is as follows:

[0021] The weighting coefficient α is positively correlated with the magnitude of the spatial error E. When E increases, the value of α increases accordingly to strengthen the correction.

[0022] The weighting coefficient β is positively correlated with the magnitude of the equipment speed v. When v increases, the value of β increases accordingly to suppress overshoot.

[0023] The weighting coefficient γ is positively correlated with the magnitude of the angle deviation θ. When θ increases, the value of γ increases accordingly to prioritize attitude correction.

[0024] Furthermore, in S500, the correlation analysis employs a Kalman filter algorithm to perform data fusion and filtering on the real-time location data, pressure data, temperature data, and external environment data, thereby improving the accuracy and reliability of the location verification results.

[0025] Furthermore, the fusion of the visual recognition system and the laser guidance system is achieved by registering the image coordinate system of the visual system and the ranging coordinate system of the laser system into the same world coordinate system, thereby constituting the pose perception.

[0026] Compared with the prior art, the present invention has the following advantages:

[0027] This method for installing large equipment in confined indoor spaces utilizes an integrated laser guidance system and visual recognition system. It acquires real-time spatial position deviation information of the equipment and accurately calculates its current position and angle based on feedback data, achieving precise docking. Compared to traditional technologies, this invention avoids errors from manual adjustments and enables high-precision positioning in complex, confined spaces. An adaptive adjustment algorithm dynamically optimizes the equipment's path, speed, and angle, automatically adjusting its motion strategy to cope with different environmental changes and improve installation accuracy. The dynamic adjustment mechanism of weighting coefficients, based on changes in error, equipment speed, and angle, ensures precise docking while avoiding over-adjustment, thus improving installation efficiency.

[0028] By using a real-time correlation analysis Kalman filter algorithm, which comprehensively considers real-time location data, pressure data, temperature data, and external environmental data, the equipment can continuously optimize its path and speed under changing environmental conditions. This ensures stable and safe installation and automatically generates an installation completion report, improving the reliability and efficiency of equipment installation. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion, according to the present invention.

[0030] Figure 2 This is a flowchart of Embodiment 1 of the adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion of the present invention. Detailed Implementation

[0031] 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.

[0032] Please see Figure 1 , Figure 1 This is a flowchart illustrating the adaptive installation method for large equipment in confined indoor spaces based on multi-source information fusion, as shown in the figure. It includes:

[0033] S100. Real-time spatial pose perception steps: A real-time pose perception field for the device in a narrow space is constructed by fusing a visual recognition system and a laser guidance system. The visual recognition system obtains the spatial position deviation information of the device through image feature matching, and the laser guidance system performs precise ranging by emitting and receiving reflected laser signals. After the data of the two are fused, high-precision spatial position coordinates and attitude angle data of the device are output.

[0034] S200. First-level position servo adjustment steps: Based on the spatial position deviation information output by S100, an adaptive adjustment algorithm is used to generate the first control command to automatically adjust the position, angle and movement path of the device, driving the device to initially approach the target position;

[0035] S300. Installation Interface Status Monitoring Steps: During the equipment movement and initial positioning process, the contact pressure between the equipment and the installation base interface, as well as the temperature changes of key components of the equipment, are monitored in real time by pressure sensors and temperature sensors installed on the equipment to obtain pressure data and temperature data; when the pressure data exceeds the preset safety threshold or the temperature data is abnormal, an alarm is generated and a second control command is triggered to adjust the movement mode of the equipment or stop the installation.

[0036] S400. Third-level multi-objective optimization and fine-tuning step: Construct a second adaptive controller with spatial error, motion state, and interface state as common optimization objectives. The input of this controller is the real-time position and angle data continuously output from S100 and the pressure and temperature data continuously output from step S300, which are input into the adaptive control algorithm. The adaptive control algorithm dynamically generates a third control command based on the fused multi-source feedback data, and performs closed-loop optimization and adjustment of the equipment's speed, movement path, and assembly sequence to achieve precise and stable movement of the equipment to the target position in a narrow space while ensuring installation safety.

[0037] S500. Installation Quality Closed-Loop Verification Steps: After the equipment is moved to the theoretical target position according to the S400 steps, the final real-time position data, pressure data, and temperature data in S100 and S300 are correlated and analyzed with the external environment data to automatically detect and verify the installation results. When the verification results meet the preset tolerance standards, the installation is deemed successful and an installation completion report is automatically generated.

[0038] Example 1: Adaptive Installation of Large Compressors on Concrete Foundations

[0039] like Figure 2 As shown, a large compressor is installed on a concrete foundation. The specific implementation process is as follows:

[0040] 1. System Setup and Initialization

[0041] First, the integrated boot module is fixedly installed around the installation area. Specifically:

[0042] Visual recognition system: Employs at least two high-resolution industrial cameras to cover the entire movement path of the equipment and the target foundation area from different angles. During system initialization, the cameras are calibrated, and the target image features (such as the relative positions of the base bolt holes and the foundation embedded parts) of the equipment when it is finally in place are pre-input.

[0043] Laser guidance system: Multiple laser trackers or lidar units are deployed within the space, with their scanning range covering the entire path from the entrance to the target location. This system establishes a precise mapping with the world coordinate system, providing an absolute coordinate reference for the entire space.

[0044] Sensor installation: Integrate pressure sensor arrays (e.g., multiple load cells evenly distributed under the equipment base) onto the equipment support points or actuators (such as lifting legs). Attach temperature sensors (such as PT100) to critical components of the equipment that are prone to heat generation, such as motor bearings.

[0045] The control center is connected to the actuators (such as multiple high-precision servo electric cylinders) to complete system self-testing and zero-point calibration.

[0046] 2. Specific steps for installation

[0047] S100. Real-time Spatial Pose Sensing

[0048] Once the equipment begins to move, the visual recognition system acquires real-time image data of the equipment through image sensors and compares the equipment's position with the target position to generate spatial position deviation information. Simultaneously, the laser guidance system emits and receives reflected signals via a laser beam, measuring the spatial distance between the equipment and the target position in real time. The control center integrates the visual attitude data and the laser's absolute position data to output real-time six-degree-of-freedom pose data of the equipment.

[0049] S200. First-level position servo adjustment

[0050] After receiving the pose data from S1, the control center first calculates the spatial Euclidean error E. At this stage, the goal of the first adaptive controller is to reduce E with high efficiency. Based on the calculation results, the controller generates first-level control commands, directing the four servo electric cylinders to work in coordination, pushing the entire device towards the target position at a relatively fast speed. This stage is similar to "coarse positioning," aiming to quickly eliminate most of the positional deviation.

[0051] S300 Installation Interface Status Monitoring

[0052] During equipment movement and S2 adjustments, the pressure sensor array monitors the force on each support leg in real time. Ideally, the pressure on all four support legs should be balanced. Temperature sensors continuously monitor the motor temperature. All this data is independently transmitted to the safety monitoring thread in the control center. A combined pressure and temperature sensor system is set up; the pressure sensors monitor the contact pressure between the equipment and the mounting surface, while the temperature sensors monitor temperature changes. By obtaining pressure and temperature data, when the contact pressure between the equipment and the mounting surface is too high or the temperature is abnormal, an alarm is automatically issued and the adjustment device is triggered to adjust the equipment's movement or stop installation, preventing damage to the equipment due to uneven pressure or abnormal temperature.

[0053] S400. Second-level multi-objective optimization and fine-tuning

[0054] When the equipment moves to a position within the error range of less than 10mm via step S2, the system automatically switches to fine-tuning mode. At this point, the second adaptive controller begins operation. Based on the pressure and temperature data detected in step S3, it dynamically adjusts the equipment installation process using an adaptive control algorithm. The algorithm adjusts the equipment's speed, movement path, and assembly sequence based on feedback data to ensure the equipment maintains optimal operating condition throughout the installation process, preventing damage and ensuring efficient and precise installation even in confined spaces. The adaptive adjustment algorithm dynamically optimizes the adjustment strategy based on the equipment's real-time motion state and changes in the spatial environment, improving installation accuracy and efficiency. The adaptive adjustment algorithm adjusts the equipment's position, speed, and angle using the following mathematical model:

[0055] S41. Calculate the spatial error between the current position of the device and the target position:

[0056]

[0057] Where E is the spatial error. The three-dimensional coordinates of the current position of the device. The three-dimensional coordinates of the target location;

[0058] S42. Based on the spatial error and the current speed and angle of the equipment, adjust the movement path, speed and angle of the equipment using an adaptive control algorithm;

[0059] S43. The adaptive adjustment algorithm dynamically updates the weight coefficients α, β, and γ to optimize the equipment's path, speed, and installation accuracy. The adaptive control algorithm minimizes the installation error through the following optimization objective function:

[0060] min[αE 2 +βv 2 +γθ 2 ]

[0061] Where: α represents the weight of the error E on the optimization target, β represents the weight of the velocity v on the optimization target, and γ represents the weight of the angle θ on the optimization target. v is the velocity vector of the device, and θ is the relative angle between the device and the target position. The weight coefficients α, β, and γ are all greater than 0 and are dynamically adjusted according to the real-time environment and operational requirements of the device. The weight coefficient α is related to the error E. If the error E is large (i.e., the deviation between the device and the target position is large), the weight coefficient of α will increase, enhancing the correction effect on the error and ensuring that the device can approach the target position more quickly. When the device approaches the target, the weight coefficient of α can decrease to avoid over-adjustment. The weight coefficient β is related to the device velocity v. When the device speed is too high, it may lead to difficulties in precise positioning. In this case, β will increase to limit the excessive influence of velocity v, thereby reducing unnecessary speed adjustments. When the speed is too slow, β will decrease, allowing for higher device movement speeds to improve efficiency. The weight coefficient γ is related to the angle θ between the device and the target position. If the angular deviation of the device is large, the weight coefficient of γ will increase to ensure that the device can quickly adjust the angle and avoid the device failing to dock precisely due to angular errors. When the angle error is small, the weighting coefficient of γ can be reduced, reducing the need for angle correction and making the adjustment process smoother.

[0062] S500. Installation Quality Closed-Loop Verification:

[0063] After the equipment is in place, the system remains still for 10 seconds to allow the data to stabilize.

[0064] The real-time location, pressure, and temperature data in S1 and S3 are correlated with external environmental data. Based on the analysis results, the installation path and speed of the equipment are optimized. After the equipment is installed, automatic detection and position verification are performed to ensure accurate alignment of the equipment and generate an installation completion report. The correlation analysis uses the Kalman filter algorithm.

[0065] Example 2: Dynamic Algorithm Adjustment Example

[0066] In this embodiment, each step of the equipment installation process is dynamically optimized and adjusted based on real-time sensor data and environmental changes. The adaptive control algorithm analyzes feedback data to automatically adjust the equipment's speed, path, and assembly sequence, ensuring that the equipment maintains optimal operating condition throughout the installation process, preventing equipment damage, and ensuring that the equipment can be installed efficiently and accurately in confined spaces.

[0067] Calculate the spatial error between the current position of the calculation device and the target position;

[0068] The three-dimensional coordinates of the device are obtained through a laser guidance system and a visual recognition system. ) and the three-dimensional coordinates of the target location ( The error between the equipment position and the target position is calculated using the Euclidean distance formula:

[0069]

[0070] For example, if the current location of the device is The target location is , The calculation error is:

[0071]

[0072] Adjust the equipment's path, speed, and angle based on spatial errors and the equipment's current speed and angle.

[0073] Based on the calculated error E, the current speed of the device v = 0.1 m / s (assuming the current speed of the device is 0.1 m / s), and the relative angle θ between the device and the target position (assuming the angle deviation is 5°), ∘ The adaptive adjustment algorithm uses the following optimization objective function for adjustment:

[0074] min[αE 2 +βv 2 +γθ 2 ]

[0075] Assuming weight coefficients α=10, β=5, and γ=2, the calculated objective function value is:

[0076]

[0077] The adaptive control algorithm minimizes this objective function by adjusting the path, speed, and angle. Because the error E is large, the weight of α is large, while the effects of speed V and angle θ are relatively small.

[0078] Through this dynamic self-tuning of weight coefficients based on real-time feedback, the adaptive control algorithm of this invention can flexibly cope with various complex situations that arise during installation, achieving balance and optimization among multiple objectives.

[0079] 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 variations 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 method for adaptive installation of large equipment in narrow indoor spaces based on multi-source information fusion, characterized in that, include: S100. Real-time spatial pose perception steps: A real-time pose perception field for the device in a narrow space is constructed by fusing a visual recognition system and a laser guidance system. The visual recognition system obtains the spatial position deviation information of the device through image feature matching, and the laser guidance system performs precise ranging by emitting and receiving reflected laser signals. After the data of the two are fused, high-precision spatial position coordinates and attitude angle data of the device are output. S200. First-level position servo adjustment steps: Based on the spatial position deviation information output by S100, an adaptive adjustment algorithm is used to generate the first control command to automatically adjust the position, angle and movement path of the device, driving the device to initially approach the target position; S300. Installation Interface Status Monitoring Steps: During the equipment movement and initial positioning process, the contact pressure between the equipment and the installation base interface, as well as the temperature changes of key components of the equipment, are monitored in real time by pressure sensors and temperature sensors installed on the equipment to obtain pressure data and temperature data; when the pressure data exceeds the preset safety threshold or the temperature data is abnormal, an alarm is generated and a second control command is triggered to adjust the movement mode of the equipment or stop the installation. S400. Third-level multi-objective optimization and fine-tuning step: Construct a second adaptive controller with spatial error, motion state, and interface state as common optimization objectives. The input of this controller is the real-time position and angle data continuously output from S100 and the pressure and temperature data continuously output from step S300, which are input into the adaptive control algorithm. The adaptive control algorithm dynamically generates a third control command based on the fused multi-source feedback data, and performs closed-loop optimization and adjustment of the equipment's speed, movement path, and assembly sequence to achieve precise and stable movement of the equipment to the target position in a narrow space while ensuring installation safety. S500. Installation Quality Closed-Loop Verification Steps: After the equipment is moved to the theoretical target position according to the S400 steps, the final real-time position data, pressure data, and temperature data in S100 and S300 are correlated and analyzed with the external environment data to automatically detect and verify the installation results. When the verification results meet the preset tolerance standards, the installation is deemed successful and an installation completion report is automatically generated.

2. The adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion according to claim 1, characterized in that: In S200 and S400, the adaptive adjustment algorithm and the adaptive control algorithm adopt an optimization objective function and achieve adjustment through the following steps: S41. Based on the S100 data, calculate the spatial error E between the current position of the device and the target position, using the following formula: Among them, (x) d y d , z d (x) represents the three-dimensional coordinates of the device's current position. t y t , z t () represents the three-dimensional coordinates of the target location; S42. Minimize the optimization objective min[αE] 2 +βv 2 +γθ 2 Generate control commands. Where α is the weighting coefficient of spatial error E, β is the weighting coefficient of equipment speed v, and γ is the weighting coefficient of the relative angle θ between the equipment and the target position, and α,β,γ>0. S43. Based on the real-time data of S100 and S300, dynamically update the weight coefficients α, β, γ.

3. The adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion according to claim 2, characterized in that: The dynamic update strategy for the weight coefficients α, β, γ is as follows: The weighting coefficient α is positively correlated with the magnitude of the spatial error E. When E increases, the value of α increases accordingly to strengthen the correction. The weighting coefficient β is positively correlated with the magnitude of the equipment speed v. When v increases, the value of β increases accordingly to suppress overshoot. The weighting coefficient γ is positively correlated with the magnitude of the angle deviation θ. When θ increases, the value of γ increases accordingly to prioritize attitude correction.

4. The adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion according to claim 1, characterized in that: In S500, the correlation analysis uses the Kalman filter algorithm to perform data fusion and filtering on the real-time location data, pressure data, temperature data and external environment data, so as to improve the accuracy and reliability of the location verification results.

5. The adaptive installation method for large equipment in narrow indoor spaces based on multi-source information fusion according to claim 1, characterized in that: The fusion of the visual recognition system and the laser guidance system is achieved by registering the image coordinate system of the visual system and the ranging coordinate system of the laser system into the same world coordinate system, thereby constituting the pose perception.

Images