A control system and operation method of a double-steel-wire detection robot for a bridge-bottom box girder

Abstract

The application discloses a kind of bridge bottom box girder double steel wire detection robot control system and operating method, for the problems of low efficiency, insufficient precision, weak environmental adaptability of existing bridge detection, the system includes host computer, lower computer and motion control system, closed loop network is constructed by TCP / UDP, serial port and other multi-protocol redundant communication, supports automatic and manual dual operation mode.Motion control system adopts PID closed-loop control and encoder-odometer fusion positioning, realizes accurate movement along double steel wire, multi-pose shooting and obstacle avoidance;Core module has double backup design, with power switching and communication link redundancy mechanism, to ensure stable operation.The application realizes automatic detection of box girder interior without dead angle, improves detection efficiency and data reliability, adapts to complex working environment, reduces manual risk, and has important engineering application value.

Description

Technical Field

[0001] This invention relates to the field of intelligent inspection equipment and robot control technology, specifically to a bridge box-type cavity inspection robot based on a flexible cable track and its operation method. Background Technology

[0002] As a core component of transportation infrastructure, the structural safety of bridges directly affects the safety of people's lives and property and the continuity of transportation. The box girder at the bottom of the bridge, as a key load-bearing component, is exposed to a complex outdoor environment for extended periods. Its internal structure is prone to defects such as coating runs, bubbles, component corrosion, paint peeling, weld cracks, and loose connections due to environmental erosion and stress. If these defects are not detected and addressed in a timely manner, they will gradually affect the bridge's load-bearing capacity and may even lead to structural failure. Therefore, regular, comprehensive, and accurate inspection of the box girder's interior is crucial.

[0003] Currently, there are two main types of bridge inspection methods: one is traditional manual inspection, where inspectors need to enter the box girder or work with the aid of elevated equipment. This is not only labor-intensive and inefficient, but the results are also highly dependent on the inspectors' experience, making it subjective and difficult to guarantee the comprehensiveness and accuracy of the inspection. In addition, the confined environment inside the box girder also poses safety hazards. The other type is intelligent inspection, including acoustic detection, vibration signal analysis, image processing, infrared thermal imaging, and drone inspection. Although these methods reduce reliance on manual methods to some extent, they still have significant limitations: acoustic detection and vibration signal analysis are sensitive to environmental noise and have high signal processing complexity; image processing is easily affected by insufficient light and dust interference inside the box girder, limiting the accuracy of the inspection; drone inspection is difficult to adapt to the confined space inside the box girder and the special working environment of double steel wire support, and it cannot achieve precise mobile inspection along a fixed path. Summary of the Invention

[0004] Based on the aforementioned issues, and considering the specific scenario of double-wire inspection of bridge box girders, there is an urgent need for a robot control system capable of precise motion control, stable data transmission, and multi-module collaborative operation. This system would address the shortcomings of existing technologies in terms of adaptability, accuracy, and reliability, thereby improving the efficiency and quality of bridge inspection.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] A control system for a double-wire inspection robot for bridge box girders is provided. The system includes an upper computer, a lower computer, and a motion control system. The three components form a closed-loop control network through multi-protocol redundant communication. It is adapted to the double-wire operation scenario of bridge box girders and realizes full-process control from command input to robot operation execution and data feedback. The host computer serves as the core of user interaction, integrating a control input module, a data and video display module, a data storage module, a communication module, and a log module. The host computer receives user operation commands and sends them to the slave computer. At the same time, it receives the dual-wire status video stream, sensor data, and equipment status information uploaded by the slave computer, and completes data display, storage, and anomaly recording. The lower-level machine, acting as an intermediate switching layer, includes a motion control module, a data storage module, a log module, a communication module, and a parsing module. The lower-level machine parses the upper-level machine's instructions and converts them into messages that the motion control system can recognize. At the same time, it collects data from the dual-wire positioning sensor of the motion control system, the robotic arm's posture data, and video frames, and reports them to the upper-level machine after processing. The motion control system uses a microcontroller as the main controller and includes a motion control algorithm module, a data acquisition module, a data storage module, a communication module, and a log module. The motion control system receives instructions from the lower-level machine and drives the robot hardware to perform motion along the double steel wire. It also collects data from peripherals such as encoders, lidar, batteries, and double steel wire tension sensors and feeds it back to the lower-level machine. The host computer and the slave computer establish a main communication link through TCP+UDP protocol and wireless bridge, with wired network cable as backup link. The slave computer communicates with the motion control system through serial port protocol, and the motion control system communicates with peripherals through RS485 / RS232 protocol, forming a multi-layered, highly reliable communication network.

[0007] In the preferred embodiment, the host computer's control input module adopts a dual backup design of physical handle and virtual button, ensuring consistent control logic; the data and video display module includes a robot control submodule, an image viewing and stitching submodule, a video monitoring submodule, and a dual-wire status early warning submodule, which displays the robotic arm's posture, captured images, battery voltage, communication status, and dual-wire detection anomaly information in real time; The physical handle is the primary input device, which listens for forward, backward, turn, and start / stop button signals of the robotic arm in real time and converts them into control logic. The virtual buttons serve as a backup solution when the physical peripheral fails. They are triggered through the host computer's UI interface to ensure control continuity. The control signals are cached by the data storage module and then sent to the lower-level computer through the communication module.

[0008] In the preferred embodiment, the motion control module of the lower-level machine supports seamless switching between manual and automatic modes, and the switching logic is triggered by protocol control codes; In manual mode, discrete commands issued by the host computer are converted into corresponding messages to precisely control the robot's step-by-step movement along the double steel wire, adjust the robotic arm's posture, and trigger photo capture. In automatic mode, after receiving the one-click start command from the host computer, the robot calls the preset double steel wire detection process message and controls the robot to autonomously complete a series of operations such as double steel wire positioning, moving along the steel wire, taking pictures in multiple postures, and uploading data. The lower-level machine's log module monitors the operating status of each functional module in real time, records information such as abnormal instruction transmission, data parsing failure, abnormal tension of the double steel wires, and equipment failure, forming a traceable operation log. The log includes a frame header, address code, protocol control code, exception type, and timestamp.

[0009] In the preferred embodiment, the motion control system adopts a four-layer architecture consisting of a perception layer, a decision layer, an execution layer, and a transmission layer, which is suitable for dual-wire working environments. The perception layer accesses peripherals via multiple protocols, including an extended arm passive encoder, LiDAR, dual wire tension sensor, battery voltage sensor, and robotic arm posture sensor, to collect raw data such as motor speed, robot position along the wire, environmental obstacles, dual wire tension, and battery voltage. The decision-making level integrates multi-sensor data through the motion control algorithm module to plan the optimal working path and robotic arm posture sequence along the double steel wire, thereby avoiding interference with the internal structure of the box girder; The execution layer drives the motor and robotic arm to perform movements according to the decision instructions, and controls the high-precision camera to complete the photography of the double steel wire and box girder components; The transport layer enables bidirectional data transmission with the lower-level machine, ensuring real-time command issuance and data feedback with a transmission delay of ≤100ms.

[0010] In the preferred embodiment, the motion control algorithm module employs a proportional-derivative-integral closed-loop control algorithm, integrating encoder positioning data, dual-wire tension data, and lidar obstacle avoidance data to achieve precise motion control along the dual-wire. The control signal calculation formula is as follows: ; in, For motor control output signals, This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients. The real-time deviation between the target position and the actual position is calculated by fusing encoder positioning deviation and dual-wire tension compensation deviation. When the lidar detects the distance to the obstacle hour, To preset a safe distance threshold, with a value of 5-10cm, the algorithm automatically triggers obstacle avoidance logic and adjusts the control signal to make the robot shift laterally along the double steel wire or decelerate and stop.

[0011] In the preferred embodiment, the communication network architecture is built upon a host computer communication module, a slave computer communication module, and a motion control system communication module, employing multi-link redundancy and frame structure verification design. The main communication link between the host computer and the slave computer communication module, built through a bridge and TCP / UDP protocol, is the core, responsible for high-bandwidth, real-time data transmission of video streams, control commands, etc. The wired network cable between the host computer and the slave computer serves as a backup link, which automatically switches when the main link fails three consecutive checks to ensure that core data is not lost. In the serial port protocol communication between the lower-level communication module and the motion control system communication module, the communication frame includes a frame header, address code, protocol control code, life signal, service protocol, check code and frame tail. The frame header is fixed at 0xFE and the frame tail is fixed at 0xA5. The life signal is a random number generated by the upper-level computer and is used to verify the validity of the communication link. The check code is calculated by the CRC8 algorithm. The architecture integrates a distributed caching function, which temporarily caches the transmitted sensor data and video frames through the data storage modules of the host computer and the slave computer. The caching time is ≥30s, reducing the impact of the complex environment under the bridge on data transmission.

[0012] In the preferred embodiment, the data storage module of the host computer adopts a dual structure design of driver layer-physical layer, with one-to-one correspondence between the driver layer and the physical layer, and a dual backup strategy of local + remote for critical data. The driver layer receives the raw data transmitted by the communication module, parses it, performs CRC8 verification, and encapsulates it into a standardized data format; The physical layer classifies and stores control command data, sensor timing data, video frame data, and log data. Video frame data is named according to timestamp-wire number-work location. Sensor data is stored in blocks according to 10Hz frequency. Video frame data is backed up to local storage devices and remote servers in real time, with a backup delay of ≤5s.

[0013] In the preferred embodiment, the robot localization employs a fusion algorithm combining an extended-arm passive encoder and an odometry system, adapting to the precise localization requirements of dual-wire detection. The formula for calculating the movement distance is as follows: ; in, This represents the actual distance (in meters) the robot travels along the double steel wires. The passive encoder outputs the rotation angle (rad). Let be the radius (m) of the motor drive wheel. Encoder resolution (lines / revolution). This refers to the motor reduction ratio; When the robot has moved 90% of the target position, a linear deceleration program is initiated, and the speed change during deceleration satisfies the following: ( ); in For the target speed, The preset deceleration time is set to 0.5-1s to ensure that the robot stops smoothly along the double steel wires and avoids the vibration of the steel wires affecting the detection accuracy.

[0014] In the preferred embodiment, the robot control system is equipped with a dual power supply module consisting of a main power supply and a backup power supply. The main power supply uses a 48V rechargeable lithium battery, and the backup power supply uses a 12V emergency battery, which is suitable for scenarios where there is no external power supply under the bridge. The motion control system monitors the main power supply level in real time. and voltage When detected or When the power switching logic is triggered automatically, the backup power supply is seamlessly connected. At the same time, the host computer issues an audible and visual alarm, and the robot performs an emergency reset operation: the robotic arm returns to the initial position, and the robot slowly moves along the double steel wire to the safe stopping point.

[0015] In the preferred embodiment, the system operates in both automatic and manual modes, and both modes include communication handshake and exception handling procedures. The automatic operation mode includes the following steps: S1. System initialization, odometer zeroing, main and backup power self-test, detection of peripherals such as dual steel wire tension sensor and encoder, communication handshake between host computer and slave computer, and between slave computer and motion control system in sequence, and verification of life signal consistency. S2, Determine the main power supply and If the peripheral device status is normal, or if it does not meet the requirements, the backup power supply will be switched on and an alarm will be issued to terminate the operation; if it meets the requirements, proceed to the next step. S3, The robot moves at the target speed. Moving along two steel wires, the distance traveled is collected in real time by an odometer and a passive encoder. ; S4, when Reach the preset work position At that time, the robot performs a deceleration and stop operation; S5. The robotic arm adjusts sequentially according to the preset posture sequence in the configuration file. After each posture positioning is completed, the camera is triggered to take pictures of the double steel wire and box girder components, and the shooting data is uploaded to the lower computer in real time. S6. After taking photos in all preset poses, the robotic arm resets to its initial position; S7. Repeat steps S3-S6 until the preset working mileage is completed or a power abnormality or communication interruption is detected. The robot then returns to the safe docking point and stops working. The manual operation mode includes the following steps: T1. Users input operation commands via a physical handle or virtual buttons on the host computer. Commands include forward, backward, turn, robotic arm adjustment, and taking a picture. T2. The host computer encapsulates the instructions into standardized data frames, including CRC8 checksums and life signals, and sends them to the slave computer via TCP+UDP protocol. T3: The lower-level machine parses the data frame, and after verifying that the check code and life signal are correct, it converts them into a message that the motion control system can recognize, and sends it down through the serial port protocol. T4. After receiving the message, the motion control system calculates and generates control signals through the motion control algorithm module, which then drive the motor and robotic arm to perform the corresponding actions. T5: The motion control system collects the operating data of the peripheral devices and the status of the double steel wires in real time, and feeds it back to the lower computer via the communication module, and then the lower computer reports it to the upper computer. T6: The host computer displays the equipment's operating status, video feed, and sensor data in real time, and synchronously stores operation data and operation logs. When three consecutive communication verification anomalies are detected, it automatically triggers communication link switching or emergency shutdown.

[0016] A control system and operation method for a double-wire inspection robot for bridge bottom box girders, the beneficial effects of which include, but are not limited to, the following: 1. Supports seamless switching between manual and automatic modes. In automatic mode, the robot can autonomously move along the double steel wires, position itself, take photos in multiple poses, and upload data without human intervention inside the box girder. Compared with traditional manual inspection, efficiency is greatly improved, while avoiding the subjectivity and limitations of manual operation, achieving full-area inspection without blind spots. 2. A proportional-derivative-integral closed-loop control algorithm is adopted, which integrates encoder and odometer positioning data, combined with dual steel wire tension compensation and lidar obstacle avoidance logic, so that the robot's movement and positioning error is controlled at the millimeter level; a linear deceleration program ensures a smooth stop before taking pictures, and the collected data is standardized and packaged with local + remote dual backup to provide reliable data support for defect analysis. 3. The communication architecture adopts a multi-link redundancy and frame structure verification design, with automatic switching between primary and backup links to resist signal interference in the complex environment under the bridge; control input, power supply, and data storage all adopt a dual backup design, coupled with an emergency reset function, to ensure safe equipment docking and no data loss in the event of a sudden failure, thus ensuring continuous operation. 4. Specifically designed for double steel wire scenarios of box girder bridge bottom, the motion control system adopts a four-layer architecture, and adapts to complex environments such as narrow spaces and insufficient light through multi-sensor fusion; the lidar obstacle avoidance function can automatically avoid interference from the internal structure of the box girder, and the distributed cache design ensures continuous data transmission, so it can be adapted to the detection of bridges of different specifications without large-scale modifications. 5. The modular and layered architecture makes each functional module independent, which facilitates development, testing and later maintenance; the system reserves sensor expansion interfaces, which can add detection modules as needed to adapt to more complex detection scenarios and enhance the reusability of equipment. 6. The automated process completely eliminates the safety hazards of manual entry into the box girder. The host computer monitors the robot's operating status and detection data in real time. In case of power failure, it automatically switches to backup power and issues an alarm. The robot performs operations such as mechanical arm reset and safe docking, comprehensively ensuring the safety of personnel, equipment and bridge structure. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the control system for the double steel wire inspection robot of the bridge bottom box girder of the present invention; Figure 2 This is a flowchart illustrating the operation method of the double-wire inspection robot control system for the bridge bottom box girder of the present invention. Detailed Implementation

[0018] like Figure 1 As shown, a control system for a double-wire inspection robot for a bridge box girder is provided. The system includes an upper computer, a lower computer, and a motion control system. The three components form a closed-loop control network through multi-protocol redundant communication, which is adapted to the double-wire operation scenario of the bridge box girder and realizes full-process control from command input to robot operation execution and data feedback. The host computer uses an industrial-grade tablet, Advantech TPC-1551H, as the core of user interaction. It integrates a control input module, a data and video display module, a data storage module, a communication module, and a log module. The host computer receives user operation commands and sends them to the slave computer via the 192.168.1.10-20 network segment. At the same time, it receives 1920×1080 resolution video streams of the dual steel wire status, sensor data, and equipment status information uploaded by the slave computer. Data display, storage, and anomaly recording are completed through a local 512GB SSD. The log storage format is JSON. The lower-level machine uses the Advantech IPC-6806 industrial control computer as an intermediate transfer layer, which includes a motion control module, a data storage module, a log module, a communication module, and a parsing module. The lower-level machine parses the upper-level machine's instructions through industrial Ethernet and converts them into messages that the motion control system can recognize at a baud rate of 9600bps. At the same time, it collects data from the dual wire positioning sensors of the motion control system, the robot arm's posture data, and 30fps video frames. After CRC8 verification, the data is reported to the upper-level machine. The motion control system uses an STM32H743VI microcontroller as the main controller and includes a motion control algorithm module, a data acquisition module, a data storage module, a communication module, and a log module. The motion control system receives instructions from the lower-level machine and drives the robot hardware to perform motion along the double steel wire. It also collects data from peripherals such as the E6B2-CWZ6C encoder, YDLIDAR X4 lidar, 48V lithium battery, and NS-WL100 double steel wire tension sensor and feeds it back to the lower-level machine. The host computer and the slave computer establish a main communication link through TCP port 8080, UDP port 8081, and the industrial wireless bridge Huawei AirEngine5760-22W. The Cat5e wired network cable serves as a backup link. The slave computer communicates with the motion control system via the Modbus serial protocol. The motion control system communicates with peripherals via RS485 and RS232 protocols. The RS485 protocol has a baud rate of 115200bps, and the RS232 protocol has a baud rate of 9600bps, forming a multi-layered, highly reliable communication network.

[0019] In the preferred embodiment, the host computer's control input module adopts a dual-backup design of physical controller and virtual buttons. The physical controller uses an Xbox Wireless Controller, which communicates via Bluetooth 5.0, and the control logic is consistent with that of the virtual buttons. The data and video display module includes a robot control submodule, an image viewing and stitching submodule, a video monitoring submodule, and a dual-wire status warning submodule. The UI interface is divided into a left video area, a right data area, and a bottom warning area. The left video area has two monitoring video windows and one photo preview window. The right data area displays the robot arm posture, dual-wire tension value, battery voltage, and communication status. The robot arm posture display accuracy is ±0.1°, and the battery voltage display accuracy is 0.1V. The bottom warning area displays detected abnormalities through red pop-up windows and displays the captured image and dual-wire detection abnormality information in real time. The physical handle is the primary input device, which listens for forward, backward, turn, and start / stop button signals of the robotic arm in real time. The button response delay is no more than 50ms and is converted into control logic. The virtual buttons are triggered by the function area on the left side of the host computer's UI interface. The layout corresponds one-to-one with the buttons on the physical handle, serving as a backup plan in case of physical peripheral failure and ensuring control continuity. The control signals are buffered for 10 seconds by the data storage module and then sent to the lower computer through the communication module.

[0020] In the preferred embodiment, the motion control module of the lower-level machine supports seamless switching between manual and automatic modes. The switching logic is triggered by the protocol control code, with control code 0x01 corresponding to manual mode and control code 0x02 corresponding to automatic mode. In manual mode, discrete instructions issued by the host computer are converted into corresponding messages of 16 bytes in length, which precisely control the robot to move in 50cm steps along the double steel wire, adjust the tilt angle of the robotic arm within the range of -30° to 60° and trigger the photo taking, with a shutter delay of no more than 100ms. In automatic mode, after receiving the one-click start command 0x03 from the host computer, the robot calls the preset double steel wire detection process message and controls the robot to autonomously complete the continuous operation of double steel wire positioning, moving along the steel wire, taking pictures in multiple postures, and uploading data. The double steel wire positioning error does not exceed 2mm, the moving speed along the steel wire is 0.3m / s at a constant speed, and the pictures are taken at 3 angles at each working position. The lower-level machine's log module monitors the operating status of each functional module in real time, recording information such as instruction transmission anomalies, data parsing failures, double wire tension exceeding the 200 to 300N threshold, and equipment malfunctions, forming a traceable operating log. The log includes a frame header, address code, protocol control code, exception type, and timestamp. The frame header is fixed at 0xFE, the address code corresponds to 0x01 for the left wire and 0x02 for the right wire, the exception type is coded from 0x01 to 0x08, and the timestamp is accurate to milliseconds.

[0021] In the preferred embodiment, the motion control system adopts a four-layer architecture consisting of a perception layer, a decision layer, an execution layer, and a transmission layer, which is suitable for dual-wire working environments. The perception layer accesses peripherals via multiple protocols, including an extended arm passive encoder, LiDAR, dual-wire tension sensor, battery voltage sensor, and robotic arm posture sensor. The extended arm passive encoder has a resolution of 1024 lines / revolution. The LiDAR has a detection range of 0.1 to 10m and a ranging accuracy of ±2cm. The dual-wire tension sensor has a measurement range of 0 to 500N and an accuracy of ±1N. The battery voltage sensor has a measurement range of 40 to 50V. The robotic arm posture sensor uses an MPU6050 to collect raw data such as motor speed, robot position along the wire, environmental obstacles, dual-wire tension, and battery voltage. The raw data sampling frequency is 10Hz. The decision-making level integrates multi-sensor data through the motion control algorithm module to plan the optimal working path and robotic arm posture sequence along the double steel wire. The deviation of the optimal working path does not exceed 5mm. Each working point is preset with 3 to 5 robotic arm posture sequences to avoid interference with the internal structure of the box girder and maintain a safe distance of not less than 10cm from the internal structure of the box girder. The execution layer drives the 57BLDC DC geared motor and the 6-DOF robotic arm to perform movements according to the decision instructions, and controls the 1920×1080 pixel high-precision camera Basler acA1920-40gm to complete the photography of the double steel wire and box girder components. The transport layer enables bidirectional data transmission with the lower-level machine, ensuring real-time command issuance and data feedback with a transmission delay of ≤100ms.

[0022] In the preferred embodiment, the motion control algorithm module employs a proportional-derivative-integral closed-loop control algorithm, integrating encoder positioning data, dual-wire tension data, and lidar obstacle avoidance data to achieve precise motion control along the dual-wire. The control signal calculation formula is as follows: ; Where is the motor control output signal, measured in V. This is a proportionality coefficient, with a value ranging from 0.2 to 0.5. This is the integral coefficient, with a value ranging from 0.02 to 0.1. is the differential coefficient, with a value ranging from 0.05 to 0.2. This represents the real-time deviation between the target position and the actual position, expressed in meters (m). It incorporates the encoder positioning deviation and the dual-wire tension compensation deviation, with a dual-wire tension compensation deviation coefficient of 0.02 N. -1 m; When the lidar detects the distance to the obstacle hour, The preset safe distance threshold is set to 8cm. The algorithm automatically triggers the obstacle avoidance logic, adjusts the control signal to make the robot shift laterally along the double steel wire, with the offset not exceeding 10cm, or decelerates and stops, with a deceleration of 0.1m / s².

[0023] In the preferred embodiment, the communication network architecture is built upon a host computer communication module, a slave computer communication module, and a motion control system communication module, employing multi-link redundancy and frame structure verification design. The system uses a main communication link between the host computer and the slave computer communication module, built via a bridge and TCP / UDP protocols. TCP is used to transmit control commands, and UDP is used to transmit video streams, handling high-bandwidth, real-time data transmission such as video streams and control commands. A Cat5e wired network cable between the host computer and the slave computer serves as a backup link. The main link automatically switches when it fails three consecutive verifications within 500ms, with a switching delay of no more than 300ms, ensuring that core data is not lost. In the serial communication protocol between the lower-level communication module and the motion control system communication module, the communication frame includes a frame header, address code, protocol control code, life signal, service protocol, checksum, and frame trailer. The frame header is fixed at 0xFE, and the frame trailer is fixed at 0xA5. Each of the frame header, address code, protocol control code, life signal, checksum, and frame trailer occupies 1 byte, while the service protocol occupies 10 bytes. The life signal is a random number generated by the upper-level computer in the range of 0x00~0xFF, used to verify the validity of the communication link. The checksum is calculated using the CRC8 algorithm, where the CRC8 algorithm polynomial is x... 8 +x 5 +x 4 +1; The architecture integrates a distributed caching function, which temporarily caches the transmitted sensor data and video frames through the local SSD of the host computer and the 1TB SATA hard drive of the slave computer. The caching time is ≥30s. The cache directory is created hierarchically by date and job number to reduce the impact of the complex environment under the bridge on data transmission.

[0024] In the preferred embodiment, the data storage module of the host computer adopts a dual structure design of driver layer-physical layer, with one-to-one correspondence between the driver layer and the physical layer, and a dual backup strategy of local + remote for critical data. The driver layer receives the raw data transmitted by the communication module, parses it in the order of "frame header-address code-data segment-checksum-frame tail", and encapsulates it into a standardized data format after CRC8 verification. The standardized data format includes timestamp, device ID, parameter type, value, and verification result field. The physical layer stores control command data, sensor timing data, video frame data, and log data in a categorized manner. Sensor timing data includes dual wire tension, encoder positioning, battery voltage, etc. Video frame data is named according to "timestamp-wire number-operation position", with a naming format such as 20251130142030-W1-P005.jpg. Sensor data is stored in blocks at a frequency of 10Hz, with 100 data entries per block. Video frame data is backed up to a remote server in real time via FTP protocol. The remote server IP is 192.168.1.100, port 21, and the backup delay is ≤5s.

[0025] In the preferred embodiment, the robot localization employs a fusion algorithm combining an extended-arm passive encoder and an odometry system, adapting to the precise localization requirements of dual-wire detection. The formula for calculating the movement distance is as follows: ; in, This represents the actual distance the robot travels along the double steel wires, in meters (m). This is the output rotation angle of the passive encoder, in rad. The radius of the motor drive wheel is 0.05m. This represents the encoder resolution, with a value of 1024 lines per revolution. The motor reduction ratio is set to 10. When the robot has moved 90% of the target position, a linear deceleration program is initiated, and the speed change during deceleration satisfies the following: ( ); in The target velocity is set to 0.3 m / s. The preset deceleration time is set to 0.8s to ensure that the robot stops smoothly along the double steel wires and avoids the vibration amplitude of the steel wires exceeding 0.5mm, which would affect the detection accuracy.

[0026] In the preferred embodiment, the robot control system is equipped with a dual power supply module consisting of a main power supply and a backup power supply. The main power supply is a 20Ah capacity 48V rechargeable lithium battery with a cycle life of no less than 1000 cycles. The backup power supply is a 5Ah capacity 12V emergency battery with a battery life of no less than 2 hours, which is suitable for scenarios where there is no external power supply under the bridge. The motion control system monitors the main power supply level in real time. and voltage Sampling frequency 1Hz, when detected or When the power switching logic is triggered automatically, the power switching time is no more than 10ms, and the backup power supply is seamlessly connected. At the same time, the host computer issues an alarm through the LTE-1101J audible and visual alarm. The alarm sound is no less than 85dB and the light flashing frequency is 2Hz. The robot performs an emergency reset operation: the robotic arm returns to the initial position with a pitch angle of 0° and a roll angle of 0° at a speed of 0.1m / s. The robot moves slowly along the double steel wire at a speed of 0.2m / s and moves along the double steel wire towards the starting point of the operation to the preset safe stopping point 5m away.

[0027] Example: like Figure 2 As shown in the figure, this embodiment elaborates on the operation method of the double steel wire inspection robot control system for bridge bottom box girder, including automatic operation mode and manual operation mode. Both modes integrate communication handshake and abnormal handling process to ensure accurate and stable operation.

[0028] I. Implementation Steps for Automatic Operation Mode S1: System initialization and communication handshake; S1.1, the host computer, the slave computer, and the motion control system are powered on and started in sequence. The host computer loads the detection configuration file (including parameters such as mileage, shooting interval, and safety threshold), and the odometer is automatically zeroed. S1.2 The main and backup power supplies perform self-tests. The motion control system detects the voltage and power of the main power supply, and the lower-level machine detects the connection status and initial parameters of peripherals such as dual steel wire tension sensors, encoders, lidar, and cameras. S1.3 Communication handshake process start: The host computer sends a handshake frame to the slave computer (including frame header 0xFE, address code 0x01, random number of life signal 0x00~0xFF, check code, and frame tail 0xA5). After the slave computer verifies that the life signal is consistent with the preset rules, it returns a response frame, completing the handshake between the host computer and the slave computer. Subsequently, the slave computer sends a handshake frame to the motion control system. After the motion control system verifies that the handshake is successful, it returns a response frame, completing the handshake between the slave computer and the motion control system. The entire handshake timeout is 1 second. If the timeout occurs, it will retry 3 times. If the retry fails, an audible and visual alarm will be triggered. S2: Determination of operating conditions; S2.1 The motion control system collects the main power supply power Q and voltage U in real time, determines that Q≥20% and U≥45V, and at the same time confirms that the feedback value of the dual steel wire tension sensor is within the range of 200~300N, the encoder has no signal loss, and the laser radar is measuring distance normally, that is, the peripheral device is in normal status. S2.2 If all the above conditions are met, the system enters the standby state; if any condition is not met, the main and backup power switching is immediately triggered (switching time ≤ 10ms), the host computer issues a continuous 85dB audible and visual alarm, the robot terminates the operation, and the robotic arm returns to the initial position. S3: The robot moves along the double steel wires; S3.1 The host computer sends an automatic operation command. After receiving the command, the motion control system starts the proportional-derivative-integral closed-loop control algorithm and drives the motor to move the robot forward at a constant speed along the double steel wire at the target speed of 0.3m / s. S3.2 The odometer and the passive encoder of the extension arm synchronously collect the moving distance s at a frequency of 10Hz. The encoder resolution is 1024 lines / revolution, the motor reduction ratio is 10, and the actual position of the robot is calculated in real time through the positioning algorithm. The data is uploaded to the lower computer every 500ms, and then synchronized from the lower computer to the upper computer. S4: Precise positioning and deceleration / stop; S4.1 The preset work position interval is 1m. When the collected moving distance s reaches the preset work position s0 (such as 1m, 2m, 3m...), the lower computer issues a deceleration command. S4.2 The motion control system starts the linear deceleration program and decelerates according to the speed formula v(t)=0.3-(0.3 / 0.8)×t (t∈[0,0.8s]) until the robot stops smoothly. When it stops, the vibration amplitude of the steel wire is ≤0.5mm and the positioning error is ≤2mm. S5: Robotic arm posture adjustment and shooting; S5.1 After the robot stops, the robotic arm reads the preset posture sequence (pitch angle 0°, 30°, 60°) from the configuration file, and adjusts the posture sequentially at a speed of 0.1m / s. After each posture positioning is completed (posture accuracy ±0.1°), the high-precision camera is triggered to take a picture. S5.2 The camera captures images of double steel wires and box girder components (coating, welds, connectors) at a resolution of 1920×1080 and a frame rate of 30fps. The captured data is uploaded to the lower computer in real time after being verified by CRC8. The lower computer stores the data and synchronizes it to the upper computer for display. S6: Robotic arm reset; S6.1 After completing the shooting of all postures at the current working position, the robotic arm returns along the original path and resets to the initial position with a pitch angle of 0° and a roll angle of 0°. The reset time is ≤3s. S7: Cyclic operation and abnormal termination; S7.1 After the robotic arm is reset, the robot repeats steps S3-S6 and continues to move along the double steel wire towards the preset work endpoint (e.g., 100m). S7.2 Real-time monitoring of power supply status, communication status, and peripheral device status during operation: If Q<20% or U<45V is detected, immediately switch to backup power and return to the safe docking point; if three consecutive communication verification anomalies are detected, automatically switch to the backup communication link; if the distance of the LiDAR to the obstacle is detected to be <8cm, trigger obstacle avoidance logic or emergency stop; if the preset 100m operation mileage is completed, the robot returns to the starting point safe docking point along the original path, stops the operation, and stores all operation data.

[0029] II. Implementation steps for manual operation mode; T1: Input of operation instructions; T1.1 Users input operation commands through the physical handle of the host computer or the virtual buttons of the UI interface. The physical handle responds first. The correspondence between the buttons and commands is as follows: the left joystick controls forward / backward / turning, the right button controls the start / stop of the robotic arm and attitude adjustment, and the middle button controls taking pictures. T1.2 The virtual button layout is consistent with the physical gamepad. When the Bluetooth communication of the physical gamepad is interrupted, it will automatically switch to virtual button input to ensure control continuity. The button response delay is ≤50ms. T2: Command encapsulation and transmission; T2.1 After receiving the operation command, the host computer encapsulates it into a standardized data frame. The data frame structure is as follows: frame header 0xFE, address code, protocol control code, life signal, service protocol, and checksum (calculated using CRC8 algorithm, polynomial x). 8 +x 5 +x 4 +1, 1 byte), frame end 0xA5; T2.2 After encapsulation, the host computer sends the data frame to the slave computer through TCP port 8080 and UDP port 8081. The sending timeout is 500ms. If the timeout occurs, the data frame will be retransmitted twice. T3: Command parsing, conversion, and distribution; T3.1 After receiving the data frame, the lower-level machine first checks the integrity of the frame header and frame tail, then verifies the check code through the CRC8 algorithm, and finally compares the consistency between the life signal and the value sent by the upper-level machine. T3.2 After verification, the lower-level machine converts the standardized data frame into a serial port protocol message that the motion control system can recognize, and sends it to the motion control system at a baud rate of 9600bps; if the verification fails, the lower-level machine returns an error message to the upper-level machine, and the upper-level machine resends the command. T4: Command execution and motion control; T4.1 After receiving the message, the motion control system parses it through the motion control algorithm module and calls the corresponding control logic according to the instruction type: the forward / backward instruction triggers the motor to run at the target speed of 0.2~0.5m / s, the steering instruction controls the speed difference between the left and right motors to achieve steering, the robotic arm adjustment instruction drives the robotic arm to adjust its posture at a preset speed, and the photo instruction triggers the camera to take a picture. T4.2 The motion control algorithm module adopts PID closed-loop control with a proportional coefficient of 0.2~0.5, an integral coefficient of 0.02~0.1, and a derivative coefficient of 0.05~0.2 to ensure motion accuracy; T5: Data Collection and Feedback; T5.1 The motion control system collects peripheral operating data at a frequency of 10Hz, including motor speed, robot position, double wire tension, battery voltage, robotic arm posture, etc., and simultaneously collects image data captured by the camera. T5.2 After the collected data is encapsulated, it is fed back to the lower computer via RS485 protocol. The lower computer classifies and stores the data and then synchronously uploads it to the upper computer. The feedback delay is ≤100ms. T6: Status monitoring and anomaly handling; T6.1 The host computer displays the equipment's operating status, video feed, and sensor data in real time, and synchronously stores job data and operation logs in JSON format. The logs include information such as instruction type, execution time, and equipment status. T6.2 The host computer monitors the communication link status in real time. If three consecutive communication verification anomalies are detected, the main and backup communication links will be automatically switched. If the anomalies persist after the switch, or if a peripheral device failure or power failure is detected, an emergency stop command will be issued immediately, the robot will stop moving, the robotic arm will return to its initial position, and the host computer will issue an audible and visual alarm.

[0030] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention; no reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A control system for a double-wire inspection robot for bridge bottom box girders, characterized in that: The system includes a host computer, a slave computer, and a motion control system. The three components form a closed-loop control network through multi-protocol redundant communication, which is adapted to the double steel wire operation scenario of the box girder under the bridge and realizes full-process control from command input to robot operation execution and data feedback. The host computer serves as the core of user interaction, integrating a control input module, a data and video display module, a data storage module, a communication module, and a log module. The host computer receives user operation commands and sends them to the slave computer. At the same time, it receives the dual-wire status video stream, sensor data, and equipment status information uploaded by the slave computer, and completes data display, storage, and anomaly recording. The lower-level machine is an intermediate transfer layer, which includes a motion control module, a data storage module, a log module, a communication module, and a parsing module. The lower-level machine parses the upper-level machine's instructions and converts them into messages that the motion control system can recognize. At the same time, it collects data from the dual-wire positioning sensor of the motion control system, the robotic arm's posture data, and video frames, and reports them to the upper-level machine after processing. The motion control system uses a microcontroller as the main controller and includes a motion control algorithm module, a data acquisition module, a data storage module, a communication module, and a log module. The motion control system receives instructions from the lower-level machine and drives the robot hardware to perform motion along the double steel wire. It also collects data from peripherals such as encoders, lidar, batteries, and double steel wire tension sensors and feeds it back to the lower-level machine. The host computer and the slave computer establish a main communication link through TCP+UDP protocol and wireless bridge, with wired network cable as backup link. The slave computer communicates with the motion control system through serial port protocol, and the motion control system communicates with peripherals through RS485 / RS232 protocol, forming a multi-layered, highly reliable communication network.

2. The double-wire inspection robot control system for bridge bottom box girder according to claim 1, characterized in that: The host computer's control input module adopts a dual-backup design of physical handle and virtual button, ensuring consistent control logic. The data and video display module includes a robot control submodule, an image viewing and stitching submodule, a video monitoring submodule, and a dual-wire status early warning submodule, which displays the robot arm's posture, captured images, battery voltage, communication status, and abnormal information detected by the dual-wire in real time. The physical handle is the primary input device, which listens for forward, backward, turn, and start / stop button signals of the robotic arm in real time and converts them into control logic. The virtual buttons serve as a backup solution when the physical peripheral fails. They are triggered through the host computer's UI interface to ensure control continuity. The control signals are cached by the data storage module and then sent to the lower-level computer through the communication module.

3. The double-wire inspection robot control system for bridge bottom box girder according to claim 1, characterized in that: The motion control module of the lower-level machine supports seamless switching between manual and automatic modes, and the switching logic is triggered by protocol control codes; In manual mode, discrete commands issued by the host computer are converted into corresponding messages to precisely control the robot's step-by-step movement along the double steel wire, adjust the robotic arm's posture, and trigger photo capture. In automatic mode, after receiving the one-click start command from the host computer, the robot calls the preset double steel wire detection process message and controls the robot to autonomously complete a series of operations such as double steel wire positioning, moving along the steel wire, taking pictures in multiple postures, and uploading data. The lower-level machine's log module monitors the operating status of each functional module in real time, records information such as abnormal instruction transmission, data parsing failure, abnormal tension of the double steel wires, and equipment failure, forming a traceable operation log. The log includes a frame header, address code, protocol control code, exception type, and timestamp.

4. The double-wire inspection robot control system for bridge bottom box girder according to claim 1, characterized in that: The motion control system adopts a four-layer architecture consisting of a perception layer, a decision layer, an execution layer, and a transmission layer, and is adapted to dual-wire working environments. The perception layer accesses peripherals via multiple protocols, including an extended arm passive encoder, LiDAR, dual wire tension sensor, battery voltage sensor, and robotic arm posture sensor, to collect raw data such as motor speed, robot position along the wire, environmental obstacles, dual wire tension, and battery voltage. The decision-making level integrates multi-sensor data through the motion control algorithm module to plan the optimal working path and robotic arm posture sequence along the double steel wire, thereby avoiding interference with the internal structure of the box girder; The execution layer drives the motor and robotic arm to perform movements according to the decision instructions, and controls the high-precision camera to complete the photography of the double steel wire and box girder components; The transport layer enables bidirectional data transmission with the lower-level machine, ensuring real-time command issuance and data feedback with a transmission delay of ≤100ms.

5. The double-wire inspection robot control system for bridge bottom box girder according to claim 4, characterized in that: The motion control algorithm module employs a proportional-derivative-integral closed-loop control algorithm, integrating encoder positioning data, dual-wire tension data, and lidar obstacle avoidance data to achieve precise motion control along the dual-wire. The control signal calculation formula is as follows: ； in, For motor control output signals, This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients. The real-time deviation between the target position and the actual position is calculated by fusing encoder positioning deviation and dual-wire tension compensation deviation. When the lidar detects the distance to the obstacle hour, To preset a safe distance threshold, with a value of 5-10cm, the algorithm automatically triggers obstacle avoidance logic and adjusts the control signal to make the robot shift laterally along the double steel wire or decelerate and stop.

6. The double-wire inspection robot control system for bridge bottom box girder according to claim 1, characterized in that: The communication network architecture is built upon a host computer communication module, a slave computer communication module, and a motion control system communication module, employing multi-link redundancy and frame structure verification design. The main communication link between the host computer and the slave computer communication module, built through a bridge and TCP / UDP protocol, is the core, responsible for high-bandwidth, real-time data transmission of video streams, control commands, etc. The wired network cable between the host computer and the slave computer serves as a backup link, which automatically switches when the main link fails three consecutive checks to ensure that core data is not lost. In the serial port protocol communication between the lower-level communication module and the motion control system communication module, the communication frame includes a frame header, address code, protocol control code, life signal, service protocol, check code and frame tail. The frame header is fixed at 0xFE and the frame tail is fixed at 0xA5. The life signal is a random number generated by the upper-level computer and is used to verify the validity of the communication link. The check code is calculated by the CRC8 algorithm. The architecture integrates a distributed caching function, which temporarily caches the transmitted sensor data and video frames through the data storage modules of the host computer and the slave computer. The caching time is ≥30s, reducing the impact of the complex environment under the bridge on data transmission.

7. The double-wire inspection robot control system for bridge bottom box girder according to claim 1, characterized in that: The host computer's data storage module adopts a dual-structure design of driver layer-physical layer, with one-to-one correspondence between the driver layer and the physical layer, and critical data adopts a dual backup strategy of local + remote backup. The driver layer receives the raw data transmitted by the communication module, parses it, performs CRC8 verification, and encapsulates it into a standardized data format; The physical layer classifies and stores control command data, sensor timing data, video frame data, and log data. Video frame data is named according to timestamp-wire number-work location. Sensor data is stored in blocks according to 10Hz frequency. Video frame data is backed up to local storage devices and remote servers in real time, with a backup delay of ≤5s.

8. The double-wire inspection robot control system for bridge bottom box girder according to claim 1, characterized in that: The robot localization employs a fusion algorithm combining an extended-arm passive encoder and an odometry system, adapting to the precise positioning requirements of dual-wire detection. The formula for calculating the movement distance is as follows: ； in, This represents the actual distance (in meters) the robot travels along the double steel wires. The passive encoder outputs the rotation angle (rad). Let be the radius (m) of the motor drive wheel. Encoder resolution (lines / revolution). This refers to the motor reduction ratio; When the robot has moved 90% of the target position, a linear deceleration program is initiated, and the speed change during deceleration satisfies the following: ( ); in For the target speed, The preset deceleration time is set to 0.5-1s to ensure that the robot stops smoothly along the double steel wires and avoids the vibration of the steel wires affecting the detection accuracy.

9. The double-wire inspection robot control system for bridge bottom box girder according to claim 1, characterized in that: The robot control system is equipped with dual power supply modules: a main power supply and a backup power supply. The main power supply uses a 48V rechargeable lithium battery, and the backup power supply uses a 12V emergency battery, which is suitable for scenarios where there is no external power supply under the bridge. The motion control system monitors the main power supply level in real time. and voltage When detected or When the power switching logic is triggered automatically, the backup power supply is seamlessly connected. At the same time, the host computer issues an audible and visual alarm, and the robot performs an emergency reset operation: the robotic arm returns to the initial position, and the robot slowly moves along the double steel wire to the safe stopping point.

10. The double-wire inspection robot control system for bridge bottom box girder according to any one of claims 1 to 9, characterized in that: The system operates in both automatic and manual modes, and both modes include communication handshake and exception handling procedures. The automatic operation mode includes the following steps: S1. System initialization, odometer zeroing, main and backup power self-test, detection of peripherals such as dual steel wire tension sensor and encoder, communication handshake between host computer and slave computer, and between slave computer and motion control system in sequence, and verification of life signal consistency. S2, Determine the main power supply and If the peripheral device status is normal, or if it does not meet the requirements, the backup power supply will be switched on and an alarm will be issued to terminate the operation; if it meets the requirements, proceed to the next step. S3, The robot moves at the target speed. Moving along two steel wires, the distance traveled is collected in real time by an odometer and a passive encoder. ; S4, when Reach the preset work position At that time, the robot performs a deceleration and stop operation; S5. The robotic arm adjusts sequentially according to the preset posture sequence in the configuration file. After each posture positioning is completed, the camera is triggered to take pictures of the double steel wire and box girder components, and the shooting data is uploaded to the lower computer in real time. S6. After taking photos in all preset poses, the robotic arm resets to its initial position; S7. Repeat steps S3-S6 until the preset working mileage is completed or a power abnormality or communication interruption is detected. The robot then returns to the safe docking point and stops working. The manual operation mode includes the following steps: T1. Users input operation commands via a physical handle or virtual buttons on the host computer. Commands include forward, backward, turn, robotic arm adjustment, and taking a picture. T2. The host computer encapsulates the instructions into standardized data frames, including CRC8 checksums and life signals, and sends them to the slave computer via TCP+UDP protocol. T3: The lower-level machine parses the data frame, and after verifying that the check code and life signal are correct, it converts them into a message that the motion control system can recognize, and sends it down through the serial port protocol. T4. After receiving the message, the motion control system calculates and generates control signals through the motion control algorithm module, which then drive the motor and robotic arm to perform the corresponding actions. T5: The motion control system collects the operating data of the peripheral devices and the status of the double steel wires in real time, and feeds it back to the lower computer via the communication module, and then the lower computer reports it to the upper computer. T6: The host computer displays the equipment's operating status, video feed, and sensor data in real time, and synchronously stores operation data and operation logs. When three consecutive communication verification anomalies are detected, it automatically triggers communication link switching or emergency shutdown.

Images