Tunnel ROV based on sonar centering cruise method and system

Abstract

The application discloses a tunnel ROV centering cruising method and system based on sonar, and belongs to the technical field of radio navigation. The method comprises the following steps: S1, echo data output by a forward-looking sonar of an ROV traveling along a tunnel is acquired in real time, and adaptive data analysis is performed frame by frame, a normalized intensity matrix after distance compensation, a sonar range and a detection threshold value obtained based on maximum intensity value weighting are extracted therefrom; S2, a horizontal sector is divided into a left half region, a right half region and a forward region, then target searching is performed and left wall distance, right wall distance and forward distance are calculated; and then a lateral movement speed instruction for controlling the ROV to move in the center is obtained by taking the difference between the left wall distance and the right wall distance as the input of a PID controller, and meanwhile, an advance speed instruction is planned according to the adaptive segmentation, and the centering cruising and forward obstacle avoidance are realized by a flight control. The application can enable the ROV to always travel safely near the center line in the tunnel cross section, and improve the safety and autonomy of the inspection operation under the water filling state of the water conveying tunnel.

Description

Technical Field

[0001] This invention belongs to the field of radio navigation technology, specifically relating to a control method and system for positioning and centering navigation in tunnels based on forward-looking sonar signals. Background Technology

[0002] In pumped storage power stations and large-scale water conveyance projects, long-distance, deeply buried, and complex-section water conveyance tunnels have become a common form of critical infrastructure. These tunnels typically operate under high-lift, high-flow conditions, with extremely high requirements for operational safety. The quality of their inner lining, the development of structural cracks, and the status of leakage and siltation directly affect unit efficiency, structural safety, and life-cycle maintenance costs. Due to safety and regulatory requirements, water conveyance tunnels need to undergo regular, detailed inspections and assessments, including local damage detection, lining void identification, siltation volume assessment, and foreign object blockage investigation. However, traditional inspection methods often rely on manual or simple tools to enter after shutdown, venting, and drainage. This approach suffers from significant drawbacks such as long construction periods, substantial downtime losses, harsh working environments (risks of oxygen deficiency, humidity, slippage, and collapse), and high personnel safety risks, making it difficult to meet the engineering requirements of modern power stations for "high-frequency inspections, detailed assessments, and rapid resumption of work."

[0003] With the development of Remotely Operated Vehicles (ROVs) and intelligent sensing technologies, ROV-based "non-stop inspection while filled with water" is gradually becoming a trend. Deploying ROVs directly into tunnels for inspection can avoid large-scale emptying operations, significantly reducing maintenance costs and safety risks. Existing engineering practices have demonstrated the feasibility of this approach in long-distance hydraulic tunnels.

[0004] However, practical research has revealed that most current cases rely on customized heavy-duty platforms, high-precision inertial navigation systems, long-distance fiber optic cables, and dedicated host computer software for data processing and manual interpretation. Sonar (also known as SONAR) information primarily serves offline mapping and structural diagnosis. During navigation, manual operation or engineer experience is dominant, and multibeam sonar echoes are not directly used to form a universal automatic centering control law. While this model is effective in large-scale engineering projects, its high system cost, complex integration, and high technical barriers make it difficult to promote and apply in small and medium-sized power plants or conventional modular ROV platforms. This highlights the real gap in a lightweight, standardized, and reusable method for automatic centering and obstacle avoidance control in tunnels.

[0005] In terms of positioning and navigation, GPS signals are typically unavailable within water conveyance tunnels, and mature external absolute positioning benchmarks are lacking. The system must rely on onboard IMUs, DVLs (Doppler velocimeters), and pressure depth gauges for trajectory calculation. However, inertial units and DVLs inevitably accumulate errors over long distances and durations. When the tunnel walls are smooth, the flow velocity is low, or the beam incidence angle is unsatisfactory, the DVL echo quality deteriorates, beam drops, and measurement instability further weaken navigation accuracy. In narrow tunnels, this positional error not only affects mapping accuracy but also directly impacts whether the tunnel remains within a safe passageway. If the error exceeds the tunnel's allowable offset, it may lead to a misjudgment of centering when the control system is actually close to the tunnel wall. This makes traditional path-tracking strategies relying solely on inertial navigation / DVL tracks insufficient to provide the ROV with sufficiently reliable near-field geometric constraints, failing to guarantee its continued safe navigation close to the tunnel centerline.

[0006] To address this issue, some research and engineering practices have begun to explore using ranging sonar to assist in centering. A typical approach involves deploying single or multiple ranging sonar beams on the left and right sides, or above and below the ROV, to measure the relative wall distance in real time. Based on the "left-right distance difference" or "wall normal distance error," an offset is constructed, and then the lateral thrust is adjusted through simple proportional or PID control to achieve centering correction. This method has certain effects in regular and symmetrical cross sections, but it also has obvious limitations: (1) It relies on fixed geometric assumptions and manually calibrated thresholds. Once the tunnel cross section changes from a circle to a horseshoe shape, from a full cross section to a local variable cross section, or the sonar installation angle, working range, and gain parameters change, the algorithm needs to be recalibrated, resulting in poor generalization ability; (2) The ranging sonar itself only provides sparse distance information, making it difficult to perform robust analysis on complex wall structures (pits, tunnel entrances, lateral components) and multi-target scenarios (walls + silt + obstacles); (3) Some schemes do not make reasonable use of the depth holding and attitude stabilization modes of the flight control system, forcibly coupling longitudinal / depth / lateral control into a single control loop, resulting in highly sensitive control parameters, a large workload for parameter tuning, and problems such as oscillation, overcorrection, or instability in the actual sea trial environment. The work carried out in the existing technology has proven the rationality of "using sonar geometric information to constrain the trajectory" from different perspectives, but it also shows that there is still a technical gap in realizing robust and adaptive multi-beam sonar closed-loop control on a general ROV platform.

[0007] Against this backdrop, given the "deep water, long distance, confined space, and complex structure" of water conveyance tunnels, there is an urgent need for a control strategy that can directly utilize the raw echo data from forward-looking multibeam imaging sonar to adaptively analyze the distances between the left and right walls, the tunnel profile, and information on obstacles ahead, and effectively coordinate with the existing flight control depth / attitude control of ROVs to achieve an integrated control strategy of automatic centering and active obstacle avoidance. Forward-looking multibeam imaging sonar has natural advantages over single-beam ranging sonar in terms of spatial coverage, resolution, and field of view. Its two-dimensional fan-shaped imaging not only includes the angular distribution and intensity structure of the echoes from the left and right walls, but also retains the spatial characteristics of potential obstacles and silted targets, providing a rich information foundation for directly constructing "tunnel centerline perception" and "safe passage corridor estimation" in the image / echo domain. If a robust method for wall contour extraction and symmetry axis estimation can be built at the level of raw sonar data, and transformed into a simple, online offset error signal, and combined with the existing depth hold and attitude hold modes of flight control, and the lateral / direction control is closed-loop separately, then a "plug-and-play" enhancement of tunnel centering flight can be achieved without significantly modifying the existing hardware and control architecture.

[0008] Therefore, the research on the automatic centering cruise method for tunnel ROVs based on forward-looking multibeam imaging sonar has clear engineering needs and research value. On the one hand, it addresses the real-world application scenarios of typical major hydraulic infrastructure projects, solving the pain points of low efficiency and poor safety of traditional manual inspections and the insufficient environmental adaptability of existing ROV inspection technologies. On the other hand, this method involves multidisciplinary issues such as underwater robot environmental perception, multibeam sonar image analysis, unstructured tunnel geometric modeling, and the coordination of constraint control and flight control systems. It provides a scalable technical path for "intelligent navigation and control of underwater robots in GPS-denied and narrow constrained spaces," and has significant theoretical and application prospects for promoting the large-scale application of underwater autonomous systems in enclosed space scenarios such as hydraulic engineering, nuclear power plant cooling water channels, and municipal utility tunnels. Summary of the Invention

[0009] The purpose of this invention is to address the ROV cruise control requirements in the unique environment of deep-buried, long-distance, tortuous, and confined water conveyance tunnels by systematically solving three key problems encountered during cruise: First, how to estimate the distances to the left and right walls and the forward distance in real time from the raw echoes using only the forward-looking multibeam imaging sonar installed at the front of the ROV, without obtaining GPS or other global positioning signals and without relying on externally deployed positioning base stations, and then construct a lateral centering error characterizing the degree of deviation of the ROV from the tunnel centerline, so that the ROV always navigates safely close to the centerline within the tunnel cross-section; Second, considering the acoustic challenges in the engineering site... The sonar's working range, sampling resolution, bit width format, time gain, and transmit power are dynamically adjusted according to water quality, tunnel size, target reflection characteristics, and operational requirements. The challenge lies in how to adaptively parse sonar UDP packets and stably extract distance information for control, ensuring the algorithm's reliability and versatility under complex conditions. Thirdly, at the control execution level, how to effectively coordinate the extracted distance information with the ROV's flight control depth holding, attitude holding, and thruster allocation mechanisms to form a simple yet adjustable lateral centering control and forward obstacle avoidance control algorithm, improving the safety, autonomy, and engineering applicability of inspection operations in water-filled tunnels. To address these three problems in the prior art, this invention provides a sonar-based centering cruise method and system for tunnel ROVs.

[0010] The specific technical solution adopted in this invention is as follows:

[0011] In a first aspect, the present invention provides a sonar-based centering cruise method for tunnel ROVs, comprising:

[0012] S1. During the process of the underwater robot ROV traveling along the tunnel, the echo data output by the forward-looking sonar of the ROV is acquired in real time and adaptive data parsing is performed frame by frame to extract the normalized intensity matrix after distance compensation, the sonar range, and the detection threshold value obtained by weighting based on the maximum intensity value.

[0013] S2. Based on the normalized intensity matrix of the latest echo data frame, the horizontal sector is divided into left, right, and forward regions according to the beam index. Based on the detection threshold value, the sampling line that first exceeds the threshold is searched and the left wall distance, right wall distance, and forward distance are obtained by combining the total number of sampling lines and the sonar range conversion. Then, the lateral centering deviation of the ROV relative to the tunnel center is calculated based on the left wall distance and right wall distance, and the deviation is input into the PID controller to obtain the lateral speed command to control the ROV to move towards the tunnel centerline. At the same time, the forward speed command is obtained by adaptive segmentation planning based on the forward distance. The obtained lateral speed command, forward speed command, target depth, and target heading are sent to the flight control on the ROV in real time to control the ROV to cruise along the tunnel centerline and avoid obstacles forward.

[0014] As a preferred embodiment of the first aspect above, when performing the adaptive data parsing, the parsing process for the echo data frame with an unknown data format in the first frame is as follows:

[0015] In all candidate bit-width formats, the strategy of prioritizing the verification of high bit widths and gradually reducing the bit width after verification failure is adopted. Under each bit-width format, the combination of candidate header length and candidate line length is traversed to determine whether it satisfies the dual constraints that the total data length after deducting the header is divisible by the line length and the header synchronization word is matched. The header length, line length and bit-width format that satisfy the dual constraints are locked as parsing parameters and recorded in the buffer for parsing the current and subsequent echo data frames.

[0016] Based on the locked analytical parameters, the header and two-dimensional intensity matrix are extracted from the echo data frame. If the header contains range information, the sonar range is read directly. Otherwise, the round-trip time delay formula for sound waves needs to be used, combined with the sampling frequency of the forward-looking sonar. Total number of sampled lines in the current echo data frame Sonar range Online estimation is performed; at the same time, distance-compensated normalization is applied to the extracted two-dimensional intensity matrix to obtain a normalized intensity matrix. The maximum intensity value is then extracted from the normalized intensity matrix and multiplied by a preset weighting factor to obtain the detection threshold value used for target finding by thresholding method.

[0017] As a preferred embodiment of the first aspect, the normalized intensity matrix contains 512 beams. When dividing the horizontal sector according to the beam index, beams 0-255 are assigned to the left half, beams 256-511 to the right half, and beams 128-383 to the forward region.

[0018] As a preferred embodiment of the first aspect above, the calculation methods for the left wall distance, right wall distance, and forward distance are as follows:

[0019] For the left half-zone, right half-zone, and forward zone, the sampling row number that first exceeds the detection threshold is searched from near to far. The sampling row number is multiplied by the single-row distance step size obtained by dividing the sonar range by the total number of sampling rows to obtain the effective echo distance of the corresponding zone. If there is no sampling row number exceeding the detection threshold in any zone, the effective echo distance of this zone is directly set as the sonar range. Finally, the effective echo distances of the left half-zone, right half-zone, and forward zone are used as the left wall distance, right wall distance, and forward distance, respectively.

[0020] As a preferred embodiment of the first aspect, in the PID controller, the difference obtained by subtracting the right wall distance from the left wall distance is scaled and used as the lateral centering deviation input to the PID controller to characterize the degree of lateral deviation of the ROV relative to the tunnel centerline. The output of the PID controller is inverted and symmetrically limited to obtain the lateral movement speed command that controls the ROV to move towards the tunnel centerline.

[0021] As a preferred embodiment of the first aspect mentioned above, the flight control operates in a depth-fixed mode, with the host computer periodically issuing the target depth and target heading. The flight control utilizes its own depth-holding and attitude-holding functions to maintain the longitudinal attitude and axial orientation of the ROV. Meanwhile, the lateral and forward movements of the ROV are decoupled separately, and the flight control performs centering constraints and forward control based on the lateral speed command and forward speed command.

[0022] As a preferred embodiment of the first aspect above, when performing adaptive segmented planning on the forward speed command, the range of forward distance values ​​is pre-divided into three segments, and speed control is performed on each segment separately:

[0023] If the forward distance is greater than the maximum distance threshold that does not require speed control, the forward speed command is set to proceed at a constant speed according to the preset cruising speed.

[0024] If the forward distance is less than the minimum distance threshold required to approach a stop, then the forward speed command is set to 0;

[0025] If the forward distance is between the maximum distance threshold and the minimum distance threshold, the forward speed command is set to decrease linearly from the cruise speed, with the forward speed decreasing as it gets closer to the minimum distance threshold.

[0026] In a second aspect, the present invention provides a computer program product, including a computer program / instruction, which, when executed by a processor, enables the implementation of the sonar-based centering cruise method for tunnel ROVs as described in any of the first aspects above.

[0027] Thirdly, the present invention provides a computer electronic device, which includes a memory and a processor;

[0028] The memory is used to store computer programs;

[0029] The processor is configured to, when executing the computer program, implement the sonar-based centering cruise method for tunnel ROVs as described in any of the first aspects above.

[0030] Fourthly, this invention provides a tunnel ROV control system based on sonar for centered navigation, comprising an underwater robot ROV and a host computer. A forward-looking multibeam imaging sonar is fixedly installed at the front of the ROV, and the horizontal sector of the sonar completely covers the left and right walls of the tunnel. The ROV is equipped with a flight controller with built-in depth and attitude holding functions. The host computer and the ROV are connected for data and command exchange. The host computer carries a computer program that, when executed, enables the tunnel ROV to perform sonar-based centered navigation as described in any of the first aspects, thereby controlling the ROV to navigate centered along the tunnel and avoid obstacles forward.

[0031] Compared with existing tunnel navigation solutions that rely on GPS, inertial navigation high-precision combined navigation or single-beam ranging sonar, this invention has many beneficial effects and can substantially improve engineering feasibility, applicable environment range and control reliability.

[0032] First, this invention, at the perception level, relies solely on forward-looking multibeam imaging sonar installed at the front of the ROV to construct geometric constraints for centered navigation, eliminating the need for GPS signals, ground base stations, and long-distance fiber optic odometer systems. This makes it suitable for environments lacking external navigation references, such as enclosed, deeply buried, and tortuous water conveyance tunnels. By directly utilizing the echo distribution of the left and right walls and forward obstacle information, it uses the "environment itself" as a reference to achieve self-referential navigation control within the tunnel. This reduces dependence on external infrastructure and complex calibration conditions, facilitating rapid migration and deployment across different power plants and tunnel projects.

[0033] Secondly, the adaptive sonar data parsing method proposed in this invention can automatically identify the current data structure and range parameters when the sonar's operating range, bit width format, sampling resolution, and gain strategy change, without requiring manual reconfiguration of the protocol or writing of specific driver code. By searching the header length, line length, and data bit width, and combining gain normalization and adaptive threshold detection, the distance between the left and right walls and the minimum forward distance are stably extracted, significantly reducing integration complexity and maintenance costs. This enhances the system's versatility in adapting to different models and configurations of forward-looking multibeam sonars, and is more robust and scalable than the traditional "fixed protocol + manual parameter tuning" approach.

[0034] Furthermore, in terms of control strategy, this invention employs lateral PID control based on the distance difference between the left and right walls and segmented velocity planning based on forward distance. All key parameters have clear physical meanings (such as tunnel radius, warning distance, safety distance, maximum speed, etc.), allowing engineers to intuitively set and adjust them according to actual working conditions, avoiding abstract, complex, and difficult-to-explain black-box control structures. The lateral error constructed through the distance difference directly reflects the direction and degree of ROV deviation from the centerline, and PID control achieves smooth correction. Forward segmented velocity planning, while maintaining cruising efficiency, can achieve early deceleration and stopping before collision risks occur, improving the intrinsic safety level during long-distance operation in tunnels. Compared with schemes based solely on single-beam ranging or solely on inertial navigation tracks, this invention can provide continuous and reliable near-field constraint information in complex geometric environments, reducing the risk of "appearing centered but actually hugging the wall."

[0035] Furthermore, this invention achieves a friendly decoupling from existing flight control systems in its system integration architecture: the host computer is responsible for sonar data parsing and generating planar control quantities (forward speed, lateral speed, target heading), while the flight control system continues to handle low-level closed-loop control such as depth hold, attitude stabilization, and thruster allocation, without requiring modification of the flight control firmware or disruption of existing safety mechanisms. Through standard MAVLink or RC overlay interfaces, this invention can be integrated as an "external algorithm module" into existing flight control systems such as ArduSub and conventional modular ROV platforms. This layered decoupling design not only reduces development and verification costs but also facilitates reuse on different platforms, contributing to the formation of a universal and engineered tunnel automatic cruise solution.

[0036] Finally, the technical approach adopted in this invention—"directly extracting geometric features from the multibeam intensity matrix, avoiding complex point cloud reconstruction and heavy SLAM"—significantly reduces computational resource requirements while ensuring accuracy in perceiving tunnel walls and obstacles. This makes it suitable for deployment in common industrial environments, and better meets the requirements for long-term continuous operation and ease of maintenance. In summary, this invention surpasses existing technologies in terms of eliminating the need for external positioning, simple structure, interpretable parameters, ease of adjustment, ease of integration, and high security, demonstrating outstanding engineering application value and widespread potential. Attached Figure Description

[0037] Figure 1 A schematic diagram illustrating the steps of a sonar-based centering cruise method for tunnel ROVs;

[0038] Figure 2 Flowchart for adaptive parsing of sonar data;

[0039] Figure 3 The flowchart shows the extraction process for left wall distance, right wall distance, and forward distance.

[0040] Figure 4 Flowchart of the tunnel centering cruise control;

[0041] Figure 5 This is a schematic diagram of the key functional modules of the centering cruise control system running in the host computer.

[0042] Figure 6 This is a schematic diagram of the structure of a computer electronic device;

[0043] Figure 7 This is a schematic diagram of the control system for a tunnel ROV that uses sonar for centered navigation. Detailed Implementation

[0044] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below. Technical features in various embodiments of the present invention can be combined accordingly without mutual conflict.

[0045] In the description of this invention, it should be understood that when an element is considered to be "connected" to another element, it can be a direct connection to the other element or an indirect connection, i.e., there is an intermediate element. Conversely, when an element is said to be "directly" connected to another element, there is no intermediate element.

[0046] This invention constructs a hierarchical control architecture of forward-looking multibeam sonar, host computer, and depth-controlled flight, enabling automatic centering and obstacle avoidance without external positioning in tunnel scenarios. In this hierarchical control architecture, data is sourced from the forward-looking multibeam sonar sensor, eliminating the influence of long-term error accumulation from other sensors such as inertial navigation systems that could cause trajectory deviations. Furthermore, by searching for header length, line length, and data bit width, adaptive parsing of the multibeam sonar UDP data frame structure is achieved, adapting to different ranges and intensity format variations. Then, this invention combines gain normalization and relative threshold-based adaptive sonar intensity detection to stably extract the distances between the left and right walls of the tunnel and obstacles ahead. Based on a rapid calculation method using the "first exceeding threshold line index" for the minimum distances between the left and right walls and the forward direction, geometric constraints are directly generated from the intensity matrix, avoiding point cloud reconstruction. Finally, the lateral control adopts a distance difference driven PID algorithm, and the forward control adopts a piecewise velocity planning, forming a decoupled control integration strategy in which "the host computer is responsible for planar navigation and the flight control is responsible for depth and attitude stabilization". This allows the host computer to only be responsible for planar control, while depth and attitude are handled by the flight control, which is suitable for forward sonar-assisted tunnel cruise scenarios.

[0047] The following is a detailed description of the specific implementation of the sonar-based centering cruise method for tunnel ROVs of the present invention.

[0048] like Figure 1 As shown, in a preferred embodiment of the present invention, a sonar-based centering cruise method for tunnel ROVs is provided, which includes steps S1 and S2, which are described using a host computer as the execution entity. The specific implementation of the two steps is described below.

[0049] S1. During the process of the underwater robot ROV traveling along the tunnel, the echo data output by the forward-looking sonar of the ROV is acquired in real time and adaptive data parsing is performed frame by frame to extract the normalized intensity matrix after distance compensation, the sonar range, and the detection threshold value obtained by weighting based on the maximum intensity value.

[0050] It should be noted that the tunnel in this invention generally refers to a water conveyance tunnel. The underwater robot (ROV) can travel along the water conveyance tunnel. Its front end is equipped with a forward-looking multibeam sonar. The horizontal sector of the forward-looking multibeam sonar should completely cover the left and right walls of the tunnel, and the vertical sector should cover the area in front of the ROV and the key area at the bottom of the tunnel. The sonar control box sends the echo data to the host computer via UDP. The host computer listens on a fixed port and continuously receives the echo data frames.

[0051] Since the sonar's operating range, bit width format (16-bit or 8-bit), sampling resolution, and gain strategy can all change, traditional methods require writing manual configuration protocols or specific driver code for specific sonar models to identify key information such as UDP data structures and range parameters, and to parse received UDP packets. In this invention, an adaptive data parsing method is designed that can perform blind protocol parsing even when the data format is unknown. Specifically, when performing this adaptive data parsing, upon receiving the first frame of echo data uploaded via UDP, the data format of this echo data frame is unknown; therefore, blind parsing is performed according to the following parsing process:

[0052] S11. The host computer pre-stores the synchronization words, bit width formats, header lengths, and row lengths of common sonars, and records them as candidate bit width formats, candidate header lengths, and candidate row lengths. Among all candidate bit width formats, a strategy of prioritizing the verification of high bit widths and decreasing the bit width level by level after verification failure is adopted. Under each bit width format, the combination of candidate header lengths and candidate row lengths is traversed to determine whether it satisfies the dual constraints that the total data length after deducting the header is divisible by the row length and that the header synchronization word matches. The header length, row length, and bit width format that satisfy the dual constraints are locked as parsing parameters and recorded in the buffer for parsing the current and subsequent echo data frames.

[0053] It should be noted that since the common bit width format for sonar is generally 16 bits or 8 bits, the system follows the 16-bit precision priority principle. It first attempts to set the bit width to 16 bits, then iterates through the combinations of candidate header lengths and candidate line lengths to determine if a combination satisfying the double constraint exists. If no such combination is found in the 16-bit bit width format, it continues to attempt to set the bit width to 8 bits, then iterates through the combinations of candidate header lengths and candidate line lengths to determine if a combination satisfying the double constraint exists. Of course, if other bit width formats exist, the same approach can be used, prioritizing the verification of higher bit widths before lower bit widths.

[0054] It should be noted that there are two constraints in the above double constraint. The first constraint is that the total length of the data after deducting the header is divisible by the line length. The second constraint is that the header can match the synchronization word.

[0055] For the first constraint, assuming the total length of the echo data frame is L and the candidate header length is Header, since the candidate row length Row is determined by the number of candidate beams (nBeams) and the bit width, the value of each echo sample point is usually stored with a fixed bit width. With a 16-bit bit width, each sample point occupies 2 bytes, and with an 8-bit bit width, each sample point occupies 1 byte. Therefore, with a 16-bit bit width, the candidate row length Row = nBeams * 2, and with an 8-bit bit width, the candidate row length Row = nBeams * 1. Thus, the first constraint actually requires determining whether (L - Header) is divisible by Row under the current bit width format, header length (Header), and row length (Row).

[0056] Regarding the second constraint, since the header typically contains a fixed synchronization word (e.g., 0xAA55 or 0x5A5A), if a combination of bit width format, candidate header length (Header), and candidate line length (Row) satisfies the first constraint, the system further reads the bytes at the candidate header position from the data frame based on the header, and then matches them with the synchronization word library of a known sonar model. If a synchronization word for a known sonar model is successfully matched, the system checks whether the combination of bit width format, candidate header length (Header), and candidate line length (Row) in the current combination matches the bit width format, header length, and line length of this known sonar model. If a match is found, the second constraint is considered satisfied. Thus, both constraints are satisfied, and the parsing parameter combination can be locked; otherwise, the next set of parsing parameter combinations needs to be traversed.

[0057] It should also be noted that once a set of parsing parameters is locked and recorded in the cache, there is no need to re-execute the combination traversal process of step S11 above. The parsing parameters in the cache can be directly called to parse the echo data frame. If the parsing fails (e.g., the data cannot be correctly segmented), it is determined that the sonar may have changed in terms of working range, bit width format, sampling resolution and gain strategy, and the combination traversal process of step S11 needs to be re-executed.

[0058] S12. According to the locked analytical parameters, extract the header and two-dimensional intensity matrix from the echo data frame. If the header contains range information, directly read the sonar range. Otherwise, the round-trip time delay formula for sound waves needs to be used, combined with the sampling frequency of the forward-looking sonar. Total number of sampled lines in the current echo data frame Sonar range Online estimation is performed; at the same time, distance-compensated normalization is applied to the extracted two-dimensional intensity matrix to obtain a normalized intensity matrix. The maximum intensity value is then extracted from the normalized intensity matrix and multiplied by a preset weighting factor to obtain the detection threshold value used for target finding by thresholding method.

[0059] It should be noted that if the sonar range cannot be directly read from the header... Then it is necessary to adjust the sonar range. Online estimation can be performed by first combining the sampling frequency of the forward-looking sonar. Total number of sampled lines in the current echo data frame The total time window length for continuous sampling of the echo signal from a single beam (single line data) by the sonar receiver is calculated, and then the corresponding sonar range is estimated using the formula for calculating the round-trip time delay of acoustic waves. Its calculation formula can be expressed as:

[0060]

[0061] The normalized intensity matrix is ​​obtained by first extracting a two-dimensional intensity matrix from the echo data frame and then compensating using the gain information carried in each row of data. For the original extracted two-dimensional intensity matrix, each row carries gain information. The host computer performs distance-compensated normalization processing on the two-dimensional intensity matrix, i.e., time-variable gain (TVG) compensation, to eliminate longitudinal intensity unevenness caused by sound wave propagation attenuation. TVG compensation is an existing technology, and the specific compensation formula will not be elaborated here. After obtaining the normalized intensity matrix through TVG compensation, the maximum intensity value can be extracted from the normalized intensity matrix. And according to the formula (k is a weighting factor, typically k=0.2) Generate adaptive target detection threshold values This detection threshold can be dynamically adjusted according to the acoustic characteristics of the current water area, ensuring consistency in echo judgment standards across different frames and operating conditions. This allows for automatic analysis of echo data and calculation of key parameters. A complete flowchart of step S1 can be found here. Figure 2 As shown, step S2 can be executed based on the calculation results of this step.

[0062] S2. Based on the normalized intensity matrix of the latest echo data frame, the horizontal sector is divided into left, right, and forward regions according to the beam index. Based on the detection threshold value, the sampling line that first exceeds the threshold is searched and the left wall distance, right wall distance, and forward distance are obtained by combining the total number of sampling lines and sonar range conversion. Then, the lateral centering deviation of the ROV relative to the tunnel center is calculated based on the left wall distance and right wall distance, and the lateral centering deviation is input into the PID controller to obtain the lateral speed command to control the ROV to move towards the tunnel centerline. At the same time, the forward speed command is obtained by adaptive segmentation planning based on the forward distance. The obtained lateral speed command, forward speed command, target depth, and target heading are sent to the flight control on the ROV in real time to control the ROV to cruise along the tunnel centerline and avoid obstacles forward.

[0063] It should be noted that the above-described operation of dividing the horizontal sector into a left half, a right half, and a forward zone needs to be optimized based on the total number of beams. In the embodiment of this invention, for 512 beams, the normalized intensity matrix corresponds to 512 beams, and the beam index numbers range from 0 to 511. Therefore, when partitioning the horizontal sector according to the beam index, beams 0-255 are assigned to the left half, beams 256-511 to the right half, and beams 128-383 in the middle are assigned to the forward zone. The left half is used to detect the distance from the ROV to the left wall of the water conveyance tunnel, the right half is used to detect the distance from the ROV to the right wall of the water conveyance tunnel, and the forward zone is used to detect the distance from the ROV to obstacles directly in front, thereby reducing the interference of strong echoes from the tunnel sidewalls on the forward criterion.

[0064] Once the left, right, and forward regions are divided, the target distance of each region can be quickly calculated based on the "first over-threshold row index" by combining the normalized intensity matrix and the detection threshold, thus stably extracting the distance between the left and right walls of the tunnel and the obstacle in front.

[0065] In an embodiment of the present invention, the target distance of each region—the left half, the right half, and the forward region—is quickly calculated based on the "first exceeding threshold row index," specifically the method for calculating the left wall distance, the right wall distance, and the forward distance:

[0066] For the left half, right half, and forward region, the search is performed from near to far to find the first time the detection threshold is exceeded. The sampling row number is then multiplied by the sonar range. Divide by the total number of sampled rows The obtained single-row distance step size yields the effective echo distance of the corresponding region. If no sampling row number in any region exceeds the detection threshold, the effective echo distance of that region is directly set as the sonar range. Finally, the effective echo distances of the left half-region, right half-region, and forward region are used as the left wall distance, right wall distance, and forward distance, respectively. For a detailed procedure for determining the three distances, please refer to [link to relevant documentation]. Figure 3 As shown, the search proceeds sequentially from nearest to farthest, traversing each sampling row in the three regions to determine if there exists an intensity exceeding the threshold. The intensity value, if the intensity in the left region exceeds The beam then records the sampling row number where the first echo appears on the left side of that row. If the right region has more than The beam then records the sampling row number where the first echo appears on the right. If the forward region contains more than The beam then records the sampling row number where the first echo appears. To maintain generality, for any region among the left, right, and forward regions, if the sampling row number of the first recorded echo is idx, then its actual distance... The following formula can be used for conversion:

[0067] (1)

[0068] The distances calculated from the left half, right half, and forward half are thus obtained. These can be recorded as the distance from the left wall. Distance from the right wall and forward distance When no valid echo is detected in a certain direction, the distance in that direction is directly taken as the sonar range. This is equivalent to "no obvious target or obstacle seen," ensuring the continuity and conservatism of the control logic. This process simplifies the multibeam intensity distribution to a geometric quantity directly related to tunnel constraints, eliminating the need for point cloud reconstruction and facilitating real-time control.

[0069] Since the cross-section of a water conveyance tunnel can usually be considered approximately regular (circular, horseshoe-shaped, or smooth variable cross-section) in engineering, when the distance to the left wall is obtained... Distance from the right wall Then, the lateral centering deviation of the ROV relative to the tunnel center can be calculated to effectively reflect the degree to which the ROV deviates from the centerline. In the embodiments of the present invention, since different tunnel sizes and sonar ranges can cause significant differences in the absolute value of this distance deviation, it can be scaled. Therefore, the lateral centering deviation is defined here as the distance from the left wall. Distance from the right wall The scaled value of the obtained difference is used to characterize the degree of lateral deviation of the ROV relative to the tunnel centerline. The lateral centering deviation is defined as follows:

[0070] (2)

[0071] in The scaling factor is selected based on the actual tunnel radius and sonar range to map the distance difference to a reasonable control level.

[0072] The above-mentioned lateral centering deviation The data can be further input to a PID controller, whose output can be inverted and symmetrically limited to obtain a lateral speed command that controls the ROV to move towards the tunnel centerline. Based on The definition is that when the left distance is small and the right distance is large, This indicates that the ROV is biased towards the left wall and should be corrected by shifting it to the right; conversely, it is biased towards the right wall. Therefore, in embodiments of the present invention, the following formula can be used to... The lateral speed command is obtained by inputting it into the PID controller. :

[0073] (3)

[0074] The expression for the aforementioned PID controller maintains a negative feedback relationship where the sign of the lateral centering deviation is opposite to the direction of the lateral speed correction. The PID controller itself is existing technology. The proportional term of the PID is used for rapid correction, the integral term is used to compensate for continuous bias (and integral limiting prevents excessive accumulation leading to system overshoot), and the derivative term is used to suppress high-frequency oscillations and control jitter. To avoid excessive lateral speed, the lateral speed command can be symmetrically limited to ensure controllable movement amplitude in complex tunnel environments and to prevent the introduction of new collision risks.

[0075] In addition to lateral movement control, this invention also requires adaptive segmented planning of forward speed commands. In embodiments of this invention, the forward distance range can be pre-divided into three segments based on two pre-set distance thresholds (the maximum distance threshold without speed control and the minimum distance threshold requiring near-stopping), and speed control can be applied to each segment separately.

[0076] If the forward distance is greater than the maximum distance threshold that does not require speed control, the forward speed command is set to proceed at a constant speed according to the preset cruising speed.

[0077] If the forward distance is less than the minimum distance threshold required to approach a stop, then the forward speed command is set to 0;

[0078] If the forward distance is between the maximum distance threshold and the minimum distance threshold, the forward speed command is set to decrease linearly from the cruise speed, with the forward speed decreasing as it gets closer to the minimum distance threshold.

[0079] Specifically, as a preferred example, based on the forward distance When performing adaptive forward speed planning, the maximum distance threshold without speed control is set to 5m, and the minimum distance threshold for approaching a stop is set to 2m. Therefore, the three-stage cruise speed adjustment mode based on forward distance can be set as follows:

[0080] when When the speed is greater than 5 m, proceed at the preset cruising speed. Proceed at a constant speed;

[0081] when When the distance falls within the 2-5 m range, the forward speed decreases linearly with decreasing distance. The closer to 5m, the closer the forward speed is. ,on the contrary (The closer to 2m, the closer the forward speed is to 0), thus achieving smooth deceleration;

[0082] when When the distance is less than 2 meters, the forward speed is reduced to 0 to avoid the ROV colliding with obstacles or tunnel components ahead. This segmented speed strategy achieves adaptive speed planning by maintaining high speed at a distance, slowing down as it approaches, and coming to a stop when it is almost at its closest point.

[0083] It should also be noted that the aforementioned cruise speed adjustment mode based on forward distance is executed under stable cruise conditions. However, during the ROV startup phase, to overcome hydrostatic resistance and mechanical dead zones, the host computer will output a relatively large forward speed command for a short period as a "static-breaking pulse," causing the ROV to transition from a stationary state to a stable cruise state, and then switch to the cruise speed adjustment mode based on forward distance. Therefore, the overall control flowchart for ROV centered cruise in a tunnel can be found in [reference needed]. Figure 4 As shown.

[0084] It should be noted that, since the flight control system of the underwater ROV itself has underlying control logic such as depth stabilization, attitude stabilization, and thruster allocation, the lateral speed command obtained by the host computer in this invention... The forward speed command is only used to control the ROV's lateral movement in the horizontal plane. In other words, the ROV's flight control operates in depth-hold mode. The host computer periodically issues target depth and heading based on user settings, utilizing the flight control's own depth and attitude hold functions to maintain the ROV's longitudinal attitude and axial pointing. Simultaneously, the ROV's lateral and forward movements are decoupled. The flight control uses the lateral speed command obtained from the PID controller for centering constraints, and the forward speed command obtained from adaptive segmented programming for forward control. Thus, the flight control continues to handle the underlying closed-loop control of depth holding, attitude stabilization, and thruster allocation without modifying the flight control firmware or disrupting existing safety mechanisms. The lateral and forward speed commands generated by this invention are interfaced via a standard MAVLink or RC overlay interface, allowing this invention to be integrated into existing flight control systems such as ArduSub and conventional modular ROV platforms as an external algorithm module.

[0085] It should be noted that steps S1 and S2 above can both run on a host computer as computer programs or software functional modules. The key functional modules of this centering cruise control system can be decomposed into a sonar acquisition module, an adaptive analysis module, a centering and obstacle avoidance control module, and a flight control execution module, such as... Figure 5 As shown.

[0086] Therefore, based on the same inventive concept, such as Figure 6 As shown, the present invention also provides a computer electronic device corresponding to the sonar-based centering cruise method for tunnel ROVs provided in the above embodiments, which includes a memory and a processor and can be used as the aforementioned host computer;

[0087] The memory is used to store computer programs;

[0088] The processor is configured to implement the sonar-based centering cruise method for tunnel ROVs as described above when executing the computer program.

[0089] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0090] Therefore, based on the same inventive concept, the present invention provides a computer-readable storage medium corresponding to the sonar-based centering cruise method for tunnel ROVs. The storage medium stores a computer program, which, when executed by a processor, enables the implementation of the sonar-based centering cruise method for tunnel ROVs as described above.

[0091] Therefore, based on the same inventive concept, the present invention provides a computer program product, including a computer program / instruction, which, when executed by a processor, can realize the sonar-based centering cruise method for tunnel ROVs as described above.

[0092] Specifically, in the computer-readable storage medium of the above three embodiments, the stored computer program is executed by a processor, which can perform the aforementioned steps S1 to S3.

[0093] It is understood that the aforementioned storage media may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Furthermore, the storage media may also be various media capable of storing program code, such as USB flash drives, external hard drives, magnetic disks, or optical discs.

[0094] It is understood that the processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0095] It should also be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. In the embodiments provided in this application, the division of steps or modules in the system and method is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple modules or steps may be combined or integrated together, and a module or step may also be split.

[0096] In another embodiment of the present invention, the host computer of the centering cruise method described in the above-mentioned operating steps S1 and S2 can be further integrated with an underwater ROV to form a tunnel ROV control system based on sonar for centering cruise. In this system, a forward-looking multibeam imaging sonar is fixedly installed at the front end of the ROV, and the horizontal sector of the sonar completely covers the left and right walls of the tunnel. The ROV is equipped with a flight controller with built-in depth and attitude holding functions. The host computer and the ROV form a communication connection to exchange data and commands. The host computer is equipped with a computer program. When the computer program is executed, it can realize the tunnel ROV centering cruise method based on sonar as described above, thereby controlling the ROV to cruise in the center of the tunnel and avoid obstacles forward.

[0097] The present invention will further illustrate the detailed implementation and technical effects of the sonar-based centering cruise method for tunnel ROVs and the sonar-based centering cruise system for tunnel ROVs shown in steps S1 and S2 below through a specific embodiment, so as to facilitate understanding of the essence of the present invention.

[0098] Example

[0099] In this embodiment, a water conveyance tunnel with a radius of approximately 2.5 meters is used as the application scenario. The tunnel ROV control system, based on sonar for centered navigation, is as follows: Figure 7 As shown, the front end of the underwater ROV is fixedly equipped with a Tritech M750d forward-looking multibeam sonar, which provides complete horizontal sector coverage of the left and right walls of the tunnel, and vertical sector coverage of the area in front of the ROV and key areas at the bottom of the tunnel. The ROV itself is equipped with an ArduSub flight controller, which has built-in depth and attitude holding functions; the host computer is deployed on the ground control console, which interacts with the ROV via Ethernet to exchange data and commands, forming a control system.

[0100] The forward-looking multibeam imaging sonar M750d is set to operate at a frequency of 1 MHz and a range of 10 m to acquire forward-looking fan-shaped echo data that balances resolution and real-time performance. The sonar output is transmitted via UDP (UDP port: 192.168.2.90) to a designated port on the host computer (port set to 192.168.4.4). After receiving the first frame of echo data, the host computer, without manually configuring a fixed protocol, directly follows the adaptive data parsing method described in steps S11 and S12 of this invention, referring to... Figure 2 The illustrated process automatically searches for combinations of bit width format, header length, and line length that satisfy dual constraints on the data frame, ultimately determining the resolving parameter configuration (header length, line length, and bit width format) for the 512-beam array. In this embodiment, refer to... Figure 2As shown, there are only two bit-width modes: 16-bit and 8-bit. The search prioritizes the 16-bit mode. If a parameter combination satisfying the dual constraints cannot be found, the search continues with the 8-bit mode. If a parameter combination satisfying the dual constraints still cannot be found, the default settings are executed or a rollback is performed, and the user is notified via an error message for manual verification. Once the parsing parameters are locked, the system immediately switches to cached operation mode, locking the data parsing protocol state. For each echo data frame, the known parsing parameters in the cache are used for direct parsing to reduce computational overhead. However, if the known parsing parameters in the cache fail to correctly segment the data when directly parsing the echo data frame, it is determined that the sonar settings may have changed. The system then exits cached operation mode and re-executes adaptive data parsing to search for parsing parameters. Therefore, when switching between different ranges or sonar internal gain curves, no manual parsing protocol settings are required; the host computer can still correctly parse the current parsing parameters using the data length and gain field.

[0101] When the latest echo data frame is obtained through analysis, the range-compensated normalized intensity matrix and sonar range are extracted from it. and the detection threshold value obtained by weighting based on the maximum intensity value. In this embodiment, within each control cycle, the host computer calculates the global maximum value of the normalized intensity matrix and calculates the detection threshold value using 0.2 as a weighting coefficient. This allows for an adaptive distinction between effective echoes and background noise.

[0102] To facilitate control logic design, this embodiment divides the entire sector into three parts based on the indices 0-511 of the 512 horizontal beams: beams 0-255 form the left region, and beams 256-511 form the right region. In the forward region, one-quarter of the beams at each end are excluded, leaving only the middle sector (beams 128-383) for detecting obstacles directly ahead, thus reducing interference from strong echoes from the tunnel sidewalls on the forward criterion. The host computer processing frequency is set to approximately 10 Hz to match the ROV control cycle. Furthermore, the ROV and host computer are time-synchronized to ensure the accuracy of the experimental data.

[0103] Subsequently, for the left, right, and forward sectors, the sampling rows that first exceed the threshold are searched from near to far, referring to the aforementioned... Figure 3 The process shown involves sequentially traversing each sampling row in the three sectors from nearest to farthest to search for any intensity exceeding the threshold. The intensity value was recorded, and the sampling row number where the first echo appeared was recorded. Then, using the aforementioned general formula The distances calculated for the left half, right half, and forward half were obtained respectively. That is, the distance to the left wall. Distance from the right wall and forward distance Then, using the preset scaling factor C, according to the formula... The calculated lateral offset is normalized and input into the PID controller to generate the lateral speed command. It is sent to the flight controller via MAVLink / RC overlay.

[0104] Additionally, the forward speed during stable cruise. Then according to Adopting a three-stage management approach, when When the speed is greater than 5 m, proceed at the preset cruising speed. Proceed at a constant speed; when When the distance falls within the 2-5 m range, the forward speed decreases linearly with decreasing distance. The closer to 5m, the closer the forward speed is. ,on the contrary The closer to 2m, the closer the forward speed is to 0, achieving smooth deceleration; when When the distance is less than 2 m, the forward speed is reduced to 0 to avoid the ROV colliding with obstacles or tunnel components in front. During the ROV startup phase, the ROV needs to overcome the static resistance by using a short "breaking the static" forward pulse. After startup, it can switch to the control mode of the stable cruise state described above.

[0105] Ultimately, the analysis of the test results in this embodiment shows that, with an ROV tunnel radius of 2.5 meters, the lateral deviation of the ROV relative to the centerline during its movement remains stable within 0.2 meters; the ROV automatically decelerates when the distance to an obstacle is less than 5 meters, and stops moving when the distance is less than 2 meters, effectively avoiding collisions. This verifies the feasibility and effectiveness of the invention, proving that automatic centering and obstacle avoidance control within tunnels can be achieved solely with forward-looking multibeam sonar, demonstrating high engineering practical value.

[0106] The embodiments described above are merely some preferred implementations of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, all technical solutions obtained through equivalent substitution or transformation fall within the protection scope of the present invention.

Claims

1. A sonar-based centering cruise method for tunnel ROVs, characterized in that, include: S1. During the process of the underwater robot ROV traveling along the tunnel, the echo data output by the forward-looking sonar of the ROV is acquired in real time and adaptive data parsing is performed frame by frame to extract the normalized intensity matrix after distance compensation, the sonar range, and the detection threshold value obtained by weighting based on the maximum intensity value. S2. Based on the normalized intensity matrix of the latest echo data frame, the horizontal sector is divided into left, right, and forward regions according to the beam index. Based on the detection threshold value, the sampling line that first exceeds the threshold is searched and the left wall distance, right wall distance, and forward distance are obtained by combining the total number of sampling lines and the sonar range conversion. Then, the lateral centering deviation of the ROV relative to the tunnel center is calculated based on the left wall distance and right wall distance, and the deviation is input into the PID controller to obtain the lateral speed command to control the ROV to move towards the tunnel centerline. At the same time, the forward speed command is obtained by adaptive segmentation planning based on the forward distance. The obtained lateral speed command, forward speed command, target depth, and target heading are sent to the flight control on the ROV in real time to control the ROV to cruise along the tunnel centerline and avoid obstacles forward.

2. The sonar-based centering cruise method for tunnel ROVs as described in claim 1, characterized in that, When performing the adaptive data parsing, for the first frame of echo data with an unknown data format, the parsing process is as follows: In all candidate bit-width formats, the strategy of prioritizing the verification of high bit widths and gradually reducing the bit width after verification failure is adopted. Under each bit-width format, the combination of candidate header length and candidate line length is traversed to determine whether it satisfies the dual constraints that the total data length after deducting the header is divisible by the line length and the header synchronization word is matched. The header length, line length and bit-width format that satisfy the dual constraints are locked as parsing parameters and recorded in the buffer for parsing the current and subsequent echo data frames. Based on the locked analytical parameters, the header and two-dimensional intensity matrix are extracted from the echo data frame. If the header contains range information, the sonar range is read directly. Otherwise, the round-trip time delay formula for sound waves needs to be used, combined with the sampling frequency of the forward-looking sonar. Total number of sampled lines in the current echo data frame Sonar range Online estimation is performed; at the same time, distance-compensated normalization is applied to the extracted two-dimensional intensity matrix to obtain a normalized intensity matrix. The maximum intensity value is then extracted from the normalized intensity matrix and multiplied by a preset weighting factor to obtain the detection threshold value used for target finding by thresholding method.

3. The sonar-based centering cruise method for tunnel ROVs as described in claim 1, characterized in that, The normalized intensity matrix contains 512 beams. When dividing the horizontal sector according to the beam index, beams 0-255 are assigned to the left half, beams 256-511 to the right half, and beams 128-383 to the forward region.

4. The sonar-based centering cruise method for tunnel ROVs as described in claim 1, characterized in that, The calculation methods for the left wall distance, right wall distance, and forward distance are as follows: For the left half-zone, right half-zone, and forward zone, the sampling row number that first exceeds the detection threshold is searched from near to far. The sampling row number is multiplied by the single-row distance step size obtained by dividing the sonar range by the total number of sampling rows to obtain the effective echo distance of the corresponding zone. If there is no sampling row number exceeding the detection threshold in any zone, the effective echo distance of this zone is directly set as the sonar range. Finally, the effective echo distances of the left half-zone, right half-zone, and forward zone are used as the left wall distance, right wall distance, and forward distance, respectively.

5. The sonar-based centering cruise method for tunnel ROVs as described in claim 1, characterized in that, In the PID controller, the difference between the left wall distance and the right wall distance is scaled and used as the lateral centering deviation input to the PID controller to characterize the degree of lateral deviation of the ROV relative to the tunnel centerline. The output of the PID controller is inverted and symmetrically limited to obtain the lateral speed command that controls the ROV to move towards the tunnel centerline.

6. The sonar-based centering cruise method for tunnel ROVs as described in claim 1, characterized in that, The flight control system operates in a depth-holding mode. The host computer periodically sends out the target depth and target heading. The flight control system uses its own depth-holding and attitude-holding functions to maintain the longitudinal attitude and axial orientation of the ROV. At the same time, the lateral movement and forward movement of the ROV are decoupled separately, and the flight control system performs centering constraints and forward control according to the lateral speed command and forward speed command.

7. The sonar-based centering cruise method for tunnel ROVs as described in claim 1, characterized in that, When performing adaptive segmented planning on the forward speed command, the range of forward distance values ​​is pre-divided into three segments, and speed control is performed on each segment separately: If the forward distance is greater than the maximum distance threshold that does not require speed control, the forward speed command is set to proceed at a constant speed according to the preset cruising speed. If the forward distance is less than the minimum distance threshold required to approach a stop, then the forward speed command is set to 0; If the forward distance is between the maximum distance threshold and the minimum distance threshold, the forward speed command is set to decrease linearly from the cruise speed, with the forward speed decreasing as it gets closer to the minimum distance threshold.

8. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it can realize the sonar-based centering cruise method for tunnel ROVs as described in any one of claims 1 to 7.

9. A computer electronic device, characterized in that, Including memory and processor; The memory is used to store computer programs; The processor is configured to, when executing the computer program, implement the sonar-based centering cruise method for tunnel ROVs as described in any one of claims 1 to 7.

10. A tunnel ROV control system based on sonar for centered navigation, characterized in that, The system includes an underwater ROV and a host computer. The ROV is equipped with a forward-looking multibeam sonar fixed at its front end, and the horizontal sector of the sonar completely covers the left and right walls of the tunnel. The ROV is equipped with a flight controller with built-in depth and attitude keeping functions. The host computer and the ROV are connected to communicate to exchange data and commands. The host computer is equipped with a computer program. When the computer program is executed, it can realize the sonar-based centering cruise method of the tunnel ROV as described in any one of claims 1 to 7, thereby controlling the ROV to cruise in the center of the tunnel and avoid obstacles forward.

Images