Collaborative anti-collision system and device of automatic driving cabin cleaning machine and portal crane for cabin

Abstract

The invention discloses an automatic driving cabin cleaning machine and portal crane collaborative anti-collision system and device for a cabin, and relates to the technical field of ship automation control. The cooperative anti-collision system of the automatic driving cabin cleaning machine and the portal crane for the cabin comprises an initial blind area analysis module, a position blind area analysis module and a path conflict analysis module. Whether the position data is obtained or not is judged through the first blind area recognition analysis coefficient of the automatic driving cabin cleaning machine, if the position data is obtained, whether the path planning data is obtained or not is judged based on the second blind area recognition coefficient obtained through the position data, and if the path planning data is obtained, path planning is completed. If yes, a path planning analysis coefficient is obtained through the obtained path planning data, and whether path optimization measures are taken or not is judged according to the path planning analysis coefficient, so that the effect of improving the anti-collision performance of collaborative operation of the automatic driving cabin cleaning machine and the portal crane is achieved. The problem that in the prior art, when an automatic driving cabin cleaning machine and a portal crane cooperatively prevent collision, blind area recognition is not sensitive is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of ship automation control, and in particular to a system and device for cooperatively preventing collisions between an automatic driving cabin cleaning machine and a door crane for a cabin. Background Art

[0002] With the development of large-scale ships and automation technology, cabin cleaning operations and cargo loading and unloading are gradually moving towards intelligence. The collaborative operation of automatic cabin cleaning machines and gantry cranes has become the key to improving operational efficiency and safety. However, the space in the cabin is limited and the environment is complex, and the risk of collision between equipment has become a major hidden danger. Therefore, the automatic cabin cleaning machine and gantry crane collaborative anti-collision system and device designed for the cabin, combined with sensors, path planning and intelligent control technology, can effectively ensure operational safety and realize an unmanned, intelligent and efficient collaborative operation mode.

[0003] Existing cabin cleaning and cargo loading and unloading operations mostly rely on manual operation or semi-automatic equipment, such as independent operation of cabin cleaning machines and gantry cranes. Due to the narrow cabin space and complex environment, path conflicts are prone to occur between operations of equipment. In addition, the existing anti-collision system mainly relies on simple distance alarms, which cannot achieve dynamic path planning and equipment coordination, and it is difficult to meet the needs of modern cabin intelligent operations.

[0004] For example, the invention patent with publication number CN115826589A discloses a method for preventing collisions between vehicles, including: step 1: the control system receives at least one transport instruction from a starting point to a destination; step 2: the control system analyzes and sorts a number of transport instructions; step 3: the control system transmits the transport instructions selected after analysis and sorting to the transport vehicle; step 4: the control system determines whether the current transport vehicle can be transported safely; step 5: the transport vehicle performs the transport task according to the determination result of step 4, so that it can intelligently choose to execute the current transport instruction or replace the current transport instruction and execute other instructions without path interference.

[0005] For example, the invention patent with announcement number CN115576360B announces a method for preventing collisions of a bucket wheel excavator, which includes using a laser detection system to obtain road condition information, making an initial adjustment based on the information, correcting the initial adjustment based on dust particle content and air humidity data, detecting the obstacle position again after a first preset time T1, determining the obstacle position and motion state, and adjusting the operating state again; obtaining the height information of the material pile or obstacle, obtaining the minimum distance between the boom and the material pile or obstacle, correcting the minimum distance based on illumination data and dust particle content data, and comparing the correction result with the first preset safety distance, and the central control PLC adjusts the boom state based on the result.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems: In the prior art, due to the obstruction of the gantry crane boom or cargo accumulation, the automatic driving cargo cleaning machine cannot sense the position of the gantry crane boom, or the gantry crane cannot sense the moving position of the automatic driving cargo cleaning machine, resulting in the problem of insensitivity to blind spot recognition when the automatic driving cargo cleaning machine and the gantry crane cooperate to avoid collision. Summary of the invention

[0007] The embodiments of the present application solve the problem of insensitivity to blind spot recognition when the automatic driving tank cleaning machine and the gantry crane cooperate to avoid collision in the prior art by providing a system and device for the coordinated anti-collision of the automatic driving tank cleaning machine and the gantry crane for the cabin, thereby improving the anti-collision performance of the coordinated operation of the automatic driving tank cleaning machine and the gantry crane.

[0008] An embodiment of the present application provides a collaborative anti-collision system for an automatic cabin cleaning machine and a gantry crane for a cabin, comprising: an initial blind spot analysis module, a position blind spot analysis module and a path conflict analysis module: wherein the initial blind spot analysis module is used to obtain a first blind spot identification analysis coefficient through blind spot identification data of the automatic cabin cleaning machine in the area to be collaboratively operated, and to determine whether to obtain position data based on the first blind spot identification analysis coefficient, and the first blind spot identification analysis coefficient is used to initially quantify the probability that there is no blind spot during the collaborative operation of the automatic cabin cleaning machine and the gantry crane; the position blind spot analysis module is used to determine whether to obtain path planning data based on a second blind spot identification coefficient obtained from the position data if the position data is obtained, and the second blind spot identification coefficient is used to quantify the probability that there is no blind spot based on the position of the actual device; the path conflict analysis module is used to obtain a path planning analysis coefficient through the obtained path planning data if the path planning data is obtained, and to determine whether to take path optimization measures based on the path planning analysis coefficient, and the path planning analysis coefficient is used to quantify the probability that there is no conflict in the collaborative operation path of the automatic cabin cleaning machine and the gantry crane.

[0009] Furthermore, the blind spot identification data includes point cloud density, point cloud range, scanning frequency, image clarity and proportion of obscured area; the scanning frequency indicates the number of times the laser radar scans the collaborative operation area per second; the proportion of obscured area indicates the proportion of the collaborative operation area in the camera's field of view that cannot be perceived due to obstacle obstruction to the total camera's field of view; the position data includes dynamic map data and equipment position data, and the position data indicates the position difference between the gantry crane and the automatic driving cabin cleaning machine; the dynamic map data includes map coverage, map update frequency, map resolution and data sparseness rate; the map coverage indicates the known collaborative operation area in the dynamic map The equipment location data includes positioning error, signal loss duration and relative position error; the positioning error indicates the deviation between the actual position and the perceived position of the automatic cabin cleaning machine; the relative position error indicates the deviation between the actual relative distance between the automatic cabin cleaning machine and the gantry crane and the relative distance perceived by the system; the path planning data includes path overlap rate, obstacle avoidance distance and obstacle avoidance triggering frequency; the path overlap rate indicates the proportion of the intersection area of ​​the automatic cabin cleaning machine and the gantry crane in the path planning to the total path; the obstacle avoidance triggering frequency indicates the frequency at which the automatic cabin cleaning machine and the gantry crane trigger obstacle avoidance actions due to the detection of obstacles.

[0010] Furthermore, a specific method for obtaining the first blind spot identification analysis coefficient is as follows: number the collaborative operation areas and obtain reference blind spot identification weights from a preset database; perform data preprocessing on the blind spot identification data, wherein the data preprocessing includes data standardization and data normalization; obtain the first blind spot identification analysis coefficient based on the blind spot identification data after data preprocessing and the reference blind spot identification weights; the reference blind spot identification weights include point cloud density weights, point cloud range weights, scanning frequency weights, image clarity weights and obscured area ratio weights.

[0011] Furthermore, the specific process of determining whether to obtain location data based on the first blind spot identification and analysis coefficient is as follows: obtain a first blind spot threshold from a preset database and compare it with the first blind spot identification and analysis coefficient: if the first blind spot identification and analysis coefficient is not less than the first blind spot threshold obtained from the preset database, then the location data is not obtained, and a first-level indicator light feedback is performed, and the first-level indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; if the first blind spot identification and analysis coefficient is less than the first blind spot threshold obtained from the preset database, then the location data is obtained through a location data tool to obtain a second blind spot identification coefficient, and the location data tool includes a dynamic map tool and a device location tool, and the dynamic map tool includes a SLAM tool, a lidar, a point cloud tool, a map analysis tool and a log analysis tool, and the device location tool includes an RTK GPS and a log analysis tool.

[0012] Furthermore, the specific process of the second blind spot identification coefficient is as follows: obtaining reference data from a preset database, the reference data including reference location data weights, reference dynamic map weights and reference device location weights; the reference location data weights including dynamic map weights and device location weights; the reference dynamic map weights including map coverage weights, map update frequency weights, map resolution weights and data sparsity weights; the reference device location weights including positioning error weights, signal loss duration weights and relative location error weights; performing data preprocessing on the location data, the data preprocessing is used to de-unitize the location data and unify the dimensions; obtaining a dynamic map coefficient based on dynamic map data and corresponding reference dynamic map weights, the dynamic map coefficient is used to determine the probability that there is no blind spot between the automatic driving cabin cleaning machine and the door crane through a dynamic map; obtaining an equipment location coefficient based on the equipment location data and the corresponding reference equipment location weights, the equipment location coefficient is used to quantify the probability of a blind spot in the equipment location; obtaining a second blind spot identification coefficient according to the dynamic map coefficient, the equipment location coefficient and the reference location data weights.

[0013] Furthermore, the specific process of determining whether to obtain path planning data based on the second blind spot identification coefficient obtained based on the position data is as follows: obtaining the second blind spot threshold from the preset database and comparing it with the second blind spot identification analysis coefficient: if the second blind spot identification analysis coefficient is not less than the second blind spot threshold obtained from the preset database, then the path planning data is not obtained and a secondary indicator light feedback is performed, and the secondary indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; if the second blind spot identification analysis coefficient is less than the second blind spot threshold obtained from the preset database, then the path planning data is obtained to obtain the path planning analysis coefficient; the path planning data is obtained through a path planning tool, and the path planning tool includes MATLAB, equipment operation logs, and obstacle avoidance log records.

[0014] Furthermore, the specific method of obtaining the path planning analysis coefficient is as follows: obtaining reference path planning data and reference path planning weights from a preset database; the reference path planning data includes a path overlap rate limit, a reference obstacle avoidance distance, and an obstacle avoidance trigger frequency limit; the reference path planning weights include a path overlap rate weight, an obstacle avoidance distance weight, and an obstacle avoidance trigger frequency weight; the path planning analysis coefficient is obtained by processing the path planning data and the corresponding reference path planning data separately and multiplying them with the corresponding reference path planning weights.

[0015] Furthermore, the specific limit expression of the path planning analysis coefficient is as follows: ; In the formula, n represents the number of the collaborative operation area, , Indicates the total number of collaborative work areas, represents the path overlap rate of the nth collaborative operation area, represents the obstacle avoidance distance of the nth collaborative operation area, Indicates the obstacle avoidance trigger frequency of the nth collaborative operation area, represents the path overlap rate limit, Indicates the reference obstacle avoidance distance, Indicates the obstacle avoidance trigger frequency limit. represents the path overlap weight, represents the obstacle avoidance distance weight, Indicates the obstacle avoidance trigger frequency weight, Represents the path planning analysis coefficient of the nth collaborative operation area.

[0016] Furthermore, the specific process of judging whether to take path optimization measures according to the path planning analysis coefficient is as follows: obtaining the path planning threshold from the preset database and comparing it with the path planning analysis coefficient: if the path planning analysis coefficient is not less than the path planning threshold obtained from the preset database, no path optimization measures are taken and three-level indicator light feedback is performed, and the three-level indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; if the path planning analysis coefficient is less than the path planning threshold obtained from the preset database, path optimization measures are taken; the path optimization measures include time-space separation strategy and speed timing optimization; the time-space separation strategy means allocating the movement of the automatic driving cabin cleaning machine and the gantry crane to different time periods and collaborative operation areas; the speed timing optimization means reducing the overlap rate between the paths of the automatic driving cabin cleaning machine and the gantry crane by controlling the movement parameters of the automatic driving cabin cleaning machine and the gantry crane, and the movement parameters include speed, acceleration and start-up time.

[0017] An embodiment of the present application provides an apparatus for applying the system for collaborative collision avoidance between an automatic driving tank cleaning machine and a gantry crane for a ship's cabin, including a data acquisition device, a storage device, and a processing device; wherein the data acquisition device is used to collect blind spot identification data, position data, and path planning data, and the data acquisition device includes a laser radar, a camera, a position data tool, and a path planning tool; the storage device is used to store blind spot identification data, position data, and path planning data; the processing device is used to analyze the collaborative collision avoidance between the automatic driving tank cleaning machine and the gantry crane for the ship's cabin according to the data stored in the storage device and take corresponding path optimization measures.

[0018] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. A first blind spot recognition analysis coefficient is obtained through the blind spot recognition data of the automatic driving cabin cleaning machine to determine whether to obtain the position data. If the position data is obtained, the second blind spot recognition coefficient obtained based on the position data is used to determine whether to obtain the path planning data. If the path planning data is obtained, the path planning analysis coefficient is obtained through the obtained path planning data, and it is determined whether to take path optimization measures according to the path planning analysis coefficient, thereby reducing the probability of collision between the automatic driving cabin cleaning machine and the gantry crane, thereby improving the anti-collision performance of the automatic driving cabin cleaning machine and the gantry crane in collaborative operation, and effectively solving the problem of insensitivity to blind spot recognition when the automatic driving cabin cleaning machine and the gantry crane are in collaborative anti-collision in the prior art.

[0019] 2. By numbering the collaborative operation areas and obtaining reference blind spot recognition weights from a preset database, the blind spot recognition data is preprocessed, and then the first blind spot recognition analysis coefficient is obtained based on the preprocessed blind spot recognition data and the reference blind spot recognition weights, so as to more efficiently initially judge the probability of the existence of blind spots, and then make further blind spot judgments more timely.

[0020] 3. By obtaining reference data from a preset database and preprocessing the location data, the dynamic map coefficient is obtained based on the dynamic map data and the corresponding reference dynamic map weight, and the device location coefficient is obtained based on the device location data and the corresponding reference device location weight. Finally, the second blind spot recognition coefficient is obtained according to the dynamic map coefficient, the device location coefficient and the reference location data weight, so as to more accurately judge the probability of the existence of the blind spot, and then realize the evaluation of the path planning to reduce the probability of collision between the automatic driving cabin cleaning machine and the gantry crane. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of a coordinated anti-collision system between an automatic cabin cleaning machine and a door crane for a cabin provided in this application; Figure 2 A schematic diagram showing how the path planning analysis coefficient provided in this application varies with the path overlap rate; Figure 3 A schematic diagram showing the path planning analysis coefficients provided for this application as a function of obstacle avoidance distance; Figure 4 A schematic diagram showing how the path planning analysis coefficient provided in this application changes with the obstacle avoidance trigger frequency. DETAILED DESCRIPTION

[0022] The embodiment of the present application solves the problem of insensitivity to blind spot recognition when the automatic driving cleaning machine and the gantry crane cooperate to avoid collision in the prior art by providing an automatic driving cleaning machine and a gantry crane to avoid collision. The collaborative operation area is numbered, and a reference blind spot recognition weight is obtained from a preset database, and then the blind spot recognition data is preprocessed. Then, a first blind spot recognition analysis coefficient is obtained based on the blind spot recognition data after data preprocessing and the reference blind spot recognition weight. Based on the first blind spot recognition analysis coefficient, it is determined whether to obtain position data. If the position data is obtained, reference data is obtained from a preset database, and data preprocessing is performed on the position data. Then, a dynamic map coefficient is obtained based on the dynamic map data and the corresponding reference dynamic map weight. At the same time, a device position coefficient is obtained based on the device position data and the corresponding reference device position weight. Finally, a second blind spot recognition coefficient is obtained according to the dynamic map coefficient, the device position coefficient and the reference position data weight to determine whether to obtain path planning data. If the path planning data is obtained, a path planning analysis coefficient is obtained through the obtained path planning data, and it is determined whether to take path optimization measures according to the path planning analysis coefficient, thereby improving the performance of the automatic driving cleaning machine and the gantry crane to avoid collision in the collaborative operation.

[0023] The technical solution in the embodiment of the present application is to solve the problem of insensitivity to blind spot recognition when the above-mentioned automatic driving cabin cleaning machine and the door crane cooperate to avoid collision. The overall idea is as follows: The first blind spot recognition and analysis coefficient of the automatic driving cabin cleaning machine is used to determine whether to obtain position data. If the position data is obtained, the second blind spot recognition coefficient obtained based on the position data is used to determine whether to obtain path planning data. If the path planning data is obtained, the path planning analysis coefficient is obtained through the obtained path planning data, and the path planning analysis coefficient is used to determine whether to take path optimization measures, thereby achieving the effect of improving the anti-collision performance of the collaborative operation of the automatic driving cabin cleaning machine and the gantry crane.

[0024] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0025] like Figure 1As shown, it is a schematic diagram of the cooperative anti-collision system of the automatic driving tank cleaning machine and the gantry crane for the cabin provided by the embodiment of the present application. The cooperative anti-collision system of the automatic driving tank cleaning machine and the gantry crane for the cabin provided by the embodiment of the present application includes: an initial blind spot analysis module, a position blind spot analysis module and a path conflict analysis module: wherein the initial blind spot analysis module is used to obtain a first blind spot identification analysis coefficient through the blind spot identification data of the automatic driving tank cleaning machine in the area to be coordinated, and judge whether to obtain the position data based on the first blind spot identification analysis coefficient, and the first blind spot identification analysis coefficient is used to initially quantify the probability that there is no blind spot in the process of the coordinated operation of the automatic driving tank cleaning machine and the gantry crane; the position blind spot analysis module is used to determine whether to obtain path planning data based on the second blind spot identification coefficient obtained from the position data if the position data is obtained, and the second blind spot identification coefficient is used to quantify the probability of no blind spot based on the position of the actual device; the path conflict analysis module is used to obtain the path planning data through the obtained path planning data if the path planning data is obtained, and judge whether to take path optimization measures according to the path planning analysis coefficient, and the path planning analysis coefficient is used to quantify the probability that there is no conflict in the coordinated operation path of the automatic driving tank cleaning machine and the gantry crane.

[0026] In this embodiment, by combining the triple protection of initial blind spot analysis, position blind spot analysis and path conflict analysis, the safety, efficiency and intelligence level of the coordinated operation of the automatic driving tank cleaning machine and the gantry crane in the cabin are significantly improved; and each module works in a linked manner, from blind spot identification to path planning, gradually reducing the uncertainty factors in the coordinated operation of the automatic driving tank cleaning machine and the gantry crane, systematically reducing the collision risk caused by blind spots or path conflicts, and improving the anti-collision performance of the coordinated operation of the automatic driving tank cleaning machine and the gantry crane.

[0027] It should be added that blind spot identification data includes point cloud density, point cloud range, scanning frequency, image clarity and proportion of obscured area. Point cloud range indicates the range of field of view that the LiDAR can cover; scanning frequency indicates the number of times the LiDAR scans the collaborative operation area per second; and proportion of obscured area indicates the proportion of the collaborative operation area in the camera's field of view that cannot be perceived due to obstacles in the camera's field of view.

[0028] The point cloud density is calculated by the number of points in the point cloud file (such as .pcd or .ply) and the volume of the target area. The point cloud range and scanning frequency are directly obtained from the technical documentation of the lidar. The image clarity is obtained by measuring the distance between points using CloudCompare and evaluating the clarity. The proportion of the occluded area is obtained by the sparseness of the point cloud distribution and the detection of the occluded area.

[0029] It should be added that the location data includes dynamic map data and equipment location data. The location data indicates the location difference between the gantry crane and the automatic tank cleaning machine. Dynamic map data includes map coverage, map update frequency, map resolution and data sparsity rate. Map coverage indicates the proportion of known collaborative operation areas in the dynamic map to the total collaborative operation area; equipment location data includes positioning error, signal loss duration and relative position error; positioning error indicates the deviation between the actual position and the perceived position of the automatic tank cleaning machine; relative position error indicates the deviation between the actual relative distance between the automatic tank cleaning machine and the gantry crane and the relative distance perceived by the system.

[0030] Among them, the map coverage rate is obtained by analyzing the point cloud map generated by scanning the environment with lidar and depth camera, the map update frequency is obtained by extracting the records of map update in the equipment operation log, and the map resolution is obtained by point cloud processing tools such as PCL (Point Cloud Library) and MATLAB (Matrix Laboratory); the data sparsity rate is obtained by map data analysis tools (such as Python NumPy); the positioning error is obtained by extracting the position information in the SLAM log and comparing it with the preset working position, the signal loss duration is obtained by extracting the time period of positioning or communication signal loss in the equipment operation log and calculating the total loss duration, and the relative position error is obtained by using RTK GPS (Real-Time Kinematic Global Positioning System) or total station to measure the true relative position between devices.

[0031] It should be added that the path planning data includes path overlap rate, obstacle avoidance distance and obstacle avoidance triggering frequency; the path overlap rate indicates the proportion of the intersection area of ​​the automatic driving cabin cleaning machine and the gantry crane in the path planning to the total path; the obstacle avoidance triggering frequency indicates the frequency with which the automatic driving cabin cleaning machine and the gantry crane trigger obstacle avoidance actions due to the detection of obstacles.

[0032] Among them, the path overlap rate is obtained by using a dynamic environment map (such as a grid map or a local path planning map generated by SLAM (Simultaneous Localization and Mapping)) to analyze whether the paths of the autonomous driving cabin cleaning machine and the door crane overlap in the same area. The obstacle avoidance distance is obtained by using sensors (such as lidar, ultrasonic sensor, depth camera) to measure the real-time distance between the obstacle and the equipment. The obstacle avoidance trigger frequency is obtained by extracting the number of obstacle avoidance triggers in the log and calculating the frequency. The collection and statistics of the above data provide a data basis for subsequent data analysis.

[0033] Furthermore, a specific method for obtaining the first blind spot identification analysis coefficient is as follows: number the collaborative operation areas and obtain reference blind spot identification weights from a preset database; perform data preprocessing on the blind spot identification data, and the data preprocessing includes data standardization and data normalization; obtain the first blind spot identification analysis coefficient based on the blind spot identification data after data preprocessing and the reference blind spot identification weights; the reference blind spot identification weights include point cloud density weights, point cloud range weights, scanning frequency weights, image clarity weights and obscured area ratio weights.

[0034] The specific limiting expression of the first blind spot identification analysis coefficient is as follows: ; In the formula, n represents the number of the collaborative operation area, , Indicates the total number of collaborative work areas, represents the point cloud density of the nth collaborative operation area, Indicates the point cloud range of the nth collaborative operation area (not 0), represents the scanning frequency of the nth collaborative operation area, Indicates the image clarity of the nth collaborative operation area, Indicates the proportion of the blocked area of ​​the nth collaborative operation area, represents the point cloud density weight, represents the point cloud range weight, represents the scanning frequency weight, represents the image clarity weight, Indicates the weight of the proportion of the occluded area, Represents the first blind spot identification analysis coefficient of the nth collaborative operation area.

[0035] In this embodiment, the algorithm combines the blind spot recognition data and the reference blind spot recognition weight for comprehensive analysis to obtain the first blind spot recognition analysis coefficient, wherein the point cloud density, scanning frequency and image clarity are positively correlated with the first blind spot recognition analysis coefficient, and the proportion of the blocked area and the point cloud range are negatively correlated with the first blind spot recognition analysis coefficient; by analyzing the blind spot recognition data, it is helpful to make a more accurate preliminary judgment on whether the automatic driving cabin cleaning machine has a blind spot, thereby judging whether further blind spot judgment is needed; the first blind spot recognition analysis coefficient is used to improve the efficiency of blind spot judgment and reduce the waste of data processing resources.

[0036] Specifically, assuming that the point cloud density weight is 0.2, the point cloud range weight is 0.1, the scanning frequency weight is 0.2, the image clarity weight is 0.2, and the occluded area ratio weight is 0.3, the image clarity here is reflected by the angular resolution of the point cloud image, the unit is degree, and the data change table of the first blind spot recognition analysis coefficient is obtained in combination with the blind spot recognition data, as shown in Table 1: Table 1 Data change table of the first blind spot identification analysis coefficient

[0037] It can be seen from Table 1 that the point cloud density, scanning frequency and image clarity are positively correlated with the first blind spot recognition analysis coefficient, and the proportion of the blocked area and the point cloud range are negatively correlated with the first blind spot recognition analysis coefficient; for example, when the point cloud density is 100 (points / square meter), the scanning frequency is 5 (Hz) and the image clarity is 0.5 (degrees), as the point cloud range increases from 180 (degrees) in the first row to 200 (degrees) in the second row, the proportion of the blocked area increases from 0.3 in the first row to 0.4 in the second row, and the corresponding first blind spot recognition analysis coefficient decreases from 0.765 in the first row to 0.745 in the second row, indicating that the proportion of the blocked area and the point cloud range are negatively correlated with the first blind spot recognition analysis coefficient; and when the blocked area is 0.5 (degrees), the proportion of the blocked area increases from 0.3 in the first row to 0.4 in the second row. When the blocking area ratio is 0.5 and the point cloud range is 250 (degrees), the point cloud density increases from 300 (points / square meter) in the third row to 500 (points / square meter) in the fifth row, the scanning frequency increases from 12 (Hz) in the third row to 20 (Hz) in the fifth row, and the image clarity increases from 1.0 (degrees) in the third row to 2.0 (degrees) in the fifth row. The corresponding first blind spot recognition analysis coefficient increases from 1.031 in the third row to 1.613 in the fifth row, indicating that the point cloud density, scanning frequency and image clarity are positively correlated with the first blind spot recognition analysis coefficient. Table 1 more clearly and intuitively analyzes the influence relationship between the first blind spot recognition analysis coefficient and the blind spot recognition data, thereby obtaining a more accurate first blind spot recognition analysis coefficient.

[0038] Specifically, the parameters involved in the processing of the first blind spot identification and analysis coefficient in this algorithm are interrelated and do not exist independently. Among them, under fixed sensor performance, if the scanning range of the lidar is expanded (such as the lidar coverage distance is increased from 50 meters to 100 meters), the point collection capability of the sensor will be distributed to a larger area, and the number of points per unit volume will decrease, resulting in a decrease in point cloud density; while expanding the point cloud range, the farther the area is, the more likely it is to be blocked by obstacles (such as buildings, vehicles), resulting in an increase in the proportion of blocked areas. Lidars with higher scanning frequencies collect more points in a short time, thereby increasing the point cloud density. When the image clarity is insufficient (such as insufficient light or lower resolution), the identification and judgment of obstructions will be more difficult, affecting the overall occlusion analysis.

[0039] Specifically, the reference blind spot recognition weight is obtained from the preset database, and the point cloud density weight, point cloud range weight, scanning frequency weight, image clarity weight and obscured area ratio weight respectively represent the numerical values ​​of the influence of the point cloud density, point cloud range, scanning frequency, image clarity and obscured area ratio on the first blind spot recognition analysis coefficient. The reference blind spot recognition weight can be directly obtained by looking up the weights corresponding to the point cloud density, point cloud range, scanning frequency, image clarity and obscured area ratio in the preset database. The mapping relationship between the point cloud density, point cloud range, scanning frequency, image clarity and obscured area ratio and the weight is pre-set in the preset database, and this relationship can be regarded as a mapping set. For example, the real-time point cloud density, point cloud range, scanning frequency, image clarity and obscured area ratio are respectively input into the corresponding mapping set to output the corresponding point cloud density weight, point cloud range weight, scanning frequency weight, image clarity weight and obscured area ratio weight. This mapping relationship is one-to-one, that is, each point cloud density, point cloud range, scanning frequency, image clarity and occluded area ratio has a unique corresponding weight. In this example, the values ​​of the reference blind spot recognition weights are limited to between 0 and 1, and the sum of these weights involved is 1.

[0040] Furthermore, the specific process of determining whether to obtain location data based on the first blind spot identification and analysis coefficient is as follows: obtain a first blind spot threshold from a preset database and compare it with the first blind spot identification and analysis coefficient: if the first blind spot identification and analysis coefficient is not less than the first blind spot threshold obtained from the preset database, then the location data is not obtained, and a first-level indicator light feedback is performed, and the first-level indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; if the first blind spot identification and analysis coefficient is less than the first blind spot threshold obtained from the preset database, then the location data is obtained through a location data tool to obtain a second blind spot identification coefficient. The location data tool includes a dynamic map tool and a device location tool. The dynamic map tool includes a SLAM tool, a lidar, a point cloud tool, a map analysis tool, and a log analysis tool. The device location tool includes an RTK GPS and a log analysis tool.

[0041] In this embodiment, through the hierarchical judgment of the first blind spot identification analysis coefficient and the dynamic call of the position data tool, efficient, accurate and hierarchical processing of blind spot identification is achieved, providing strong technical support for the efficient collaborative operation of the automatic driving cabin cleaning machine and the gantry crane in the cabin.

[0042] Specifically, the first blind spot threshold is obtained from a preset database. In a specific embodiment, the blind spot identification data corresponding to the situation where a blind spot in the historical data causes a collision is substituted into the specific restriction expression of the first blind spot identification analysis coefficient to obtain the corresponding data set, and the data set is averaged to obtain the first blind spot threshold.

[0043] Furthermore, the specific process of the second blind spot identification coefficient is as follows: obtain reference data from a preset database, the reference data including reference location data weights, reference dynamic map weights and reference equipment location weights; the reference location data weights including dynamic map weights and equipment location weights; the reference dynamic map weights including map coverage weights, map update frequency weights, map resolution weights and data sparsity weights; the reference equipment location weights including positioning error weights, signal loss duration weights and relative location error weights; perform data preprocessing on the location data, the data preprocessing is used to de-unitize the location data and unify the dimensions; obtain a dynamic map coefficient based on the dynamic map data and the corresponding reference dynamic map weights, the dynamic map coefficient is used to determine the probability that there is no blind spot between the automatic driving cabin cleaning machine and the door crane through the dynamic map; obtain an equipment location coefficient based on the equipment location data and the corresponding reference equipment location weights, the equipment location coefficient is used to quantify the probability of a blind spot in the equipment location; obtain the second blind spot identification coefficient based on the dynamic map coefficient, the equipment location coefficient and the reference location data weights.

[0044] The specific limiting expression of the second blind spot recognition coefficient is as follows: ; ; ; In the formula, n represents the number of the collaborative operation area, , Indicates the total number of collaborative work areas, represents the map coverage of the nth collaborative operation area, Indicates the map update frequency of the nth collaborative operation area, Indicates the map resolution of the nth collaborative operation area, represents the data sparsity rate of the nth collaborative operation area, represents the positioning error of the nth collaborative operation area, Indicates the signal loss duration of the nth collaborative operation area. Indicates the relative position error of the nth collaborative operation area (not 0), represents the dynamic map coefficient of the nth collaborative operation area, represents the equipment location coefficient of the nth collaborative operation area, Represents the dynamic map weight, represents the device location weight, represents the map coverage weight, represents the map update frequency weight, Represents the map resolution weight, represents the data sparsity rate weight, represents the positioning error weight, Indicates the signal loss duration weight, represents the relative position error weight, Represents the second blind spot recognition coefficient of the nth collaborative operation area.

[0045] In this embodiment, the algorithm combines the dynamic map coefficient, the equipment location coefficient and the reference location data weight for comprehensive analysis to obtain the second blind spot recognition coefficient, wherein the second blind spot recognition coefficient increases with the increase of the dynamic map coefficient and the equipment location coefficient, wherein the dynamic map coefficient increases with the increase of the map coverage, the map update frequency and the map resolution, and decreases with the increase of the data sparsity rate; the equipment location coefficient decreases with the increase of the positioning error, the signal loss duration and the relative position error; by analyzing the dynamic map coefficient and the equipment location coefficient separately, it is helpful to analyze and obtain the second blind spot recognition coefficient more comprehensively and accurately. When the second blind spot recognition coefficient is larger, the possibility of the existence of a blind spot in the collaborative operation area is smaller. Through the analysis of the second blind spot recognition coefficient, the probability of the existence of a blind spot between the automatic driving cabin cleaning machine and the gantry crane in the current collaborative operation area is further analyzed, so that corresponding measures are taken to improve the collaborative collision avoidance performance between the automatic driving cabin cleaning machine and the gantry crane.

[0046] Specifically, the parameters involved in the processing of the second blind spot recognition coefficient in this algorithm are interrelated and do not exist independently. Among them, the lower the map coverage, the higher the sparseness rate is usually, because many map areas are not filled with valid data. The higher the map update frequency, the faster the map coverage may reach the ideal value, especially in dynamic environments. In maps with higher resolution, the actual physical range of each grid is smaller, and the map coverage calculation is more detailed. However, due to occlusion and insufficient sampling density, the lower the map coverage, the faster the uncovered areas in dynamic scenes will be filled, and the sparseness rate will decrease. Low, but too high a frequency may result in incomplete data fusion, resulting in an increase in the local sparsity rate; as the signal loss duration increases, the prediction error accumulates, causing the positioning error to increase over time, and the relative position error will grow linearly or even exponentially over time, depending on the sensor drift rate. The positioning error is the result of the cumulative error and is directly affected by the relative position error during each movement. If the relative error is larger, the final positioning error will also increase. The longer the signal loss time, the greater the accumulated offset of the inertial navigation. When the signal is restored, the positioning error may drop briefly, but the relative position error still exists, affecting the subsequent trajectory.

[0047] Specifically, the specific expression of the sigmoid function is as follows: ; in, represents the input of the function, Represents the output of a function.

[0048] Specifically, the dynamic map weight is obtained from the preset database, and the map coverage weight, map update frequency weight, map resolution weight and data sparse rate weight respectively represent the numerical values ​​of the influence of map coverage, map update frequency, map resolution and data sparse rate on the dynamic map coefficient. The dynamic map weight can be directly obtained by looking up the weights corresponding to the map coverage, map update frequency, map resolution and data sparse rate in the preset database. The mapping relationship between map coverage, map update frequency, map resolution and data sparse rate and weight is pre-set in the preset database, and this relationship can be regarded as a mapping set. For example, the real-time map coverage, map update frequency, map resolution and data sparse rate are respectively input into the corresponding mapping set to output the corresponding map coverage weight, map update frequency weight, map resolution weight and data sparse rate weight. This mapping relationship is one-to-one, that is, each map coverage, map update frequency, map resolution and data sparse rate has a unique corresponding weight. In this example, the values ​​of the dynamic map weights are all limited to between 0 and 1, and the sum of the map update frequency weight, map resolution weight and data sparse rate weight is 1.

[0049] Specifically, the device location weight is obtained from a preset database, and the positioning error weight, signal loss duration weight, and relative position error weight respectively represent the numerical values ​​of the influence of the positioning error, signal loss duration, and relative position error on the device location coefficient. The device location weight can be directly obtained by looking up the weights corresponding to the positioning error, signal loss duration, and relative position error in the preset database. The mapping relationship between the positioning error, signal loss duration, relative position error, and weight is pre-set in the preset database, and this relationship can be regarded as a mapping set. For example, the real-time positioning error, signal loss duration, and relative position error are respectively input into the corresponding mapping set to output the corresponding positioning error weight, signal loss duration weight, and relative position error weight. This mapping relationship is one-to-one, that is, each positioning error, signal loss duration, and relative position error has a unique corresponding weight. In this example, the values ​​of the device location weight are all limited to between 0 and 1, and the sum of the positioning error weight, signal loss duration weight, and relative position error weight is 1.

[0050] Specifically, the reference location data weight is obtained from the preset database, and represents the numerical value of the degree of influence of the device location coefficient on the second blind spot identification coefficient. This weight can be obtained directly by looking up the weight corresponding to the device location coefficient in the preset database. The mapping relationship between the device location coefficient and the weight is pre-set in the preset database, and this relationship can be regarded as a mapping set. For example, by inputting the real-time device location coefficient into this mapping set, the corresponding weight value can be found. This mapping relationship is one-to-one, that is, each device location coefficient has a unique corresponding weight. In this example, the value of the reference location data weight is limited to between 0 and 1.

[0051] Specifically, in this example, the sum of the reference location data weight and the reference dynamic map weight is 1.

[0052] Furthermore, the specific process of determining whether to obtain path planning data based on the second blind spot identification coefficient obtained from the position data is as follows: obtain the second blind spot threshold from the preset database and compare it with the second blind spot identification analysis coefficient: if the second blind spot identification analysis coefficient is not less than the second blind spot threshold obtained from the preset database, then the path planning data is not obtained and a secondary indicator light feedback is performed, and the secondary indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; if the second blind spot identification analysis coefficient is less than the second blind spot threshold obtained from the preset database, then the path planning data is obtained to obtain the path planning analysis coefficient; the path planning data is obtained through a path planning tool, and the path planning tool includes MATLAB, equipment operation logs, and obstacle avoidance log records.

[0053] In this embodiment, the intelligence and safety of the path planning for the collaborative operation of the autonomous driving cabin cleaning machine and the gantry crane are improved through dynamic judgment of the second blind spot recognition coefficient and efficient calling of the path planning tool; hierarchical decision-making is used to reduce system resource consumption, while ensuring the accuracy and real-time nature of path planning, providing important guarantees for efficient and safe collaborative operations.

[0054] Specifically, the second blind spot threshold is obtained from a preset database. In a specific embodiment, the position data corresponding to the situation where a collision occurs due to a blind spot in the historical data is substituted into the specific restriction expression of the second blind spot identification analysis coefficient to obtain the corresponding data set, and the data set is averaged to obtain the second blind spot threshold.

[0055] Furthermore, the specific method of obtaining the path planning analysis coefficient is as follows: obtain reference path planning data and reference path planning weights from a preset database; the reference path planning data includes path overlap rate limit, reference obstacle avoidance distance and obstacle avoidance trigger frequency limit; the reference path planning weights include path overlap rate weight, obstacle avoidance distance weight and obstacle avoidance trigger frequency weight; the path planning analysis coefficient is obtained by processing the path planning data and the corresponding reference path planning data separately and multiplying them with the corresponding reference path planning weights.

[0056] The specific limit expression of the path planning analysis coefficient is as follows: ; In the formula, n represents the number of the collaborative operation area, , Indicates the total number of collaborative work areas, represents the path overlap rate of the nth collaborative operation area, represents the obstacle avoidance distance of the nth collaborative operation area, Indicates the obstacle avoidance trigger frequency of the nth collaborative operation area, represents the path overlap rate limit, Indicates the reference obstacle avoidance distance, Indicates the obstacle avoidance trigger frequency limit. represents the path overlap weight, represents the obstacle avoidance distance weight, Indicates the obstacle avoidance trigger frequency weight, Represents the path planning analysis coefficient of the nth collaborative operation area.

[0057] In this embodiment, the algorithm combines the path planning data and the corresponding reference path planning data and the reference path planning weight for comprehensive analysis to obtain the path planning analysis coefficient, wherein, as the path overlap rate increases, the corresponding path planning analysis coefficient gradually decreases, as the obstacle avoidance distance gets closer to the reference obstacle avoidance distance, the corresponding path planning analysis coefficient gradually increases, and as the obstacle avoidance trigger frequency increases, the corresponding path planning analysis coefficient decreases accordingly; by analyzing the path overlap rate, obstacle avoidance distance and obstacle avoidance trigger frequency, the possibility of a collision between the automatic driving cabin cleaning machine and the gantry crane in path planning is more accurately analyzed, so that corresponding measures can be taken in time to reduce the probability of a collision between the automatic driving cabin cleaning machine and the gantry crane in path planning, thereby improving the anti-collision performance of the collaborative operation of the automatic driving cabin cleaning machine and the gantry crane.

[0058] Specifically, assuming that the path overlap rate limit is 0.8, the reference obstacle avoidance distance is 3 (meters), the obstacle avoidance trigger frequency limit is 4 (times / hour), the path overlap rate weight is 0.3, the obstacle avoidance distance weight is 0.5, and the obstacle avoidance trigger frequency weight is 0.2. Figure 2 As shown, it is a schematic diagram of the path planning analysis coefficient provided in the embodiment of the present application changing with the path overlap rate. When the obstacle avoidance distance is 3 (meters) and the obstacle avoidance trigger frequency is 2 (times / hour), as the path overlap rate increases, the image shows a downward trend, and the path planning analysis coefficient decreases accordingly, indicating that the higher the path overlap rate, the greater the possibility of conflict in path planning, and the greater the possibility of a collision between the automatic driving cabin cleaning machine and the door crane.

[0059] like Figure 3 As shown, it is a schematic diagram of the change of the path planning analysis coefficient with the obstacle avoidance distance provided in the embodiment of the present application. When the path overlap rate is 0.4 and the obstacle avoidance trigger frequency limit is 2 (times / hour), as the obstacle avoidance distance gets closer to the reference obstacle avoidance distance, the image shows an upward trend, and the path planning analysis coefficient increases accordingly, indicating that the closer the obstacle avoidance distance is to the reference obstacle avoidance distance, the smaller the possibility of conflict in the path planning, and the smaller the possibility of a collision between the automatic driving cabin cleaning machine and the door crane.

[0060] like Figure 4 As shown, it is a schematic diagram of the path planning analysis coefficient provided in the embodiment of the present application changing with the obstacle avoidance triggering frequency. When the path overlap rate is 0.4 and the obstacle avoidance distance is 3 (meters), as the obstacle avoidance triggering frequency increases, the image shows a downward trend, and the path planning analysis coefficient decreases accordingly, indicating that the higher the obstacle avoidance triggering frequency, the greater the possibility of conflict in path planning, and the greater the possibility of a collision between the automatic driving cabin cleaning machine and the door crane.

[0061] Specifically, the reference path planning weight is obtained from the preset database, and the path overlap rate weight, obstacle avoidance distance weight and obstacle avoidance trigger frequency weight respectively represent the numerical values ​​of the influence of the path overlap rate, obstacle avoidance distance and obstacle avoidance trigger frequency on the path planning analysis coefficient. The reference path planning weight can be directly obtained by looking up the weights corresponding to the path overlap rate, obstacle avoidance distance and obstacle avoidance trigger frequency in the preset database. The mapping relationship between the path overlap rate, obstacle avoidance distance and obstacle avoidance trigger frequency and the weight is pre-set in the preset database, and this relationship can be regarded as a mapping set. For example, the real-time path overlap rate, obstacle avoidance distance and obstacle avoidance trigger frequency are respectively input into the corresponding mapping set to output the corresponding positioning error weight, signal loss duration weight and relative position error weight. This mapping relationship is one-to-one, that is, each path overlap rate, obstacle avoidance distance and obstacle avoidance trigger frequency has a unique corresponding weight. In this example, the value of the reference path planning weight is limited to between 0 and 1, and the sum of the path overlap rate weight, obstacle avoidance distance weight and obstacle avoidance trigger frequency weight is 1.

[0062] Specifically, the parameters involved in the processing of the path planning analysis coefficients in this algorithm are interrelated and do not exist independently. In a dynamic environment, when the position of an obstacle changes or a new obstacle appears, the path planning needs to be adjusted frequently, resulting in a decrease in the path overlap rate and an increase in the obstacle avoidance triggering frequency. A longer obstacle avoidance distance allows the system to avoid obstacles in advance and reduce the number of triggers, but an obstacle avoidance distance that is too long may cause the planned path to detour too much, affecting the path efficiency. A longer obstacle avoidance distance provides a larger buffer zone, and the system path planning is more stable with a higher overlap rate. When the obstacle is fixed, the obstacle avoidance triggering frequency is low, the path planning is more stable, and the path overlap rate is higher. When the obstacles change frequently or new obstacles appear, the obstacle avoidance triggering frequency increases and the path overlap rate decreases.

[0063] Specifically, the reference path planning data is obtained from a preset database. In a specific embodiment, the reference path planning data is set by professional dock staff according to the specific operation conditions of the cabin.

[0064] Furthermore, the specific process of judging whether to take path optimization measures according to the path planning analysis coefficient is as follows: the path planning threshold is obtained from the preset database and compared with the path planning analysis coefficient: if the path planning analysis coefficient is not less than the path planning threshold obtained from the preset database, no path optimization measures are taken and three-level indicator light feedback is performed, and the three-level indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; if the path planning analysis coefficient is less than the path planning threshold obtained from the preset database, path optimization measures are taken; the path optimization measures include time-space separation strategy and speed timing optimization; the time-space separation strategy means allocating the movement of the automatic driving cabin cleaning machine and the gantry crane to different time periods and collaborative operation areas; the speed timing optimization means reducing the overlap rate between the paths of the automatic driving cabin cleaning machine and the gantry crane by controlling the movement parameters of the automatic driving cabin cleaning machine and the gantry crane, and the movement parameters include speed, acceleration and start-up time.

[0065] In this embodiment, the path planning analysis coefficient is used to dynamically determine whether to take path optimization measures, and the time-space separation strategy and speed timing optimization are flexibly used to significantly improve the efficiency, safety and economy of equipment collaborative operations; at the same time, the system can accurately determine optimization needs and adjust the optimization plan in real time, providing reliable and efficient path planning and optimization solutions for complex operation scenarios.

[0066] Specifically, the path planning threshold is obtained from a preset database. In a specific embodiment, the path planning data corresponding to the situation where there is a blind spot in the historical data causing a collision is substituted into the specific restriction expression of the path planning analysis coefficient to obtain the corresponding data set, and the data set is averaged to obtain the path planning threshold.

[0067] An embodiment of the present application provides an application such as a device for an automatic cabin cleaning machine and a gantry crane collaborative anti-collision system for a ship's cabin, including a data acquisition device, a storage device and a processing device; wherein the data acquisition device is used to collect blind spot identification data, position data and path planning data, and the data acquisition device includes a laser radar, a camera, a position data tool and a path planning tool; the storage device is used to store blind spot identification data, position data and path planning data; the processing device is used to analyze the automatic cabin cleaning machine and the gantry crane collaborative anti-collision system for the ship's cabin according to the data stored in the storage device and take corresponding path optimization measures.

[0068] To summarize, the embodiment of the present application obtains a first blind spot recognition analysis coefficient through the blind spot recognition data of the automatic driving cabin cleaning machine to determine whether to obtain the position data. If the position data is obtained, then the second blind spot recognition coefficient obtained based on the position data is used to determine whether to obtain the path planning data. If the path planning data is obtained, then the path planning analysis coefficient is obtained through the obtained path planning data, and it is determined whether to take path optimization measures based on the path planning analysis coefficient, thereby reducing the probability of collision between the automatic driving cabin cleaning machine and the gantry crane, thereby improving the anti-collision performance of the automatic driving cabin cleaning machine and the gantry crane in collaborative operation, and effectively solving the problem of insensitivity to blind spot recognition when the automatic driving cabin cleaning machine and the gantry crane cooperate to avoid collision in the prior art.

[0069] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0070] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0071] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0072] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An automatic cabin cleaning machine and door crane collaborative anti-collision system for a cabin, characterized in that: Including initial blind spot analysis module, location blind spot analysis module and path conflict analysis module: The initial blind spot analysis module is used to obtain a first blind spot identification analysis coefficient through blind spot identification data of the automatic driving tank cleaning machine in the area to be coordinated, and determine whether to obtain position data based on the first blind spot identification analysis coefficient. The first blind spot identification analysis coefficient is used to initially quantify the probability that there is no blind spot during the coordinated operation of the automatic driving tank cleaning machine and the gantry crane. The position blind spot analysis module is used to determine whether to obtain path planning data based on a second blind spot identification coefficient obtained from the position data if the position data is obtained, and the second blind spot identification coefficient is used to quantify the probability of no blind spot based on the position of the actual device; The path conflict analysis module is used to obtain the path planning data, obtain the path planning analysis coefficient through the acquired path planning data, and determine whether to take path optimization measures according to the path planning analysis coefficient. The path planning analysis coefficient is used to quantify the probability that there is no conflict in the collaborative operation path of the automatic driving cleaning machine and the gantry crane.

2. The automatic driving cabin cleaning machine and door crane cooperative anti-collision system for cabins according to claim 1, characterized in that: The blind spot identification data includes point cloud density, point cloud range, scanning frequency, image clarity and proportion of blocked area; The scanning frequency refers to the number of times the laser radar scans the collaborative operation area per second; The blocked area ratio indicates the ratio of the collaborative operation area in the camera field of view that cannot be perceived due to obstacles to the total camera field of view; The position data includes dynamic map data and equipment position data, and the position data represents the position difference between the gantry crane and the automatic driving tank cleaning machine; The dynamic map data includes map coverage, map update frequency, map resolution and data sparseness rate; The map coverage rate indicates the ratio of the known collaborative operation area to the total collaborative operation area in the dynamic map; The device location data includes positioning error, signal loss duration and relative position error; The positioning error represents the deviation between the actual position and the perceived position of the automatic driving tank cleaning machine; The relative position error represents the deviation between the actual relative distance between the automatic driving cleaning machine and the door crane and the relative distance perceived by the system; The path planning data includes path overlap rate, obstacle avoidance distance and obstacle avoidance trigger frequency; The path overlap rate indicates the proportion of the intersection area of ​​the automatic driving cleaning machine and the door crane in the path planning to the total path; The obstacle avoidance trigger frequency indicates the frequency at which the automatic driving cabin cleaning machine and the door crane trigger the obstacle avoidance action due to the detection of obstacles.

3. The automatic driving cabin cleaning machine and door crane cooperative anti-collision system for cabins according to claim 2, characterized in that: The specific method for obtaining the first blind spot identification analysis coefficient is as follows: Number the collaborative operation areas and obtain the reference blind area recognition weights from the preset database; Performing data preprocessing on the blind spot identification data, wherein the data preprocessing includes data standardization and data normalization; Obtaining a first blind spot recognition analysis coefficient based on the blind spot recognition data after data preprocessing and a reference blind spot recognition weight; The reference blind spot recognition weights include a point cloud density weight, a point cloud range weight, a scanning frequency weight, an image clarity weight, and an obscured area ratio weight.

4. The automatic driving tank cleaning machine and door crane cooperative anti-collision system for a ship cabin as claimed in claim 1, characterized in that: The specific process of determining whether to obtain position data based on the first blind spot identification analysis coefficient is as follows: Get the first blind spot threshold from the preset database and compare it with the first blind spot identification analysis coefficient: If the first blind spot identification analysis coefficient is not less than the first blind spot threshold obtained from the preset database, the position data is not obtained, and a first-level indicator light feedback is performed, wherein the first-level indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; If the first blind spot identification and analysis coefficient is less than the first blind spot threshold obtained from a preset database, the position data is acquired through a position data tool to obtain a second blind spot identification coefficient. The position data tool includes a dynamic map tool and a device position tool. The dynamic map tool includes a SLAM tool, a lidar, a point cloud tool, a map analysis tool and a log analysis tool. The device position tool includes an RTK GPS and a log analysis tool.

5. The automatic driving cabin cleaning machine and door crane cooperative anti-collision system for cabins according to claim 2, characterized in that: The specific process of the second blind spot identification coefficient is as follows: Acquire reference data from a preset database, the reference data including reference location data weight, reference dynamic map weight and reference device location weight; The reference location data weight includes a dynamic map weight and a device location weight; The reference dynamic map weights include map coverage weight, map update frequency weight, map resolution weight and data sparsity weight; The reference device position weight includes a positioning error weight, a signal loss duration weight and a relative position error weight; Performing data preprocessing on the position data, wherein the data preprocessing is used to de-unitize the position data and unify the dimensions; A dynamic map coefficient is obtained based on the dynamic map data and the corresponding reference dynamic map weight, wherein the dynamic map coefficient is used to determine the probability that there is no blind spot between the automatic driving tank cleaning machine and the door crane through the dynamic map; Obtaining a device location coefficient based on the device location data and a corresponding reference device location weight, wherein the device location coefficient is used to quantify the probability of a blind spot existing in the device location; The second blind spot recognition coefficient is obtained according to the dynamic map coefficient, the device position coefficient and the reference position data weight.

6. The automatic driving tank cleaning machine and door crane cooperative anti-collision system for a ship cabin as claimed in claim 1, characterized in that: The specific process of determining whether to obtain path planning data based on the second blind spot recognition coefficient obtained based on the position data is as follows: Get the second blind zone threshold from the preset database and compare it with the second blind zone identification analysis coefficient: If the second blind spot identification analysis coefficient is not less than the second blind spot threshold obtained from the preset database, the path planning data is not obtained and the secondary indicator light feedback is performed, wherein the secondary indicator light feedback is used to feedback that there is no blind spot in the current collaborative operation area; If the second blind spot identification analysis coefficient is less than the second blind spot threshold obtained from the preset database, acquiring path planning data to obtain the path planning analysis coefficient; The path planning data is obtained through a path planning tool, which includes MATLAB, equipment operation logs and obstacle avoidance log records.

7. The automatic driving tank cleaning machine and door crane cooperative anti-collision system for a ship cabin as claimed in claim 6, characterized in that: The specific method of obtaining the path planning analysis coefficient is as follows: Acquire reference path planning data and reference path planning weights from a preset database; The reference path planning data includes a path overlap rate limit, a reference obstacle avoidance distance, and an obstacle avoidance trigger frequency limit; The reference path planning weights include path overlap rate weights, obstacle avoidance distance weights, and obstacle avoidance trigger frequency weights; The path planning analysis coefficient is obtained by processing the path planning data and the corresponding reference path planning data respectively and performing a product operation with the corresponding reference path planning weight.

8. The automatic driving tank cleaning machine and door crane cooperative anti-collision system for a ship cabin as claimed in claim 7, characterized in that: The specific limit expression of the path planning analysis coefficient is as follows: ； In the formula, n represents the number of the collaborative operation area, , Indicates the total number of collaborative work areas, represents the path overlap rate of the nth collaborative operation area, represents the obstacle avoidance distance of the nth collaborative operation area, Indicates the obstacle avoidance trigger frequency of the nth collaborative operation area, represents the path overlap rate limit, Indicates the reference obstacle avoidance distance, Indicates the obstacle avoidance trigger frequency limit. represents the path overlap weight, represents the obstacle avoidance distance weight, Indicates the obstacle avoidance trigger frequency weight, Represents the path planning analysis coefficient of the nth collaborative operation area.

9. The automatic driving cabin cleaning machine and door crane cooperative anti-collision system for cabins according to claim 1, characterized in that: The specific process of judging whether to take path optimization measures according to the path planning analysis coefficient is as follows: Get the path planning threshold from the preset database and compare it with the path planning analysis coefficient: If the path planning analysis coefficient is not less than the path planning threshold obtained from the preset database, no path optimization measures are taken and three-level indicator light feedback is performed, wherein the three-level indicator light feedback is used to provide feedback that there is no blind spot in the current collaborative operation area; If the path planning analysis coefficient is less than the path planning threshold obtained from the preset database, path optimization measures are taken; The path optimization measures include time-space separation strategy and speed timing optimization; The time-space separation strategy means allocating the movements of the automatic clearing machine and the door crane to different time periods and collaborative operation areas; The speed timing optimization means reducing the overlap rate between the paths of the automatic driving cleaning machine and the gantry crane by controlling the motion parameters of the automatic driving cleaning machine and the gantry crane, wherein the motion parameters include speed, acceleration and start time.

10. A device for the coordinated anti-collision system of an automatic cabin cleaning machine and a door crane for a cabin according to any one of claims 1 to 9, characterized in that: Includes data acquisition equipment, storage equipment and processing equipment; Wherein, the data acquisition device is used to collect blind spot identification data, location data and path planning data, and the data acquisition device includes a laser radar, a camera, a location data tool and a path planning tool; The storage device is used to store blind spot identification data, location data and path planning data; The processing device is used to analyze the coordinated collision avoidance between the automatic driving cabin cleaning machine and the door crane for the cabin according to the data stored in the storage device and take corresponding path optimization measures.

Images