Robot offline comprehensive detection device and detection method

Abstract

The present invention provides a comprehensive detection device for robot offline and a detection method thereof, which belongs to the field of intelligent device production, and includes: a sound collection device and a video collection device, a detection platform located below a detection module for the robot to run and detect, the detection platform is provided with a plurality of functional detection areas for matching different detection scenes of the robot, and the functional detection area includes at least one of a step obstacle detection area, a black background detection area, a high reflective background detection area and a cliff detection area; the sound collection device, the video collection device, the robot and the control module are communicated and connected. The comprehensive detection device for robot offline provided by the present invention and the detection method thereof are designed with a plurality of relevant functional detection areas according to the needs of simulating the main functional scenes, and a combination of sound and vision is used to realize a comprehensive and all-round detection of the robot offline, and compared with the detection platform for traditional manual detection, the function and performance of the robot can be analyzed more comprehensively and accurately.

Description

Technical Field

[0001] The present invention relates to the field of intelligent equipment production, and in particular to a robot offline comprehensive detection device and a detection method thereof. Background Art

[0002] After the final assembly of robots, such as cleaning robots, is completed, a comprehensive pre-offline inspection is required. The current inspection solution on the production line is mainly to manually place the robot on a test bench, and then use visual and human hearing methods to manually identify the robot's running track and running sound on the test bench to complete the pre-offline inspection.

[0003] The current manual inspection scheme for offline robot inspection has the following main shortcomings:

[0004] First, the inspection projects carried out, such as trajectory detection and abnormal sound detection, mainly rely on the inspection personnel to visually obtain the running trajectory and the human ear to receive the noise, and to confirm the function through personal experience. The judgment standard is too subjective and cannot be well quantified, resulting in poor detection accuracy. Secondly, it is unable to meet the comprehensive inspection of multiple indicators of the robot offline, and can only detect some basic and relatively intuitive simple projects, and the detection functionality is too single. In addition, the manual inspection scheme adopted can only detect whether there is a fault or not, and cannot accurately locate and classify the fault mode and cause of the fault. Summary of the invention

[0005] The present invention provides a robot offline comprehensive detection device and a detection method thereof, which are used to solve the defects of current manual detection, such as low detection accuracy, large workload, and single detection function.

[0006] The present invention provides a comprehensive detection device for offline robots, which mainly includes but is not limited to:

[0007] A detection module, including a sound acquisition device for acquiring a sound signal of the robot and a video acquisition device for acquiring an image signal of the robot;

[0008] A detection platform, located below the detection module, for the robot to run and perform detection, the detection platform is provided with a plurality of functional detection areas for matching different detection scenes of the robot, the functional detection areas including at least one of a step obstacle detection area, a black background detection area, a high-reflective background detection area, and a cliff detection area;

[0009] The control module, the sound acquisition device, the video acquisition device, and the robot are respectively connected to the control module for communication.

[0010] According to the robot offline comprehensive detection device provided by the present invention, a preset distance is maintained between the table top of the detection table and the detection module;

[0011] The step obstacle detection area is arranged with step obstacles, the black background detection area is paved with a black material coating, the high-reflective background detection area is paved with a high-reflective material coating, and the cliff detection area includes at least two adjacent platforms with a height difference greater than a preset height difference.

[0012] The robot offline comprehensive detection device provided by the present invention further includes a display device connected to the control module;

[0013] The display device is used to visually display the audio features of the sound signal obtained during the offline detection of the robot, as well as the environmental perception map reconstruction data, actual operation trajectory and offline comprehensive detection results obtained during the offline detection of the robot.

[0014] The robot offline comprehensive detection device provided by the present invention further includes an abnormal execution mechanism that is communicatively connected to the control module;

[0015] When the control module determines that the offline comprehensive detection result for any robot is abnormal, the control module controls the abnormal execution mechanism to transfer and isolate the any robot to a preset area.

[0016] The present invention also provides a robot offline comprehensive detection method, which is applied to any of the above robot offline comprehensive detection devices, and mainly includes:

[0017] Receiving the sound signal collected by the sound collection device, the image signal collected by the video collection device, and the operation information sent by the robot;

[0018] Performing time synchronization analysis on the sound signal, the image signal and the operation information, and outputting offline comprehensive detection results of the robot;

[0019] The offline comprehensive test results include abnormal sound detection results and capability test results;

[0020] The capability detection result includes one or more of an environment map reconstruction detection result, an actual operation trajectory detection result, a cleaning route planning function detection result, and an edge obstacle avoidance function detection result.

[0021] According to the robot offline comprehensive detection method provided by the present invention, the operation information includes speed information;

[0022] The performing time synchronization analysis on the sound signal, the image signal and the operation information and outputting the offline comprehensive detection result of the robot includes:

[0023] Performing time synchronization on the sound signal and the operation information;

[0024] Extracting audio features of the sound signal;

[0025] Determine the abnormal sound detection result according to the audio feature and the speed information;

[0026] The offline comprehensive detection result is determined by combining the abnormal sound detection result and the capability detection result.

[0027] According to the robot offline comprehensive detection method provided by the present invention, the operation information also includes position information; and determining the abnormal sound detection result according to the audio feature and the speed information includes:

[0028] Determine the detection area where the robot is located at each time point according to the position information, and obtain interval speed information of the robot running in each detection area;

[0029] Dividing the audio feature into a plurality of audio feature segments according to partial image signals collected in different detection areas;

[0030] Using each audio feature segment and the interval speed information, a foreign sound detection vector is generated;

[0031] Inputting the abnormal sound detection vector into an abnormal sound detection model to obtain an interval abnormal sound detection result output by the abnormal sound detection model;

[0032] By traversing all the audio feature segments, the abnormal sound detection result is determined according to the acquired abnormal sound detection results of the multiple intervals.

[0033] According to the robot offline comprehensive detection method provided by the present invention, the environmental map reconstruction detection result is determined based on the following steps:

[0034] Receiving environment perception map reconstruction data sent by the robot during offline detection;

[0035] The environmental perception map reconstruction data is compared with the design map of the detection platform to determine the environmental perception detection result of the robot according to the comparison result.

[0036] According to the robot offline comprehensive detection method provided by the present invention, the actual running trajectory detection result is determined based on the following steps:

[0037] Determine the position information of the robot at each time point according to the operation information;

[0038] Projecting the position information to the global coordinates of the detection platform, and obtaining the position point of the robot on the detection platform at each time point;

[0039] Connect all the position points in chronological order to obtain the actual running trajectory;

[0040] The actual running track is matched with the designed running track of the detection platform to determine the actual running track detection result.

[0041] According to the robot offline comprehensive detection method provided by the present invention, the cleaning route planning function detection result is determined based on the following steps:

[0042] Acquire partial image signals, partial running trajectories, and partial running information of the robot running in each detection area on the detection table;

[0043] Based on the partial image signal, partial operation trajectory and partial operation information, it is determined whether the cleaning action, operation trajectory and operation status of the robot in each detection area meet the preset conditions, and if the preset conditions are not met, an unqualified item is determined.

[0044] According to the robot offline comprehensive detection method provided by the present invention, the edge obstacle avoidance function detection result is determined by adopting the following steps:

[0045] Obtaining the running state and running trajectory of the robot in the step obstacle detection area and the cliff detection area;

[0046] The edge obstacle avoidance function detection result of the robot is determined according to the running state and the running trajectory.

[0047] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the program, it implements any of the above-mentioned robot offline comprehensive detection methods.

[0048] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned robot offline comprehensive detection methods.

[0049] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned robot offline comprehensive detection methods.

[0050] The robot offline comprehensive detection device and detection method provided by the present invention designs multiple relevant detection areas according to the needs of simulating main functional scenarios, and adopts a combination of sound and vision to realize all-round comprehensive detection of the robot offline. Compared with the traditional manual detection test bench, the function and performance of the robot can be analyzed more comprehensively and accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 It is a cross-sectional view of the robot offline comprehensive detection device provided by the present invention.

[0053] Figure 2 It is a top view of the detection platform in the robot offline comprehensive detection device provided by the present invention.

[0054] Figure 3 This is one of the flow charts of the robot offline comprehensive detection method provided by the present invention.

[0055] Figure 4 This is the second flow chart of the robot offline comprehensive detection method provided by the present invention.

[0056] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] It should be noted that in the description of the present invention, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "include one..." do not exclude the existence of other identical elements in the process, method, article or device including the elements. The orientation or position relationship indicated by the terms "upper", "lower" and the like is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0059] Before the robot is assembled and manufactured, it needs to undergo comprehensive testing, which mainly includes the following two aspects:

[0060] 1) Abnormal sound detection mainly involves the detection of some moving parts or active parts of the robot. Taking the cleaning robot as an example, the objects of abnormal sound detection include abnormal noise detection of core moving parts such as driving wheels, motors, and sweeping brushes to ensure that there are no manufacturing defects before the robot goes offline.

[0061] 2) Capability testing, including functional and performance testing, such as environmental perception testing, environmental map reconstruction testing, cleaning route planning function testing, actual operation trajectory testing, edge obstacle avoidance function testing, etc., to ensure that the core functions and performance of the robot meet the design requirements before it goes offline.

[0062] In response to the above-mentioned inspection requirements, the current inspection solution on the production line is to use manual inspection, by placing the robot on the inspection table, and then using visual and human hearing methods to utilize the observed robot's running trajectory and heard running sounds to achieve pre-offline inspection.

[0063] Through the analysis of historical detection, it can be determined that the current manual detection scheme has many defects and shortcomings, which are mainly manifested in the following aspects:

[0064] 1) Since trajectory detection and abnormal sound detection mainly rely on personnel to confirm the function of motion trajectory and noise, there is a defect that the judgment criteria are relatively subjective and cannot be well quantified.

[0065] 2) Manual detection can only detect the functions and performance of trajectory and obstacle avoidance, but cannot detect functions such as environmental perception and map reconstruction.

[0066] 3) Manual abnormal sound detection can only detect whether there is a fault or not, but cannot locate and classify the fault mode of the abnormal sound.

[0067] 4) The production line rhythm requires that multiple sweeping robot products be tested at the same time. There are multiple test stations. Due to the interference of factory environment noise and the mutual interference of noise between test stations, the accuracy of abnormal sound detection is easily affected.

[0068] 5) Manual testing only records results, but no original data is available for tracing and storage, which is not conducive to problem location and optimization.

[0069] In view of this, the present invention provides a new robot offline comprehensive detection device, in order to solve or alleviate to a certain extent the above-mentioned problems existing in the existing manual detection solutions.

[0070] Combine the following Figure 1-Figure 5 The present invention describes a robot offline comprehensive detection device and a detection method thereof. Figure 1 is a cross-sectional view of the robot offline comprehensive detection device provided by the present invention, such as Figure 1 As shown, mainly including but not limited to:

[0071] The detection module specifically includes a sound acquisition device 4 for acquiring the robot sound signal and a video acquisition device 5 for acquiring the robot image signal;

[0072] A detection platform 2, located below the detection module, for the robot 6 to run and perform detection, the detection platform 2 is provided with a plurality of functional detection areas for matching different detection scenes of the robot 6, the functional detection areas including at least one of a step obstacle detection area, a black background detection area, a high-reflective background detection area and a cliff detection area;

[0073] The control module 3 , the sound acquisition device 4 , the video acquisition device 5 , and the robot 6 are respectively connected to the control module 3 in communication.

[0074] Optionally, a box 1 may be provided outside the above-mentioned robot offline comprehensive detection device, and the detection platform 2 may be placed at the bottom of the box 1. When any robot is offline for detection, the robot 6 runs on the detection platform 2 and maintains a communication connection with the control module 3; a plurality of different functional detection areas are defined on the detection platform 2 to match different detection scenarios for offline detection of the robot 6.

[0075] It should be noted in advance that the object detected by the offline comprehensive detection device provided by the present invention is a robot in a broad sense, and is also applicable to related mechanical equipment or electronic equipment similar to robots that need to perform autonomous path planning, movement and operation on a site. For the convenience of description, in the subsequent embodiments, the robot offline comprehensive detection device and its detection method provided by the present invention are described in detail by taking a cleaning robot as an example, which is not regarded as a specific limitation on the scope of protection of the present invention.

[0076] The shape of the box 1 can be a conventional hollow cavity structure such as a cuboid or a cylinder, or can be designed into other shapes according to actual detection needs, or even formed by appropriate modification of a room or a factory building.

[0077] The testing table 2 is placed as a whole inside the box 1. A plurality of different functional detection areas can be set on the table top of the testing table 2 according to the needs of offline detection of the cleaning robot. For example, a step obstacle detection area, a black background detection area, a high-reflective background detection area and a cliff detection area can be set to simulate different actual working environments of the cleaning robot to perform corresponding functional tests.

[0078] The control module 3 can be an industrial computer equipped with a processor and a memory of a certain capacity. The control module 3 performs comprehensive data interaction with the sound acquisition device 4, the video acquisition device 5 and the cleaning robot in a wired or wireless manner. After receiving the collected sound signals, image signals and operation information of the cleaning robot, the control module 3 can comprehensively process and analyze these data to complete the detection of the functions and performance of the cleaning robot.

[0079] The sound collection device 4 may be a common microphone, or a high-sensitivity microphone array may be selected in consideration of the need for detection accuracy, and is mainly used to collect sound signals generated by the cleaning robot during offline detection.

[0080] Generally speaking, the collected sound signals mainly include the normal operation sounds of some moving parts of the cleaning robot during its operation, such as the sound of the motor rotating, the sound generated when the sweeping brush contacts the ground, etc., and may also include external noise generated by the cleaning robot itself.

[0081] However, when the cleaning robot fails, especially when some operating parts / moving parts fail, abnormal sounds are often generated, and the sound signals collected by the sound collection device 4 must contain this part of the abnormal sound information. The sound collection device 4 sends the collected sound signals to the control module 3 in real time, so that the control module 3 analyzes these sound signals and completes the abnormal sound detection of the cleaning robot.

[0082] The video acquisition device 5 can be a conventional camera, and the video acquisition device 5 and the sound acquisition device 4 are generally fixedly mounted on the top of the box 1, so as to be able to completely capture the image signal of the cleaning robot on the entire detection table 2 during the offline detection process. In order to further improve the detection accuracy, a high-resolution camera or a high-speed camera can be used to accurately capture the image signal of the cleaning robot in different functional scenarios.

[0083] The video acquisition device 5 can be connected to the control module 3 using a video cable or a wireless network, and timely convert the collected image signal into an image signal and send it to the control module 3, so that the control module 3 can obtain the operation trajectory of the cleaning robot and combine the operation status of the cleaning robot in different application scenario areas in the detection platform 2 to perform environmental perception detection, actual operation trajectory detection, edge obstacle avoidance function detection and other capability detections.

[0084] The robot offline comprehensive detection device provided by the present invention is designed with multiple relevant detection areas according to the needs of simulating main functional scenarios, and adopts a combination of sound and vision to realize all-round comprehensive detection of the robot offline. Compared with the traditional manual detection test bench, it can analyze the function and performance of the robot more comprehensively and accurately.

[0085] Figure 2 FIG. 1 is a top view of the detection platform in the robot offline comprehensive detection device provided by the present invention, as shown in FIG. Figure 2 As shown, in the present invention, a plurality of different detection areas are defined on the detection platform 2, and the different detection areas are arranged to match different detection scenarios for offline detection of the cleaning robot.

[0086] Specifically, the detection area mainly includes: one or more of a step obstacle detection area 21, a black background detection area 22, a high-reflection background detection area 23 and a cliff detection area 24.

[0087] Among them, the detection platform 2 is placed at the bottom of the box 1, a step obstacle is arranged in the step obstacle detection area 21, a black material coating is laid in the black background detection area 22, a high-reflective material coating is laid in the high-reflective background detection area 23, and the cliff detection area 24 includes at least two adjacent table surfaces with a height difference greater than a preset height difference.

[0088] In the step obstacle detection area 21, step obstacles that may be encountered in a simulated home environment are arranged. These step obstacles can be fixed or adjustable in height to meet the testing requirements of different models of cleaning robots. The height, width and inclination angle of the steps can be set according to standards or specific test requirements to evaluate the obstacle avoidance ability, climbing ability and path planning ability of the cleaning robot when encountering steps.

[0089] In the black background detection area 22, a uniform, highly light-absorbing black material coating is applied. This coating can effectively absorb light, reduce reflection, and simulate low-light or completely dark detection scenarios. In such a background, the cleaning robot can be tested for sensor sensitivity, navigation accuracy, and effective obstacle avoidance strategies in low-light conditions.

[0090] In contrast to the black background detection area 22, the high-reflective background detection area 23 simulates a high-brightness or strong-reflective detection scene by laying a high-reflective material coating. The high-reflective material coating used can efficiently reflect light and simulate smooth surface environments such as glass and mirrors. In this area, the cleaning robot is mainly tested for its visual sensor performance when dealing with strong light reflections, its ability to prevent misjudgment, and whether it can accurately identify the surrounding environment.

[0091] The cliff detection area 24 involved in the present invention is particularly unique, including at least two adjacent countertops with a height difference greater than a preset height difference, simulating a cliff-like edge effect. This design is intended to test the cliff detection function of the cleaning robot when it encounters a seemingly unsupported edge, ensuring that the cleaning robot can identify and avoid falling in time. The edge of this area can be designed with different inclination angles and shapes to comprehensively evaluate the safety performance and boundary recognition ability of the cleaning robot. Among them, the preset height difference is set according to the design requirements of the cleaning robot.

[0092] The robot offline comprehensive detection device provided by the present invention can comprehensively and efficiently evaluate various performance indicators of the cleaning robot before it goes offline, including but not limited to abnormal sound detection, obstacle avoidance ability detection, environmental adaptability detection, sensor sensitivity detection and safety detection, etc., through multiple functional detection areas carefully designed on the detection platform, thereby ensuring that each robot leaving the factory can meet high quality standards.

[0093] In order to further optimize the detection accuracy, especially in terms of noise control, the present invention covers the interior of the box 1 with sound-absorbing material. This design is intended to reduce or eliminate the interference of the noise generated by the robot during the detection process on the external environment.

[0094] The selected sound-absorbing material may be foam plastic, glass fiber, polyester fiber, mineral wool, etc. Taking the case of covering the interior of the cavity structure with polyester fiber as an example, the cut polyester fiber sound-absorbing panels are installed one by one on the inner wall, top surface, and even the ground of the box 1, ensuring that each sound-absorbing panel is tightly connected without gaps to achieve the best sound-absorbing effect.

[0095] The present invention effectively improves the noise control capability by covering the inside of the box of the robot offline comprehensive detection device with sound-absorbing material, provides a good environment for subsequent abnormal sound detection of the sound signals collected by the sound collection equipment, and can greatly improve the detection accuracy.

[0096] If the detection range of a single sound collection device 4 and video collection device 5 covers the entire detection table 2, the installation height of the sound collection device 4 and video collection device 5 will be relatively high. In order to reduce the floor height of the box 1, the entire box structure can be partitioned. After partition detection using multiple sound collection devices 4 and video collection devices 5, data fusion is used for analysis to improve space utilization at the production line site.

[0097] As an optional embodiment, the sound acquisition device 4 and the video acquisition device 5 are both multiple;

[0098] A plurality of the sound collecting devices 4 and a plurality of the video collecting devices 5 are evenly installed in the cavity structure;

[0099] Alternatively, at least one set of the sound collection device 4 and the video collection device 5 is installed corresponding to each detection area.

[0100] Specifically, the present invention can arrange multiple sound collection devices 4 and video collection devices 5 inside the box 1, and the following two arrangement schemes can be adopted:

[0101] In the first arrangement scheme, multiple sound collection devices 4 and multiple video collection devices 5 are evenly distributed inside the cavity structure, for example, evenly arranged according to the top structure of the box. This arrangement scheme ensures that no matter where the cleaning robot is operating, it can be covered by at least one sound collection device 4 and at least one video collection device 5, thereby achieving comprehensive capture of sound signals and image signals in the entire box.

[0102] In the second layout scheme, at least one set of sound collection equipment 4 and video collection equipment 5 is installed for each functional detection area (such as the step obstacle detection area 21, the black background detection area 22, the high reflective background detection area 23 and the cliff detection area 24). This layout method focuses more on the specific needs of each functional detection area and can more accurately capture the sound signals and image signals generated by the cleaning robot when it is running in these functional detection areas.

[0103] For example, one or more microphones are installed above or on the side of each functional detection area to ensure that the specific sound signals generated by the cleaning robot when it is operating in the functional detection area can be captured. In the step obstacle detection area 21, the microphone can capture the mechanical noise of the cleaning robot when climbing the steps; in the cliff detection area 24, the microphone can capture the braking sound of the robot when it approaches the edge or the sound of the motor reducing the speed, etc. One or more cameras are installed directly above or on the side of each functional detection area to ensure that the operating status of the cleaning robot in the functional detection area and the detailed changes in the cleaning robot's movement, such as path planning, obstacle avoidance strategy, etc., can be clearly captured.

[0104] Based on the content of the above embodiment, as an optional embodiment, the robot offline comprehensive detection device provided by the present invention further includes a display device 7 which is communicatively connected to the control module 3;

[0105] The display device 7 is used to display and visualize the audio features of the sound signal obtained during the offline detection of the robot, as well as the environmental perception map reconstruction data, actual operation trajectory and offline comprehensive detection results obtained during the offline detection of the robot.

[0106] Optionally, when the offline comprehensive detection result is abnormal, the display device 7 can display the abnormal cause and the corresponding troubleshooting method in detail.

[0107] The robot offline comprehensive detection device provided by the present invention provides a more intuitive and comprehensive detection result display and diagnosis function by adding a display device 7.

[0108] The display device 7 may be a high-resolution display or a touch screen, installed at a location convenient for inspection personnel to observe, such as a control console of the inspection device or a nearby working area.

[0109] The display device 7 can be connected to the communication interface (such as HDMI, USB-C, DP, etc.) of the control module 3 to receive and display the processing results and detection data from the control module 3.

[0110] Specifically, the audio features of the sound signal obtained during the offline inspection of the cleaning robot can be displayed by the display device 7. These features may include a sound spectrum, a waveform, a sound intensity distribution, etc., to help the inspector intuitively understand the sound state of the cleaning robot during operation, such as motor noise, abnormal vibration, etc.

[0111] The display device 7 can also display the environment perception map reconstructed by the cleaning robot through its own sensors (such as laser radar, camera, etc.). These environment perception maps can include the location of obstacles, the texture and height information of the ground, lighting conditions, etc., which helps the inspection personnel understand the inspection environment where the cleaning robot is located in real time.

[0112] The actual running track of the cleaning robot during the detection process can also be displayed in real time or in playback mode through the display device 7. The control module 3 will compare the actual running track with the actually planned path and feedback the relevant comparison results to the display device 7, so that the detection personnel can evaluate the path planning ability and tracking accuracy of the cleaning robot.

[0113] The final result of the offline comprehensive inspection of the robot can also be displayed through the display device 7, including the scores of various inspection indicators, whether the inspection is passed, etc., which helps the inspection personnel to quickly understand the overall performance of the cleaning robot.

[0114] Furthermore, when the offline comprehensive test result is displayed as unqualified, the display device 7 can automatically display the specific reasons for the unqualified, including the failure mode causing the abnormal sound and the unqualified items corresponding to the unqualified capacity test. Accordingly, the corresponding troubleshooting methods can also be synchronously displayed on the display device 7. The information related to these troubleshooting methods may include error codes, fault locations, recommended repair steps, etc., to help testers quickly locate problems and take corresponding measures.

[0115] The robot offline comprehensive detection device provided by the present invention has a display device, which not only provides a richer display method for the detection results, but also significantly improves the transparency and operability of the detection process. The detection personnel can intuitively understand the robot's sound characteristics, environmental perception ability, running trajectory and comprehensive detection results, so as to more accurately evaluate the performance of the cleaning robot and quickly take troubleshooting measures when an abnormality occurs, which not only improves the detection efficiency, but also reduces the downtime and maintenance costs caused by failures.

[0116] As an optional embodiment, the present invention further provides lighting for the interior of the cavity structure by installing a lighting device 8 in the box 1. The lighting device 8 uses an LED light source, such as an LED light bar or LED panel light installed on the top, bottom or side of the box to ensure that the light can be evenly distributed inside the box 1. If the box is deep, it is also possible to consider using multi-angle or adjustable direction LED lights to reduce the shadow area.

[0117] The control module 3 can remotely control the lighting device to turn on / off, adjust the brightness, change the possible color, and illuminate the area to meet the needs of different detection scenarios.

[0118] Optionally, the robot offline comprehensive detection device provided by the present invention further includes an abnormal execution mechanism that is communicatively connected to the control module 3;

[0119] When the control module 3 determines that the offline comprehensive detection result for any robot is abnormal, the control module 3 controls the abnormal execution mechanism to transfer and isolate the any robot to a preset area.

[0120] The robot offline comprehensive detection device provided by the present invention not only has comprehensive detection functions, but also integrates an abnormal execution mechanism, which is used to automatically transfer the robot to a preset isolation area when it is detected that the robot is unqualified, thereby effectively preventing unqualified products from entering the market, while improving the overall efficiency and safety of the production line.

[0121] The abnormal actuator mainly includes three components: a gripping unit, a mobile platform, and an isolation area. The gripping unit can use high-precision electric or pneumatic grippers and is designed with flexible joints, which can stably grip and carry robots of different models and weights according to the instructions of the control module 3.

[0122] The mobile platform can be an automatic guided vehicle or a rail-mounted transport vehicle with automatic navigation and obstacle avoidance functions, and can move freely on the production line to transport abnormal robots from the inspection station to a preset isolation area. The isolation area can be a closed, clearly marked area for storing robots for further analysis or maintenance to prevent them from being mixed with normal products.

[0123] Optionally, the robot offline comprehensive detection device provided by the present invention also facilitates loading and unloading of the robot to be detected by arranging an automatic door mechanism on the box body 1.

[0124] Optionally, the robot offline comprehensive detection device provided by the present invention is further configured with an identity recognition unit for performing identity code scanning or ID recognition on the robot to be tested, so as to realize centralized management.

[0125] Figure 3This is one of the flow charts of the robot offline comprehensive detection method provided by the present invention, such as Figure 3 As shown, it mainly includes but is not limited to the following steps:

[0126] Step 31, receiving the sound signal collected by the sound collection device, the image signal collected by the video collection device, and the operation information sent by the robot;

[0127] Step 32, performing time synchronization analysis on the sound signal, the image signal and the operation information, and outputting the offline comprehensive detection result of the robot.

[0128] Specifically, the offline comprehensive detection results that can be achieved include abnormal sound detection results and capability detection results.

[0129] The abnormal sound detection result refers to a detection result that uses spectrum analysis, sound feature extraction and other technical means to monitor and analyze the sound generated by the cleaning robot during operation, so as to determine whether the cleaning robot has abnormal noise. For example, the presence of abnormal sound can be determined based on whether certain audio features (such as frequency, amplitude, etc.) in the sound signal exceed the preset abnormal sound threshold.

[0130] The capability test result refers to a test result that evaluates the function and performance of the robot in a specific task or scenario, mainly including one or more of the environmental map reconstruction test results, actual operation trajectory test results, cleaning route planning function test results and edge obstacle avoidance function test results.

[0131] The environmental map reconstruction test result refers to a test result that evaluates the cleaning robot's map construction capability (such as map accuracy, completeness, and accuracy of feature point matching) by comparing the environmental map independently explored and constructed by the cleaning robot with a preset standard map.

[0132] The actual operation trajectory detection result refers to the detection result of the robot's position information being recorded in real time by the sensors carried by the cleaning robot itself (such as lidar, etc.), and its actual operation trajectory is determined by this position information and compared with the preset path, thereby evaluating the cleaning robot's path tracking capability.

[0133] The cleaning route planning function test result refers to a test result that simulates areas with different levels of dirtiness to test whether the robot can automatically adjust its cleaning strategy according to the degree of dirtiness, and evaluate its cleaning efficiency, coverage and response capabilities.

[0134] The edge obstacle avoidance function test result refers to a test result that sets obstacles to test whether the cleaning robot can successfully avoid obstacles when moving towards the obstacle, and evaluates the rationality of its obstacle avoidance strategy, reaction speed and obstacle avoidance success rate.

[0135] In the field of smart home and cleaning automation, the performance stability and quality of cleaning robots are directly related to user experience. The present invention provides a comprehensive offline detection method for robots, which combines sound collection, video monitoring and the robot's own operation information summary, and comprehensively evaluates the working status and performance of the cleaning robot through time synchronization analysis. The following describes in detail how the present invention achieves comprehensive offline detection of robots.

[0136] In step 31, the control module will collect multiple information in real time, mainly including sound signals, image signals and the robot's own operating information.

[0137] When the cleaning robot is working, its motor, fan, roller and other parts will generate specific sound signals. Sound collection equipment such as microphones are used to collect these sound signals in real time, especially the abnormal noise contained in them, such as abnormal noise of the motor and jamming sound of the roller. The presence of these abnormal noises often indicates potential mechanical failure or wear.

[0138] In addition, high-definition cameras or visual sensors can be used to monitor the robot's movement trajectory, cleaning effect, obstacle avoidance ability, and interaction with the surrounding environment. For example, it can be observed whether the robot can effectively avoid obstacles, whether the cleaning path is reasonably planned, and whether there are any missed areas.

[0139] The sensors carried by the cleaning robot, such as lidar, IMU navigation unit, etc., and its own processor can send operating parameters to the control module in real time, such as operating voltage, operating current, control signal, speed, position information, cleaning mode, sensor status and other operating information. These operating information directly reflects the current status, energy consumption and health status of the cleaning robot.

[0140] Since the sound signal, image signal and operation information have different sources, in step 32, the control module first performs time synchronization processing on these information to ensure that all information is aligned on the same time axis for subsequent comprehensive analysis.

[0141] Among them, the analysis of the sound signal, image signal and operation information after time synchronization can be carried out by independently analyzing a certain type of information or fusing and analyzing multiple types of information according to different actual detection items.

[0142] For example, a sound recognition algorithm can be used to perform spectrum analysis, feature extraction, and pattern matching on the time-synchronized sound signal to identify abnormal noise and determine its possible fault type.

[0143] For another example, the time-synchronized image signal and the operation information can be fused and analyzed, the position information of the cleaning robot at the current moment can be determined by the operation information, and then the detection area of ​​the cleaning robot on the detection table can be determined in combination with the image signal. Through image processing and machine learning algorithms, target detection, tracking and behavior analysis can be performed on the image signal to evaluate whether the cleaning actions and operation trajectories performed by the cleaning robot in the detection area are qualified.

[0144] The method for comprehensive offline robot detection provided by the present invention operates the robot on a specially designed detection platform, on which a plurality of relevant detection areas are pre-designed according to the needs of simulating main functional scenarios, and then adopts a combination of sound and vision to realize all-round comprehensive offline robot detection. Compared with the traditional manual detection detection platform, the function and performance of the robot can be analyzed more comprehensively and accurately.

[0145] The following will explain how to perform the above-mentioned tests in conjunction with specific embodiments to obtain the environmental map reconstruction test results, the actual operation trajectory test results, the cleaning route planning function test results and the edge obstacle avoidance function test results.

[0146] In the manufacturing process of cleaning robots, comprehensive inspection before going offline is a key step to ensure their performance and quality. In addition to traditional visual inspection (through image signals) and capability testing (such as action execution effects, capability testing results obtained through video and algorithm analysis), the present invention also introduces sound signal analysis, especially in the process of abnormal sound detection, the running speed information of the cleaning robot is combined to facilitate a more comprehensive and accurate evaluation of the running status of the cleaning robot.

[0147] Figure 4 This is the second flow chart of the robot offline comprehensive detection method provided by the present invention, such as Figure 4 As shown, after obtaining the sound signal of the cleaning robot during operation obtained by the sound acquisition device, the image signal of the cleaning robot during operation obtained by the video acquisition device, and the operation information uploaded by the sensor carried by the cleaning robot itself, the timestamp technology can be used to accurately align the image signal, the sound signal and the operation information (including at least the speed data). This can be achieved by adding a timestamp to each data packet during the data acquisition stage to ensure that the three can be accurately matched in subsequent analysis.

[0148] Furthermore, audio processing software or audio feature extraction algorithms are used to extract key audio features from sound signals, which may include one or more of time domain features, frequency domain features, and time-frequency domain features, such as frequency distribution, energy spectrum, harmonic structure, etc.

[0149] Combining the extracted audio features with the synchronized speed information (such as average speed, maximum speed), machine learning models (such as support vector machine (SVM), convolutional neural network (CNN), etc.) are used for analysis.

[0150] The present invention uses speed information as part of the machine learning model input to help the machine learning model understand the difference between normal sound and abnormal sound at different speeds, thereby more accurately determining whether there is abnormal sound. For example, if the machine learning model identifies that at a certain speed, the sound characteristics emitted by the cleaning robot significantly deviate from the normal range, it is determined to be an abnormal sound.

[0151] Based on the content of the above embodiment, as an optional embodiment, the operation information further includes position information; and determining the abnormal sound detection result according to the audio feature and the speed information may also include:

[0152] Determine the detection area where the cleaning robot is located at each time point according to the position information, and obtain the interval speed information of the cleaning robot running in each of the detection areas;

[0153] Dividing the audio feature into a plurality of audio feature segments according to partial image signals collected in different detection areas;

[0154] Using each audio feature segment and the interval speed information, a foreign sound detection vector is generated;

[0155] Inputting the abnormal sound detection vector into an abnormal sound detection model to obtain an interval abnormal sound detection result output by the abnormal sound detection model;

[0156] By traversing all the audio feature segments, the abnormal sound detection result is determined according to the acquired abnormal sound detection results of the multiple intervals.

[0157] The robot offline comprehensive detection method provided by the present invention further refines and expands the abnormal sound detection process, especially introduces position information, combines the running trajectory of the cleaning robot with the audio characteristics, and realizes more accurate and comprehensive abnormal sound detection, which mainly includes the following steps:

[0158] Firstly, the operation process of the cleaning robot is divided into a plurality of intervals according to the operation track of the cleaning robot and the preset detection area.

[0159] Extract audio features such as frequency, amplitude, timbre, etc. from the synchronized sound signal. Determine the detection area where the cleaning robot is located at each time point based on the position information, and obtain the interval speed information of the robot in each detection area. Then, according to the partial image signals collected in different detection areas, divide the audio features into multiple audio feature segments, and each audio feature segment corresponds to a detection area.

[0160] Furthermore, each audio feature segment and the corresponding interval speed information are used to generate a foreign sound detection vector, which includes a combination of audio features and speed information and is used for subsequent foreign sound detection.

[0161] The generated noise detection vector is input into a pre-trained noise detection model, which can output the corresponding interval noise detection result according to the input noise detection vector.

[0162] All audio feature segments are traversed to obtain multiple interval abnormal sound detection results. Then, based on these interval abnormal sound detection results, the final abnormal sound detection result is comprehensively determined.

[0163] Finally, the abnormal sound detection results are combined with the capability detection results (such as the detection results of environmental map reconstruction, actual operation trajectory, clean route planning and edge obstacle avoidance functions) to form a comprehensive offline comprehensive detection result.

[0164] If the abnormal sound detection result is unqualified, the abnormal sound detection model will output the fault mode that the model initially judged to cause the abnormal sound. This is because when training the abnormal sound detection model, in order to improve the robustness of the model, positive and negative sample labels are used for synchronous training, and different negative samples are pre-labeled with corresponding fault mode labels.

[0165] The robot offline comprehensive detection method provided by the present invention can detect the abnormal sound of the robot more accurately and comprehensively. After introducing the position information and speed information, a more accurate abnormal sound detection vector can be generated according to the operating status and sound characteristics of the robot in different detection areas, and the abnormal sound detection can be performed using a pre-trained abnormal sound detection model. This not only improves the accuracy of abnormal sound detection, but also provides strong support for subsequent fault diagnosis. At the same time, combined with the ability test results, a comprehensive offline comprehensive test result can be formed to ensure that the robot leaving the factory has high reliability and stability.

[0166] As an optional embodiment, the following describes in detail how to reconstruct data based on the environmental perception map sent by the cleaning robot and compare it with the design map of the detection station to determine the environmental perception detection result of the robot, which specifically includes the following steps:

[0167] During the offline inspection of the cleaning robot, the robot is started and made to run on the inspection table according to the preset motion trajectory. The cleaning robot perceives the surrounding environment in real time through its built-in sensors (such as cameras, lidar, etc.) and generates environmental perception map reconstruction data. The cleaning robot transmits the collected environmental perception map reconstruction data to the control module via wired or wireless methods (such as WIFI / Bluetooth, etc.). The control module is responsible for receiving and storing these environmental perception map reconstruction data for subsequent comparison and analysis.

[0168] Before the cleaning robot goes offline for testing, a design map will be made based on the actual layout of the test bench and the distribution of obstacles. The design map describes in detail the characteristics and positional relationships of each test area in the test bench.

[0169] The received environmental perception map reconstruction data is compared with the design map. The comparison process can involve multiple aspects, such as the map size, the location and shape of each obstacle, path planning, etc. Through the comparison, the accuracy and completeness of the cleaning robot in environmental perception can be evaluated.

[0170] The specific implementation steps for comparing the environmental perception map reconstruction data with the design map of the test station specifically include:

[0171] First, the environmental perception map reconstruction data can be reconstructed into a perception map, and it can be determined that the perception map has the same scale and coordinate system as the design map. At the same time, the two maps can be gridded, with each grid representing a cell in the environment and its value representing the characteristics of the location (such as occupancy probability).

[0172] Furthermore, the similarity between the perception map and the design map can be compared based on feature comparison, including extracting significant feature points (such as corner points, edges, etc.) on the two maps.

[0173] Then, the significant feature points at the same location on the two maps are constructed into a significant feature point pair; the matching degree of each significant feature point pair is evaluated, and indicators such as matching degree and overlap degree are calculated as the comparison result.

[0174] The comparison result can accurately represent whether the environmental perception of the cleaning robot is accurate, and can distinguish possible perception errors or omissions. Therefore, it can be judged whether the environmental perception ability of the cleaning robot meets the preset standards based on the comparison result.

[0175] If the reconstructed data of the environment perception map is highly consistent with the design map and can accurately reflect the environmental characteristics of the test bench, the environment perception test result is judged to be qualified. Conversely, if there are significant differences or errors in the comparison results, the environment perception test result is judged to be unqualified.

[0176] The accuracy of the environmental map reconstruction detection results depends on multiple factors, including the robot's sensor accuracy, data processing algorithm, detection platform layout and obstacle distribution. The robot offline comprehensive detection method provided by the present invention can effectively evaluate the robot's ability in environmental perception. By comparing the results, these factors that affect environmental map reconstruction can be continuously optimized, thereby improving the robot's environmental perception ability and making it better adapt to various complex environments.

[0177] In automated production lines or testing environments, robots are often used to perform various tasks. To ensure that the robot can complete tasks efficiently and accurately, its trajectory needs to be precisely controlled. However, due to factors such as mechanical errors and sensor noise, the actual trajectory of the robot may deviate from the designed trajectory. Therefore, it is necessary to detect the actual trajectory of the robot in order to detect and correct the deviation in time.

[0178] As an optional embodiment, the robot offline comprehensive detection method provided by the present invention can realize the trajectory detection of the cleaning robot and obtain the actual operation trajectory detection result, which can be specifically implemented based on the following steps:

[0179] Determine the position information of the cleaning robot at each time point according to the operation information;

[0180] Projecting the position information to the global coordinates of the detection platform, and obtaining the position point of the cleaning robot on the detection platform at each time point;

[0181] Connect all the position points in chronological order to obtain the actual running trajectory;

[0182] The actual running track is matched with the designed running track of the test bench to determine the actual running track test result.

[0183] First, the location information of the cleaning robot during operation is collected in real time through the cleaning robot's built-in sensors (such as encoders, gyroscopes, etc.) or external positioning devices (such as laser rangefinders, visual positioning systems, etc.). This information may include the cleaning robot's coordinate position, speed, acceleration, etc.

[0184] For example, we can assume that the cleaning robot has a built-in high-precision encoder that can provide real-time feedback of coded signals, which are used to characterize changes in its joint angles. By analyzing these coded signals and combining them with the kinematic model of the cleaning robot, the three-dimensional coordinate position of the cleaning robot at each time point can be calculated.

[0185] Since the cleaning robot may run in different positions and directions, it is necessary to uniformly project the collected position information into a global coordinate system. In this way, no matter where the cleaning robot runs, its position information can be accurately represented in a unified coordinate system.

[0186] Optionally, this embodiment uses the global coordinate system of the detection platform to unify the position information, that is, the center point of the detection platform is used as the origin, the horizontal direction is the X axis, and the vertical direction is the Y axis. By calculating the relative position and posture relationship between the cleaning robot and the detection platform, the position information of the cleaning robot can be projected into this global coordinate system.

[0187] Furthermore, by connecting all the positions of the cleaning robot in the global coordinate system at each time point in chronological order, the actual running trajectory of the cleaning robot can be obtained. This actual running trajectory is a curve composed of multiple discrete points, which can intuitively reflect the path changes of the cleaning robot during operation.

[0188] Optionally, an interpolation algorithm is used to smooth the discrete points to obtain a more continuous and smooth actual running trajectory.

[0189] Finally, the actual running trajectory is matched with the designed running trajectory of the test bench. The designed trajectory is usually pre-set by engineers based on task requirements and cleaning robot performance, and it represents the ideal path that the cleaning robot should follow.

[0190] Specifically, a distance-based matching algorithm can be used to compare the deviation between the actual running trajectory and the designed trajectory. For example, for each point on the actual running trajectory, the distance between it and the nearest point on the designed trajectory is calculated, and the average of these distances is used as the matching error. A threshold is set to determine whether the matching error is within an acceptable range, thereby determining the actual running trajectory detection result.

[0191] The present invention builds a detection platform including a step obstacle detection area, a black background detection area, a high-reflective background detection area, and a cliff detection area, each of which simulates a specific cleaning challenge. For example, the step obstacle tests the robot's obstacle avoidance ability, the black background and high-reflective background tests the robot's visual recognition ability, and the cliff detection area tests the robot's edge detection and safety protection mechanism. Accordingly, according to the design specifications of the cleaning robot and user needs, the accuracy requirements of the cleaning action and running trajectory and the normal range of the running state are set.

[0192] At the same time, the present invention can capture the image signal, sound signal and operation information of the cleaning robot in real time by installing a sound collection device and a video collection device on the detection table and combining the sensors carried by the cleaning robot itself.

[0193] In this embodiment, the real-time operation video of the cleaning robot in each detection area is recorded by a video acquisition device, with special attention paid to the robot's reaction and action when facing steps, black backgrounds, highly reflective backgrounds and cliff edges. At the same time, by analyzing the operation information and recording the operation trajectory data of the cleaning robot in each detection area, key positions such as the starting point, end point, passing point and turning point can be marked. And the speed, acceleration, power consumption, motor temperature and other operation data of the cleaning robot are collected through the operation information.

[0194] Machine vision algorithms can be used to analyze image signals to evaluate whether the cleaning robot's cleaning actions in each detection area are accurate and effective, especially its adaptability in complex environments. Based on the running trajectory data, the cleaning robot's path planning capabilities in each detection area can be evaluated, including the accuracy, continuity and coverage of the path. Based on the analysis of the running information, the cleaning robot's stability when facing challenges in different detection areas can also be checked, such as whether the speed is stable, whether the power consumption is reasonable, and whether there is abnormal temperature rise in the motor.

[0195] If the robot's cleaning action, running track or running state in any detection area does not meet the preset conditions, it will be judged as unqualified. For example, if the robot cannot accurately identify and avoid steps in the step obstacle detection area, or if the robot does not clean thoroughly due to visual recognition problems in the black background detection area, or if the robot falls off the cliff instead of moving along the edge of the cliff in the cliff detection area, or if the robot is more than the preset threshold away from the edge of the cliff, it will be considered unqualified.

[0196] At the same time, it is also necessary to record the specific information of the non-conforming items, including the inspection area related to the non-conforming inspection, the non-conforming type, the problem description and possible improvement suggestions.

[0197] The robot offline comprehensive detection method provided by the present invention is targeted at a specially designed detection platform (including a step obstacle detection area, a black background detection area, a high-reflective background detection area, and a cliff detection area). By systematically collecting and analyzing the image signals, running trajectories, and running information of the robot in these areas, it can comprehensively evaluate the performance of the robot in different scenarios to ensure that it can perform well when facing actual household or commercial cleaning tasks. At the same time, combined with preset conditions (such as the accuracy of cleaning actions, the accuracy of running trajectories, and the stability of the running state), it can accurately determine whether the robot meets the design requirements, providing strong support for product quality control. And by recording the specific information of unqualified items in detail (including detection areas, unqualified types, problem descriptions, and possible improvement suggestions), it is possible to quickly locate the shortcomings in the robot's performance, which can provide a clear direction for improvement for the R&D team.

[0198] Based on the content of the above embodiment, as an optional embodiment, the edge obstacle avoidance function detection result mentioned in the above embodiment is determined by adopting the following steps:

[0199] Obtaining the running state and running trajectory of the robot in the step obstacle detection area and the cliff detection area;

[0200] The edge obstacle avoidance function detection result of the robot is determined according to the running state and the running trajectory.

[0201] The present invention provides a method for testing the edge obstacle avoidance function of a cleaning robot using a specially designed test platform, that is, by setting a step obstacle detection area and a cliff detection area on the test platform to simulate different edge environments that the robot may encounter. This method can effectively evaluate whether the robot can accurately identify and effectively avoid potential dangers when facing step obstacles and cliff-like edges.

[0202] Specifically, the step obstacle detection area can simulate steps of different heights and shapes so that the cleaning robot can identify and cross or avoid them; the cliff detection area simulates an environment similar to the edge of a cliff with a significant height difference so that the cleaning robot can identify and respond to obstacle avoidance.

[0203] Accordingly, the preset conditions for edge obstacle avoidance can be set according to the design specifications of the cleaning robot and user needs. In particular, for the step obstacle detection area and the cliff detection area, the specific obstacle avoidance actions that the cleaning robot should take when it recognizes an obstacle are set, such as crossing the steps, avoiding obstacles, emergency stopping, retreating or turning, etc.

[0204] First, the real-time operation video of the cleaning robot in the step obstacle detection area and the cliff detection area is intercepted from the image signal. The key frames in the video can be extracted through video analysis technology to observe whether its obstacle avoidance action is accurate and timely. At the same time, by extracting the distance between the cleaning robot and the step obstacle and the edge of the cliff when it is running in these two areas, as well as the dynamic information such as the speed and acceleration of the cleaning robot during the obstacle avoidance process, the working status data such as the motor status, power consumption, and safety system response are also extracted.

[0205] Furthermore, based on the intercepted image signals related to these detection areas, it is analyzed whether the cleaning robot has taken the preset obstacle avoidance action when identifying the step obstacle and the cliff edge. For example, in the step obstacle detection area, whether the cleaning robot successfully crosses or avoids the steps; in the cliff detection area, whether the cleaning robot has taken obstacle avoidance actions such as emergency stop, retreat or turn in time. Calculate the key indicators such as the average obstacle avoidance distance, maximum obstacle avoidance distance and response time of the obstacle avoidance action of the cleaning robot during the obstacle avoidance process, compare them with the preset conditions, and evaluate its obstacle avoidance accuracy and reaction speed. In addition, it is also possible to analyze whether the cleaning robot has any abnormal conditions during the obstacle avoidance process, such as motor overheating, sharp drop in power or non-response of the safety system, so as to evaluate its overall operating efficiency and stability.

[0206] If the obstacle avoidance action, sensor data or operating status of the cleaning robot in the step obstacle detection area or cliff detection area does not meet the preset conditions, it will be judged as unqualified. For example, if the robot fails to successfully cross or avoid a step obstacle when it identifies it, or fails to take obstacle avoidance action in time when it identifies the edge of a cliff, resulting in crossing the edge or collision, it will be considered unqualified.

[0207] Record the specific information of the unqualified items, including the test scenario, unqualified type (obstacle avoidance action, sensor data, operating status), problem description and possible improvement suggestions.

[0208] The robot offline comprehensive detection method provided by the present invention uses the step obstacle detection area and cliff detection area set on the detection platform to realize the edge obstacle avoidance function detection, which can systematically evaluate the robot's obstacle avoidance ability when facing different edge environments, not only improving the detection efficiency and accuracy, but also providing detailed detection data and feedback suggestions for the R&D team, which is helpful to promote the continuous improvement and optimization of product technology. At the same time, the method also ensures that the cleaning robot can perform cleaning tasks safely and efficiently in actual applications, especially when facing potential dangers, it can respond quickly to avoid accidents.

[0209] Figure 5 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 5As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communication interface 520 and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the robot offline comprehensive detection method, which is applied to the control module provided in any of the above embodiments. The method includes: receiving the sound signal collected by the sound collection device, the image signal collected by the video collection device and the operation information sent by the robot; performing time synchronization analysis on the sound signal, the image signal and the operation information, and outputting the offline comprehensive detection result of the robot. The offline comprehensive detection result includes an abnormal sound detection result and a capability detection result; the capability detection result includes one or more of an environmental map reconstruction detection result, an actual operation trajectory detection result, a cleaning route planning function detection result and an edge obstacle avoidance function detection result.

[0210] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0211] On the other hand, the present invention also provides a computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by the computer, the computer can execute the robot offline comprehensive detection method provided by the above embodiments, applied to the control module provided by any of the above embodiments, the method includes: receiving the sound signal collected by the sound acquisition device, the image signal collected by the video acquisition device and the operation information sent by the robot; performing time synchronization analysis on the sound signal, the image signal and the operation information, and outputting the offline comprehensive detection result of the robot. The offline comprehensive detection result includes an abnormal sound detection result and a capability detection result; the capability detection result includes one or more of an environmental map reconstruction detection result, an actual operation trajectory detection result, a cleaning route planning function detection result and an edge obstacle avoidance function detection result.

[0212] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the processor executes the robot offline comprehensive detection method provided in the above embodiments, and is applied to the control module provided in any of the above embodiments, the method comprising: receiving the sound signal collected by the sound acquisition device, the image signal collected by the video acquisition device, and the operation information sent by the robot; performing time synchronization analysis on the sound signal, the image signal, and the operation information, and outputting the offline comprehensive detection result of the robot. The offline comprehensive detection result includes an abnormal sound detection result and a capability detection result; the capability detection result includes one or more of an environmental map reconstruction detection result, an actual operation trajectory detection result, a cleaning route planning function detection result, and an edge obstacle avoidance function detection result.

[0213] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0214] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0215] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A robot offline comprehensive detection device, characterized in that: include: A detection module, including a sound acquisition device for acquiring a sound signal of the robot and a video acquisition device for acquiring an image signal of the robot; A detection platform, located below the detection module, for the robot to run and perform detection, the detection platform is provided with a plurality of functional detection areas for matching different detection scenes of the robot, the functional detection areas including at least one of a step obstacle detection area, a black background detection area, a high-reflective background detection area, and a cliff detection area; A control module, wherein the sound acquisition device, the video acquisition device, and the robot are respectively connected to the control module for communication; The control module receives the sound signal collected by the sound collection device, the image signal collected by the video collection device, and the operation information sent by the robot; performs time synchronization analysis on the sound signal, the image signal, and the operation information, and outputs the offline comprehensive detection result of the robot; The offline comprehensive test results include abnormal sound detection results and capability test results; The capability test result includes one or more of an environment map reconstruction test result, an actual operation trajectory test result, a cleaning route planning function test result, and an edge obstacle avoidance function test result; The operation information includes speed information, and the control module performs time synchronization analysis on the sound signal, the image signal and the operation information, and outputs the offline comprehensive detection result of the robot, which specifically includes: Performing time synchronization on the sound signal and the operation information; Extracting audio features of the sound signal; Determine the abnormal sound detection result according to the audio feature and the speed information; The offline comprehensive detection result is determined by combining the abnormal sound detection result and the capability detection result.

2. The robot offline comprehensive detection device according to claim 1 is characterized in that: A preset distance is maintained between the table top of the detection table and the detection module; The step obstacle detection area is arranged with step obstacles, the black background detection area is paved with a black material coating, the high-reflective background detection area is paved with a high-reflective material coating, and the cliff detection area includes at least two adjacent platforms with a height difference greater than a preset height difference.

3. The robot offline comprehensive detection device according to claim 1 is characterized in that: Also included is a display device connected to the control module; The display device is used to visually display the audio features of the sound signal obtained during the offline detection of the robot, as well as the environmental perception map reconstruction data, actual operation trajectory and offline comprehensive detection results obtained during the offline detection of the robot.

4. The robot offline comprehensive detection device according to claim 1 is characterized in that: Also included is an abnormality execution mechanism communicatively connected to the control module; When the control module determines that the offline comprehensive detection result for any robot is abnormal, the control module controls the abnormal execution mechanism to transfer and isolate the any robot to a preset area.

5. A comprehensive detection method for offline robots, characterized in that: The robot offline comprehensive detection device is applied to any one of claims 1-4.

6. The robot offline comprehensive detection method according to claim 5 is characterized in that: The operation information also includes position information; and determining the abnormal sound detection result according to the audio feature and the speed information includes: Determine the detection area where the robot is located at each time point according to the position information, and obtain interval speed information of the robot running in each detection area; Dividing the audio feature into a plurality of audio feature segments according to partial image signals collected in different detection areas; Using each audio feature segment and the interval speed information, a foreign sound detection vector is generated; Inputting the abnormal sound detection vector into an abnormal sound detection model to obtain an interval abnormal sound detection result output by the abnormal sound detection model; By traversing all the audio feature segments, the abnormal sound detection result is determined according to the acquired abnormal sound detection results of the multiple intervals.

7. The robot offline comprehensive detection method according to claim 5 is characterized in that: The environment map reconstruction detection result is determined based on the following steps: Receiving environment perception map reconstruction data sent by the robot during offline detection; The environmental perception map reconstruction data is compared with the design map of the detection platform to determine the environmental perception detection result of the robot according to the comparison result.

8. The robot offline comprehensive detection method according to claim 5 is characterized in that: The actual running trajectory detection result is determined based on the following steps: Determine the position information of the robot at each time point according to the operation information; Projecting the position information to the global coordinates of the detection platform, and obtaining the position point of the robot on the detection platform at each time point; Connect all the position points in chronological order to obtain the actual running trajectory; The actual running track is matched with the designed running track of the detection platform to determine the actual running track detection result.

9. The robot offline comprehensive detection method according to claim 5 is characterized in that: The cleaning route planning function detection result is determined based on the following steps: Acquire partial image signals, partial running trajectories, and partial running information of the robot running in each detection area on the detection table; Based on the partial image signal, partial operation trajectory and partial operation information, it is determined whether the cleaning action, operation trajectory and operation status of the robot in each detection area meet the preset conditions, and if the preset conditions are not met, an unqualified item is determined.

10. The robot offline comprehensive detection method according to claim 9, characterized in that: The edge obstacle avoidance function detection result is determined by the following steps: Obtaining the running state and running trajectory of the robot in the step obstacle detection area and the cliff detection area; The edge obstacle avoidance function detection result of the robot is determined according to the running state and the running trajectory.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the robot offline comprehensive detection method as described in any one of claims 5 to 10 is implemented.

12. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the robot offline comprehensive detection method as described in any one of claims 5 to 10 is implemented.

13. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the robot offline comprehensive detection method as described in any one of claims 5 to 10 is implemented.

Images