Low obstacle sensing methods, devices and computer equipment

Abstract

This application relates to a method, apparatus, and computer device for detecting low-lying obstacles. The method includes: installing a single-line lidar directly behind a robot and an ultrasonic radar directly below the single-line lidar; extracting single-line lidar data and ultrasonic radar data respectively to obtain a first obstacle target perceived by the single-line lidar and a second obstacle target perceived by the ultrasonic radar; matching the first and second obstacle targets, identifying obstacles that fail to match as low-lying obstacles, continuously tracking the low-lying obstacles and outputting their positions; and outputting the estimated position of the low-lying obstacle when it reaches the common blind spot of the single-line lidar and ultrasonic radar. This method achieves more accurate and convenient low-lying obstacle detection with fewer sensors.

Description

Technical Field

[0001] This application relates to the field of robot obstacle detection technology, and in particular to a method, apparatus and computer device for sensing low-lying obstacles. Background Technology

[0002] Sensor measurement technology often acts as the "eyes" of modern robot navigation systems. Commonly used sensors include LiDAR, millimeter-wave radar, ultrasonic radar, and stereo vision. LiDAR is often the primary sensor, offering advantages such as long detection range and stable measurement. Because laser wavelengths can reach the micrometer level, it can detect extremely small targets. However, due to the characteristics of laser measurement, there is usually a blind spot in the short-range field of view of LiDAR, requiring the use of other sensing solutions to compensate for this. Ultrasonic propagation has advantages such as slow energy consumption, long propagation distance, strong penetration, and low cost, effectively supplementing the scanning blind spot of LiDAR and making it suitable for large-scale deployment.

[0003] When a robot is moving backward, it usually moves backward along a pre-planned route. It is necessary to pay attention to obstacles that may pose safety hazards along the backward route, especially low obstacles.

[0004] The invention disclosed in CN113111905A, entitled "An Obstacle Detection Method Integrating Multi-line LiDAR and Ultrasonic Data," is a method for obstacle detection that integrates multi-line LiDAR and ultrasonic data. This invention extracts obstacles from both LiDAR point cloud data and ultrasonic data separately, and then performs target-level fusion of the obstacles. The method for extracting obstacles from ultrasonic data involves constructing linear ultrasonic obstacles based on intrinsic and extrinsic ultrasonic parameters, followed by a scattering operation to obtain point ultrasonic obstacles. The method for extracting obstacles from LiDAR data involves ground fitting and removing ground point clouds to obtain point cloud obstacles. The fusion method involves detecting point cloud obstacles near point ultrasonic obstacles; if present, they are described as point cloud obstacles; otherwise, ultrasonic obstacles are used. The obstacle distance is obtained by assigning different weights to the most recent time-series measurement results through a weighted average. It is easy to see that the obstacle distance calculated by ultrasound in the disclosed solution of this invention uses time-series data for weighting, so it can have a certain smoothing effect. However, this calculation method is only suitable for measuring static obstacles, and is not applicable to the case of ultrasonic measurement by a robot in motion.

[0005] The invention disclosed in CN108037515A, entitled "A System and Method for Fusion Information of LiDAR and Ultrasonic Radar," presents a scheme for fusing forward obstacles using ultrasonic and LiDAR systems installed at the front of a vehicle. This invention sets the multi-line radar installation position at a forward tilt of 10-15 degrees, comprising 4 or 6 ultrasonic radars, evenly installed at the bumper position. The disclosed technical solution involves: preprocessing the point cloud data of the multi-line LiDAR within a range of 3m to 80m and converting it to the vehicle's coordinate system; preprocessing the ultrasonic data within 5m and similarly converting it to the vehicle's coordinate system; rasterizing the forward area with a resolution of 20cm × 20cm; and finally, weighting the distance of the LiDAR point cloud and the target distance of the ultrasonic radar. When the weighted distance result exceeds a set value, the target grid is marked as occupied; otherwise, it is marked as unoccupied. From the perspective of the solution steps, this invention directly marks the grid based on the obstacle distance measured by laser and ultrasonic sensors. However, it does not explain how to handle situations where neither ultrasonic nor laser radar provides a measurement result (e.g., the obstacle is transparent glass). Since laser measurement results are more reliable, and the solution design requires both laser and ultrasonic radar measurement results, this is a two-way verification solution. For transparent obstacles, where only ultrasonic results are available but no radar results, this invention cannot handle this scenario.

[0006] The invention disclosed in CN109814112A, entitled "A Method and System for Fusion of Ultrasonic Radar and LiDAR Information," aims to reduce misjudgments by ultrasonic radar. The technical solution involves acquiring obstacle information from both lidar and ultrasonic radar separately, unifying the targets from both into a single reference coordinate system, and overlaying overlapping areas. If both data sources are present in a grid, the grid is considered occupied; otherwise, it is considered unoccupied. This invention also uses lidar results to verify ultrasonic results, has certain requirements regarding the material of the obstacles, and its versatility is limited. Summary of the Invention

[0007] Therefore, it is necessary to provide a method, device, and computer equipment for detecting low obstacles that requires fewer sensors and has good versatility in detecting low obstacles, in order to address the above-mentioned technical problems.

[0008] A method for sensing low-lying obstacles, the method comprising:

[0009] A single-line lidar is installed directly behind the robot, and an ultrasonic radar is installed directly below the single-line lidar; the scanning direction of the single-line lidar is horizontal; the sensing range of the ultrasonic radar covers the near-range blind zone of the single-line lidar.

[0010] The single-line lidar data and the ultrasonic radar data are extracted from the single-line lidar and the ultrasonic radar respectively to obtain the first obstacle target perceived by the single-line lidar and the second obstacle target perceived by the ultrasonic radar.

[0011] The first obstacle target and the second obstacle target are matched. The obstacle that fails to match is regarded as a low obstacle. The low obstacle is continuously tracked and its position is output. When the low obstacle reaches the common blind zone of the single-line lidar and the ultrasonic radar, the estimated position of the low obstacle is output.

[0012] In one embodiment, the method further includes: extracting a first obstacle target from the single-line lidar data using a pre-set DBSCAN algorithm.

[0013] In one embodiment, the method further includes: validating the ultrasonic radar data and filtering out noise to obtain a second obstacle target.

[0014] In one embodiment, the method further includes: obtaining the minimum effective measurement threshold and the measured value of the ultrasonic radar; and calculating the verification result based on the minimum effective measurement threshold and the measured value.

[0015]

[0016] in, Indicates the verification result, d t This represents the measurement value at time t. and This indicates the minimum and maximum effective measurement thresholds for ultrasound.

[0017] In one embodiment, the method further includes: performing a continuity check based on the check result to obtain a noise filtering result:

[0018]

[0019] Among them, v t Let Δ be the linear velocity of the robot at time t; t-a δ represents the time difference between the previous time ta and the current time t; δ represents the standard deviation of the ultrasonic radar measurement.

[0020] In one embodiment, the method further includes updating the verification result based on the measured distance recorded at the previous moment, the robot's linear velocity, and the effective measured distance at the current moment:

[0021]

[0022] Among them, w t This indicates updating the weights, d t ' represents the prediction verification result at time t. This represents the updated verification result. The formula for calculating the updated weight is:

[0023]

[0024] In one embodiment, the method further includes: obtaining the obstacle distance attribute of each target in the first obstacle target; matching the obstacle distance attribute with the second obstacle target, and treating the unmatched obstacle as a low obstacle.

[0025] In one embodiment, the method further includes: when it is determined that a low obstacle has entered the common blind zone of the single-line lidar and the ultrasonic radar, the estimated position of the low obstacle is obtained based on the robot's position at the previous moment, the robot's linear velocity, and the sampling time difference.

[0026]

[0027] Where Δt represents the sampling time difference, This represents the estimated position of the robot at time t within the common blind zone.

[0028] A low-profile obstacle sensing device, the device comprising:

[0029] The radar installation location determination module installs a single-line lidar directly behind the robot and an ultrasonic radar directly below the single-line lidar; the scanning direction of the single-line lidar is horizontal; the sensing range of the ultrasonic radar covers the near-range blind zone of the single-line lidar.

[0030] An obstacle extraction module is used to extract the single-line lidar data of the single-line lidar and the ultrasonic radar data of the ultrasonic radar respectively, to obtain the first obstacle target perceived by the single-line lidar and the second obstacle target perceived by the ultrasonic radar.

[0031] The low obstacle analysis module is used to match the first obstacle target and the second obstacle target, and to identify the obstacle that fails to match as a low obstacle. The module continuously tracks the low obstacle and outputs its position. When the low obstacle reaches the common blind zone of the single-line lidar and the ultrasonic radar, the module outputs the estimated position of the low obstacle.

[0032] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program performing the following steps:

[0033] A single-line lidar is installed directly behind the robot, and an ultrasonic radar is installed directly below the single-line lidar; the scanning direction of the single-line lidar is horizontal; the sensing range of the ultrasonic radar covers the near-range blind zone of the single-line lidar.

[0034] The single-line lidar data and the ultrasonic radar data are extracted from the single-line lidar and the ultrasonic radar respectively to obtain the first obstacle target perceived by the single-line lidar and the second obstacle target perceived by the ultrasonic radar.

[0035] The first obstacle target and the second obstacle target are matched. The obstacle that fails to match is regarded as a low obstacle. The low obstacle is continuously tracked and its position is output. When the low obstacle reaches the common blind zone of the single-line lidar and the ultrasonic radar, the estimated position of the low obstacle is output.

[0036] The aforementioned low obstacle detection method, device, and computer equipment first place a single-line lidar directly behind the robot at a certain height, ensuring its scanning direction is horizontal. This, combined with the appropriate placement of an ultrasonic radar, allows the ultrasonic radar's sensing range to cover the near-range blind zone of the single-line lidar during close-range detection. Thus, by matching data, low obstacles can be directly identified and their locations determined. This detection method is simple in structure, convenient in calculation, and adaptable to dynamic conditions and any obstacle material. Even when the obstacle enters the blind zone of the single-line lidar and ultrasonic radar, supplementary low obstacle localization can still be performed. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating a low-profile obstacle sensing method in one embodiment;

[0038] Figure 2 This is a schematic diagram of the sensor mounting structure in one embodiment;

[0039] Figure 3 This is a schematic diagram illustrating the distribution of the sensor detection range in another embodiment;

[0040] Figure 4 This is a structural block diagram of a low-profile obstacle sensing device in one embodiment;

[0041] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0043] In one embodiment, such as Figure 1 As shown, a method for sensing low-lying obstacles is provided, including the following steps:

[0044] Step 102: Install a single-line lidar directly behind the robot and an ultrasonic radar directly below the single-line lidar.

[0045] like Figure 2 As shown, the scanning direction of the single-line lidar is horizontal, and the sensing range of the ultrasonic radar covers the near-range blind zone of the single-line lidar, but there is still a common blind zone in the field of view of both.

[0046] Step 104: Extract the single-line lidar data from the single-line lidar and the ultrasonic radar data from the ultrasonic radar to obtain the first obstacle target perceived by the single-line lidar and the second obstacle target perceived by the ultrasonic radar.

[0047] Step 106: Match the first obstacle target and the second obstacle target. Treat the obstacle that fails to match as a low obstacle. Continuously track the low obstacle and output its position. When the low obstacle reaches the common blind zone of the single-line lidar and ultrasonic radar, output the estimated position of the low obstacle.

[0048] In the aforementioned low obstacle perception method, a single-line LiDAR is first placed directly behind the robot at a certain height, ensuring its scanning direction is horizontal. This is combined with a reasonably positioned ultrasonic radar. This allows the ultrasonic radar's sensing range to cover the near-range blind zone of the single-line LiDAR during close-range detection. Thus, by matching data, low obstacles can be directly identified and their locations determined. This detection method is simple in structure, convenient in calculation, and adaptable to dynamic conditions and any obstacle material. Even if the obstacle enters the blind zone of the single-line LiDAR and ultrasonic radar, supplementary low obstacle localization can still be performed.

[0049] In one embodiment, a pre-configured DBSCAN algorithm is used to extract the first obstacle target from the single-line lidar data. The DBSCAN algorithm is a density-based spatial clustering algorithm that divides points in a dataset into different clusters, where each cluster is a subset of data points that are close to each other and have a sufficiently high density under some distance metric.

[0050] Specifically, such as Figure 3 As shown, the ROI (Region of Interest) of the LiDAR point cloud is set as the overlapping observation area where the single-line LiDAR and ultrasonic sensing ranges have the same scanning field of view. Obstacles are extracted from the point cloud located within the ROI. Specifically, the DBSCAN algorithm parameters (∈, minPoints) are set to describe the density distribution of the samples. Since the single-line LiDAR mainly focuses on the distribution of obstacles at a resolution of 10cm, ∈ is set to 10cm. Secondly, the point cloud of the same object is rasterized at 10cm resolution, and the point cloud attributes are statistically analyzed to obtain a set value of minPoints = 3, indicating that the same target can be described as long as there are 3 points in the grid. After obtaining the set value, the DBSCAN algorithm is run directly to obtain the point cloud clustering results. Point clouds not included in the clustering results are isolated points or scanning noise and can be directly deleted. Finally, all the obstacle point clouds are directly output to obtain the first obstacle target o = {o1, o2, ..., o} of the single-line LiDAR. N}

[0051] In one embodiment, the ultrasonic radar data is validated and noise is filtered out to obtain the second obstacle target.

[0052] Specifically, the steps for validity verification include: obtaining the minimum effective measurement threshold and measurement value of the ultrasonic radar; and calculating the verification result based on the minimum effective measurement threshold and measurement value.

[0053]

[0054] in, Indicates the verification result, d t This represents the measurement value at time t. and This indicates the minimum and maximum effective measurement thresholds for ultrasound.

[0055] Based on the verification results, a continuity verification is performed, and the noise filtering result is as follows:

[0056]

[0057] Among them, v t Let Δ be the linear velocity of the robot at time t; t-a δ represents the time difference between the previous time ta and the current time t; δ represents the standard deviation of the ultrasonic radar measurement.

[0058] Furthermore, based on the measured distance recorded at the previous moment, the robot's linear velocity, and the effective measured distance at the current moment, the verification result is updated as follows:

[0059]

[0060] Among them, w t represents the update weight, and d t ′ represents the predicted verification result at time t, represents the updated verification result, and the calculation formula for the update weight is:

[0061]

[0062] The design of the update weight enables the system to have a certain balancing ability, that is, real-time measurement is adopted when the system error is large, and the measurement has a certain prior repair ability when the error is small. The second obstacle target is uniformly denoted as: U = {u1, u2,..., u M}.

[0063] In one embodiment, obtain the obstacle distance attribute of each target in the first obstacle target; match according to the obstacle distance attribute and the second obstacle target, and regard the obstacles with failed matching as low obstacles.

[0064] Specifically, extract the obstacle distance of the single-line lidar. First, for the first obstacle target O = {o1, o2,..., o N}, use the sensor installation information and the robot pose to calculate the distance between the lidar obstacle and the sensor, and extract the obstacle distance attribute. Then set the distance matching success threshold Match the obstacles extracted by the single-line lidar with the distance attribute with the obstacles U = {u1, u2,..., u M} described by the distance of the ultrasonic wave, and extract the ultrasonic targets with failed distance matching as low obstacles U' = {u1', u2',..., u' m}, m < M, satisfying:

[0065]

[0066] The installation height of the single-line lidar is higher than that of the ultrasonic radar, and it is in a flush-mounted manner. In the common view range of the lidar and the ultrasonic radar, the height of the obstacle target scanned by the single-line lidar is higher than that of the ultrasonic target. Therefore, the target perceived by the lidar does not belong to the low obstacle; for the target only perceived by the ultrasonic sensor and not perceived by the single-line lidar, it is determined as the low obstacle target:

[0067]

[0068] In one embodiment, when a low obstacle is determined to have entered the common blind zone of the single-line lidar and ultrasonic radar, the estimated position of the low obstacle is obtained based on the robot's position at the previous moment, the robot's linear velocity, and the sampling time difference:

[0069]

[0070] Where Δt represents the sampling time difference, This represents the estimated position of the robot at time t within the common blind zone.

[0071] Therefore, two strategies are used to output the output of low obstacles:

[0072] 1) When a low obstacle is within the sensing range of the ultrasonic wave, the real-time tracking result of the ultrasonic wave is output as the position of the low obstacle.

[0073] 2) When low-lying obstacles are located in the common blind zone of ultrasonic and lidar, the estimation results of low-lying obstacles will be output externally.

[0074] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0075] In one embodiment, such as Figure 4 As shown, a low-profile obstacle sensing device is provided, comprising: a radar installation location determination module 402, an obstacle extraction module 404, and a low-profile obstacle analysis module 406, wherein:

[0076] The radar installation position determination module 402 is used to install a single-line lidar directly behind the robot and an ultrasonic radar directly below the single-line lidar; the scanning direction of the single-line lidar is horizontal; the sensing range of the ultrasonic radar covers the near-range blind zone of the single-line lidar.

[0077] The obstacle extraction module 404 is used to extract the single-line lidar data of the single-line lidar and the ultrasonic radar data of the ultrasonic radar respectively, to obtain the first obstacle target perceived by the single-line lidar and the second obstacle target perceived by the ultrasonic radar.

[0078] The low obstacle analysis module 406 is used to match the first obstacle target and the second obstacle target, take the obstacle that fails to match as a low obstacle, continuously track the low obstacle and output its position, and output the estimated position of the low obstacle when the low obstacle reaches the common blind zone of the single-line lidar and the ultrasonic radar.

[0079] In one embodiment, the obstacle extraction module 404 is further configured to extract a first obstacle target from the single-line lidar data using a pre-set DBSCAN algorithm.

[0080] In one embodiment, the obstacle extraction module 404 is further used to perform validity verification and noise filtering on the ultrasonic radar data to obtain a second obstacle target.

[0081] In one embodiment, the obstacle extraction module 404 is further configured to acquire the minimum effective measurement threshold and measurement value of the ultrasonic radar, and calculate the verification result based on the minimum effective measurement threshold and measurement value:

[0082]

[0083] in, Indicates the verification result, d t This represents the measurement value at time t. and This indicates the minimum and maximum effective measurement thresholds for ultrasound.

[0084] In one embodiment, the obstacle extraction module 404 is further configured to perform continuity verification based on the verification result, and obtain the noise filtering result as follows:

[0085]

[0086] Among them, v t Let Δ be the linear velocity of the robot at time t; t-a δ represents the time difference between the previous time ta and the current time t; δ represents the standard deviation of the ultrasonic radar measurement.

[0087] In one embodiment, the obstacle extraction module 404 is further configured to update the verification result based on the measured distance recorded at the previous moment, the robot's linear velocity, and the effective measured distance at the current moment:

[0088]

[0089] Among them, w t This indicates updating the weights, d t ' represents the prediction verification result at time t. This represents the updated verification result. The formula for calculating the updated weight is:

[0090]

[0091] In one embodiment, the low obstacle analysis module 406 is further configured to obtain the obstacle distance attribute of each target in the first obstacle target; and to match the obstacle distance attribute with the second obstacle target, and to identify the obstacle that fails to match as a low obstacle.

[0092] In one embodiment, the low obstacle analysis module 406 is further configured to, when it is determined that a low obstacle has entered the common blind zone of the single-line lidar and the ultrasonic radar, obtain the estimated position of the low obstacle based on the robot's position at the previous moment, the robot's linear velocity, and the sampling time difference:

[0093]

[0094] Where Δt represents the sampling time difference, This represents the estimated position of the robot at time t within the common blind zone.

[0095] Specific limitations regarding the low-altitude obstacle sensing device can be found in the limitations of the low-altitude obstacle sensing method described above, and will not be repeated here. Each module in the aforementioned low-altitude obstacle sensing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0096] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores single-line lidar data and ultrasonic radar data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a low-profile obstacle detection method.

[0097] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0098] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described above.

[0099] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0100] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0101] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0102] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for sensing low-lying obstacles, characterized in that, The method includes: A single-line lidar is installed directly behind the robot, and an ultrasonic radar is installed directly below the single-line lidar; the scanning direction of the single-line lidar is horizontal; the sensing range of the ultrasonic radar covers the near-range blind zone of the single-line lidar. The single-line lidar data and the ultrasonic radar data are extracted from the single-line lidar and the ultrasonic radar respectively to obtain the first obstacle target perceived by the single-line lidar and the second obstacle target perceived by the ultrasonic radar; wherein, the ROI region of interest of the lidar point cloud is set as an observation area that overlaps with the single-line lidar and ultrasonic radar sensing ranges, and the point cloud obstacles are extracted from the point cloud located in the ROI region. The first obstacle target and the second obstacle target are matched. Obstacles that fail to match are identified as low obstacles. The low obstacles are continuously tracked and their positions are output. When the low obstacle reaches the common blind zone of the single-line lidar and the ultrasonic radar, the estimated position of the low obstacle is output. Among them, targets that are only perceived by the ultrasonic sensor and not by the single-line lidar are identified as low obstacle targets. Matching the first obstacle target and the second obstacle target, and classifying obstacles that fail to match as low-lying obstacles, including: Obtain the first obstacle target The obstacle distance attribute for each target; the obstacle distance attribute is obtained by calculating the distance between the lidar obstacle and the sensor based on the sensor installation information and the robot posture; Set the distance matching success threshold According to the obstacle distance attribute With the second obstacle target Matching: Obstacles that fail to match are treated as low-profile obstacles. N represents the number of first obstacle targets, M represents the number of second obstacle targets, and m represents the number of low obstacles, satisfying: The method further includes: When a low obstacle is determined to have entered the common blind zone of the single-line lidar and the ultrasonic radar, the estimated position of the low obstacle is obtained based on the robot's position at the previous moment, the robot's linear velocity, and the sampling time difference: in, Indicates the sampling time difference. This represents the estimated position of the robot at time t within the common blind zone. When the low obstacle is within the sensing range of the ultrasonic wave, the real-time tracking result of the ultrasonic wave is output as the position of the low obstacle. When the low obstacle is within the common blind zone of the ultrasonic wave and the lidar, the estimated position of the low obstacle is output.

2. The method according to claim 1, characterized in that, The step of extracting single-line lidar data from the single-line lidar to obtain the first obstacle target sensed by the single-line lidar includes: The first obstacle target in the single-line lidar data is extracted using a pre-set DBSCAN algorithm.

3. The method according to claim 1, characterized in that, The step of extracting the ultrasonic radar data from the ultrasonic radar to obtain the second obstacle target sensed by the ultrasonic radar includes: After validating the ultrasonic radar data and filtering out noise, the second obstacle target is obtained.

4. The method according to claim 3, characterized in that, The steps for validating the validity include: Obtain the minimum effective measurement threshold and measurement value of the ultrasonic radar. Based on the minimum effective measurement threshold and measurement value, calculate the verification result as follows: in, Indicates the verification result. d t This represents the measurement value at time t. and This indicates the minimum and maximum effective measurement thresholds for ultrasound.

5. The method according to claim 4, characterized in that, The steps for noise filtering include: Based on the verification results, a continuity verification is performed, and the noise filtering result is as follows: in, Let be the linear velocity of the robot at time t; δ represents the time difference between the previous time ta and the current time t; δ represents the standard deviation of the ultrasonic radar measurement.

6. The method according to claim 5, characterized in that, The steps for extracting the second obstacle target also include: Based on the measured distance recorded at the previous moment, the robot's linear velocity, and the effective measured distance at the current moment, the verification result is updated as follows: in, w t This indicates an update to the weights. This represents the prediction verification result at time t. This represents the updated verification result. The formula for calculating the updated weight is: 。 7. A low-profile obstacle sensing device, characterized in that, The device includes: The radar installation location determination module is used to install a single-line lidar directly behind the robot and an ultrasonic radar directly below the single-line lidar; the scanning direction of the single-line lidar is horizontal; the sensing range of the ultrasonic radar covers the near-range blind zone of the single-line lidar. The obstacle extraction module is used to extract the single-line lidar data and the ultrasonic radar data of the single-line lidar, respectively, to obtain the first obstacle target perceived by the single-line lidar and the second obstacle target perceived by the ultrasonic radar; wherein, the ROI region of interest of the lidar point cloud is set as an overlapping observation area where the single-line lidar and ultrasonic radar have the same scanning field of view, and the point cloud obstacles are extracted from the point cloud located in the ROI region. The low obstacle analysis module is used to match the first obstacle target and the second obstacle target, identify the obstacle that fails to match as a low obstacle, continuously track the low obstacle and output its position, and output the estimated position of the low obstacle when the low obstacle reaches the common blind zone of the single-line lidar and the ultrasonic radar; wherein, for targets that are only perceived by the ultrasonic sensor and not by the single-line lidar, they are identified as low obstacle targets. The low obstacle analysis module is also used to acquire the first obstacle target. The obstacle distance attribute for each target; the obstacle distance attribute is obtained by calculating the distance between the lidar obstacle and the sensor based on the sensor installation information and the robot posture; Set the distance matching success threshold According to the obstacle distance attribute With the second obstacle target Matching: Obstacles that fail to match are treated as low-profile obstacles. N represents the number of first obstacle targets, M represents the number of second obstacle targets, and m represents the number of low obstacles, satisfying: When a low obstacle is determined to have entered the common blind zone of the single-line lidar and the ultrasonic radar, the estimated position of the low obstacle is obtained based on the robot's position at the previous moment, the robot's linear velocity, and the sampling time difference: in, Indicates the sampling time difference. This represents the estimated position of the robot at time t within the common blind zone. When the low obstacle is within the sensing range of the ultrasonic wave, the real-time tracking result of the ultrasonic wave is output as the position of the low obstacle. When the low obstacle is within the common blind zone of the ultrasonic wave and the lidar, the estimated position of the low obstacle is output.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

Images