Anti-collision control system, control method thereof, processor and aerial work platform

Abstract

Embodiments of the present application provide a kind of for aerial work platform anti-collision control method, processor, anti-collision control system, aerial work platform and machine readable storage medium.Control method includes: the obstacle information detected by radar sensor is obtained;Obstacle information in the image of monitoring area collected by visual sensor is obtained;According to image recognition work platform inside work personnel whether to work;In the case where it is identified that work personnel is in work, determine whether the non-obstacle information caused by work personnel work is contained in obstacle information;In the case where it is determined that non-obstacle information is contained in obstacle information, remove non-obstacle information;According to the real obstacle information after removing non-obstacle information in obstacle information, execute anti-collision strategy.Through the working condition of work personnel in work platform obtained by visual sensor, and the information that can be identified in the process of work radar sensor false alarm, it can avoid the false alarm situation of anti-collision control system.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of aerial work platform safety control, in particular to a collision prevention control method for an aerial work platform, a processor, a collision prevention control system, an aerial work platform and a machine readable storage medium. BACKGROUND

[0002] An aerial work platform is a product for high-altitude work, equipment installation, maintenance and other mobile high-altitude work in various industries. For example, traditional aerial work platform related products can include a scissor aerial work platform, a truck-mounted aerial work platform, a boom aerial work platform, a self-propelled aerial work platform, an aluminum alloy aerial work platform, a sleeve cylinder aerial work platform, etc.

[0003] Currently, if the operation is improper or there is a blind area in the line of sight during the operation of the aerial work platform, the working bucket and the external environment may collide. Once a collision occurs, it will cause huge economic losses, and even casualties. Adding a collision prevention control system to the aerial work platform can prevent collisions from occurring through early warning, limiting movement, etc. Most of the current aerial work platform collision prevention systems are based on pure radar solutions, which obtain distance information of obstacles and the working platform through ultrasonic radar or millimeter wave radar, determine whether the working platform has a collision risk according to the distance information, and take corresponding collision prevention measures such as alarm or forced braking. However, the existing collision prevention scheme applied to the aerial work platform has the problem that the normal operation of the operator will trigger the collision prevention measures. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a collision prevention control method for an aerial work platform, a processor, a collision prevention control system, an aerial work platform and a machine readable storage medium.

[0005] In order to achieve the above-mentioned purpose, the collision prevention control method provided by the embodiments of the present application is applied to an aerial work platform, the aerial work platform includes a working platform and a collision prevention control system, the collision prevention control system includes a visual sensor for collecting images of a monitoring area and a radar sensor for detecting obstacles, and the control method includes:

[0006] Obtaining obstacle information of obstacles detected by the radar sensor;

[0007] Obtaining images of the monitoring area collected by the visual sensor, the monitoring area containing an activity area of the operator when working on the working platform;

[0008] Identifying whether the operator inside the working platform is working according to the images;

[0009] In a case where it is identified that the worker is working, it is determined whether the non-obstacle information caused by the worker working is contained in the obstacle information.

[0010] In a case where it is determined that the non-obstacle information is contained in the obstacle information, the non-obstacle information is removed.

[0011] A collision avoidance strategy is executed according to the real obstacle information in the obstacle information after the non-obstacle information is removed.

[0012] In the embodiments of the present application, identifying whether the worker is working according to the image includes:

[0013] inputting the image into a target detection model;

[0014] the target detection model outputs the identification result, and the output result includes a detected hand of the worker or a working tool category and a position relationship of the hand or the working tool bounding box relative to the working platform;

[0015] identifying whether the worker is working according to the identification result.

[0016] In the embodiments of the present application, in a case where it is identified that the worker is working, it is determined whether the non-obstacle information caused by the worker working is contained in the obstacle information, including:

[0017] determining, according to the obstacle information, an obstacle position of the detected obstacle;

[0018] determining whether there is an obstacle position with a matching position relationship in the obstacle position;

[0019] In a case where it is determined that there is an obstacle position with a matching position relationship, it is determined that the non-obstacle information is contained in the obstacle information.

[0020] In the embodiments of the present application, the radar sensor includes a first radar sensor and a second radar sensor, a first detection area of the first radar sensor and a second detection area of the second radar sensor have an overlapping detection area, and the collision avoidance control method further includes:

[0021] respectively acquiring first object information and second object information detected by the first radar sensor and the second radar sensor for a target object moving in the overlapping detection area;

[0022] generating a first motion trajectory in a first coordinate system of the first radar sensor and a second motion trajectory in a second coordinate system of the second radar sensor according to the first object information and the second object information, respectively;

[0023] The rotation angle and the translation matrix between the first motion trajectory and the second motion trajectory are determined to realize spatial alignment of the first motion trajectory and the second motion trajectory, so as to complete calibration of the first radar sensor and the second radar sensor.

[0024] In the embodiments of the present application, the anti-collision strategy is executed according to the real obstacle information obtained by removing the non-obstacle information from the obstacle information, and the anti-collision strategy includes:

[0025] The obstacle distance and the relative speed of the detected obstacle with the working platform are determined according to the real obstacle information.

[0026] The risk level is calculated according to the obstacle distance and the relative speed; and

[0027] The anti-collision strategy is executed according to the risk level.

[0028] In the embodiments of the present application, the risk level is calculated according to the obstacle distance and the relative speed, and the risk level is calculated according to the following formula:

[0029] R = min(m, f(D) + v flag )

[0030] f(D) = m - 2 * (D - 1)

[0031]

[0032] wherein R is the risk level, m is the highest risk level, D is the obstacle distance, v flag is the speed threshold.

[0033] In the embodiments of the present application, the obstacle information of the obstacle detected by the radar sensor includes:

[0034] The obstacle information of the obstacle in the target direction detected by the radar sensor is obtained.

[0035] In the embodiments of the present application, the target direction is the motion direction of the working platform.

[0036] The second aspect of the present application provides a processor configured to execute the anti-collision control method described above.

[0037] The third aspect of the present application provides an anti-collision control system applied to an aerial work platform, the aerial work platform including a working platform, and the anti-collision control system including:

[0038] a visual sensor configured to collect images of a monitoring area, the monitoring area containing an activity area of a worker when the worker is working on the working platform;

[0039] a radar sensor configured to detect obstacles; and

[0040] The processor described above.

[0041] In the embodiment of the present application, the visual sensor is installed on the working platform through the support rod.

[0042] In the embodiment of the present application, the support rod is an electric telescopic support rod.

[0043] In the embodiment of the present application, the radar sensor includes a left radar sensor, a rear radar sensor and a right radar sensor respectively located at the left side, the rear side and the right side of the working platform, wherein the rear radar sensor and the left radar sensor have a first overlapping detection area, and the rear radar sensor and the right radar sensor have a second overlapping detection area.

[0044] The fourth aspect of the present application provides a high-altitude work platform, comprising:

[0045] a working platform; and

[0046] The anti-collision control system described above.

[0047] The fifth aspect of the present application provides a machine-readable storage medium, which stores instructions, and the execution of the instructions in the processor causes the processor to implement the anti-collision control method described above.

[0048] Through the above technical solution, the working condition of the worker in the working platform is obtained by using the visual sensor, and the information that may cause the radar sensor to misreport in the working process is identified, so as to avoid the false alarm of the anti-collision control system, greatly improve the use experience of the worker on the anti-collision control system. In addition, through the arrangement of the radar sensor in multiple directions, higher detection coverage of the environment around the working platform can be realized, and the safety of the working platform can be improved.

[0049] Other features and advantages of the embodiments of the present application will be described in detail in the following specific implementation part. BRIEF DESCRIPTION OF DRAWINGS

[0050] The accompanying drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification, and are used to explain the embodiments of the present application together with the following specific implementation, but do not constitute a limitation to the embodiments of the present application. In the drawings:

[0051] Figure 1A and Figure 1B The arrangement of various sensors in the anti-collision control system for the high-altitude work platform according to the embodiments of the present application is schematically shown;

[0052] Figure 2 The flow chart of the anti-collision control method for the high-altitude work platform according to the embodiments of the present application is schematically shown; and

[0053] Figure 3 A functional block diagram of a collision avoidance control system for a work platform is shown schematically.

[0054] Legend of reference signs

[0055] 1 processor 2 visual sensor

[0056] 3 radar sensor 3-1 rear side radar sensor

[0057] 3-2 left side radar sensor 3-3 right side radar sensor

[0058] 4 work personnel identification module 5 obstacle detection module

[0059] 6 multi-sensor data fusion module 7 collision avoidance control module

[0060] 8 support pole DETAILED DESCRIPTION

[0061] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely intended to explain and illustrate the present application, and are not intended to limit the present application.

[0062] It should be noted that if the present application has a description of directionality (such as up, down, left, right, front, back, etc.), the directionality is only used to explain the relative position relationship, movement, etc. between the components in a certain posture (as shown in the drawings), and if the certain posture changes, the directionality also changes accordingly.

[0063] If the present application has a description of "first", "second", etc., the description of "first", "second", etc. is only for description purposes, and cannot be understood as indicating or implying the relative importance of the technical features indicated, or implicitly indicating the number of technical features indicated. Therefore, the features limited by "first", "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions of each embodiment can be combined with each other, but it must be based on the realization of a person skilled in the art, and when the combination of technical solutions contradicts each other or cannot be realized, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope required by the present application.

[0064] In the present application, the "left side", "right side" and "rear side" of the work platform are defined with respect to the "front side" of the work platform, and the "front side" of the work platform refers to the side that the work personnel usually faces when standing on the work platform.

[0065] Figure 1A and Figure 1B The arrangement of various sensors in the anti-collision control system for aerial work platforms according to embodiments of the present application is schematically shown. Referring to Figure 1A and Figure 1B In embodiments of the present application, an anti-collision control system for aerial work platforms is provided. The aerial work platforms can include a work platform and other components, which can include but are not limited to, for example, a main body (e.g., a travelling mechanism), a jib, a lifting mechanism, etc., depending on the type of the aerial work platforms. The anti-collision control system can include:

[0066] a vision sensor 2 configured to capture images of a monitoring area, which contains an active area of a worker when the worker is working on the work platform;

[0067] a radar sensor 3 configured to detect obstacles; and

[0068] a processor 1.

[0069] In particular, the vision sensor 2 can include at least one camera, such as a CCD camera. In one example, the vision sensor 2 can be mounted on the work platform via a support pole 8. In particular, in one example, the support pole 8 can be an electrically extendable support pole 8, the height of the vision sensor 2 can be adjusted via the extension of the support pole 8, and the support pole 8 can be retracted to the same height as the work platform or below when not in use. The vision sensor 2 can be disposed at the top of the support pole 8. The support pole 8 can be mounted on a guardrail of the work platform.

[0070] Examples of the radar sensor 3 can include but are not limited to, ultrasonic sensors, millimeter wave sensors, and laser sensors. Preferably, the radar sensor 3 can be a millimeter wave sensor. In one example, the radar sensor 3 can include a left-side radar sensor 3-2, a rear-side radar sensor 3-1, and a right-side radar sensor 3-3, which are respectively located at the left side, the rear side, and the right side of the work platform. For example, these side radar sensors 3 can be disposed at the sides of the bottom of the work platform.

[0071] In addition, in order to reduce the detection blind area of the radar sensor 3 as much as possible, in embodiments of the present application, adjacent radar sensors 3 can have overlapping detection areas. For example, the rear-side radar sensor 3-1 and the left-side radar sensor 3-2 can have a first overlapping detection area. For another example, the rear-side radar sensor 3-1 and the right-side radar sensor 3-3 can have a second overlapping detection area.

[0072] In optional embodiments of the present application, a top radar sensor 3 can be disposed at the top of the support pole 8 for detecting the area above the work platform, so as to further increase the detection range of the radar sensor 3.

[0073] In optional embodiments of the present application, visual sensors 2 (e.g., cameras) can be arranged on the sides of the working platform to further monitor the environment around the working platform. For example, visual sensors 2 can be arranged on the left side, right side and rear side of the working platform, respectively.

[0074] Examples of the processor 1 can include, but are not limited to, a single-chip microcomputer, a microprocessor, a Field Programmable Gate Array (FPGA), a Programmable Logic Controller (PLC), a Digital Signal Processing (DSP), an Application Specific Integrated Circuit (ASIC), a state machine, etc.

[0075] The processor 1 can be configured to acquire the detection signal of the radar sensor 3 and the environmental image captured by the camera, determine whether there is a target obstacle according to the detection signal and the environmental image, and perform a corresponding anti-collision measure in the case of determining that there is a target obstacle. In an example, the anti-collision measure can include alarming / reminding, for example, alarming / reminding in the form of sound / light. The anti-collision control system can further include an alarm device, and the processor 1 can send an instruction to the alarm device to instruct the alarm device to issue an alarm, such as sound / light, when it is determined that the alarm is needed. In another example, the processor 1 can control the movement of the working platform. For example, the aerial work platform can include a driving mechanism for driving the movement of the working platform, such as an arm swing mechanism, an arm luffing mechanism, a platform lifting mechanism, etc. The processor 1 can control the driving mechanism to perform a corresponding action to avoid collision between the working platform and the target obstacle according to the obstacle information (e.g., position information, speed information, etc.) of the detected target obstacle, or send an instruction to the controller of the driving mechanism, and the controller can control the driving mechanism to perform a corresponding action after receiving the instruction.

[0076] Figure 2 A flowchart of an anti-collision control method for an aerial work platform according to an embodiment of the present application is schematically shown. As shown in Figure 2 The anti-collision control method can be applied to the anti-collision control system of any of the above embodiments in the embodiments of the present application. The anti-collision control method can include the following steps.

[0077] In step S210, the obstacle information of the obstacle detected by the radar sensor is acquired.

[0078] Specifically, for example, the processor can obtain a detection signal of the radar sensor from the radar sensor, and if the radar sensor detects an obstacle, the processor can obtain obstacle information of the detected obstacle from the detection signal. For example, taking the radar sensor as a millimeter wave radar sensor as an example, the millimeter wave radar sensor sends point cloud data to the processor, and the processor can determine the position information and the speed information (relative speed) of the detected obstacle from the point cloud data.

[0079] In the embodiments of the present application, the radar sensor can adopt Figure 1A and the arrangement shown in Figure B, that is, the radar sensor 3 can include a left radar sensor 3-2, a right radar sensor 3-3, and a rear radar sensor 3-1, and the rear radar sensor 3-1 has an overlapping detection area with the left radar sensor 3-2 and the right radar sensor 3-3, respectively. For the radar sensor pair of the rear radar sensor 3-1 and the left radar sensor 3-2, or for the radar sensor pair of the rear radar sensor 3-1 and the right radar sensor 3-3, in order to facilitate the description below, the radar sensor pair is referred to as a first radar sensor and a second radar sensor, and then the first detection area of the first radar sensor has an overlapping detection area with the second detection area of the second radar sensor. In order to avoid the case that the detection of an obstacle in the overlapping detection area causes both radar sensor pairs to report the same obstacle, the first radar sensor and the second radar sensor can be calibrated.

[0080] Specifically, the calibration method can include:

[0081] respectively obtaining first object information and second object information detected by the first radar sensor and the second radar sensor for a target object moving in the overlapping detection area;

[0082] generating a first motion trajectory in a first coordinate system of the first radar sensor and a second motion trajectory in a second coordinate system of the second radar sensor according to the first object information and the second object information, respectively;

[0083] determining a rotation angle and a translation matrix between the first motion trajectory and the second motion trajectory to realize spatial alignment of the first motion trajectory and the second motion trajectory, thereby completing calibration of the first radar sensor and the second radar sensor.

[0084] More specifically, the target object moves in the overlapping detection area of the first radar sensor and the second radar sensor (represented by radar A and radar B, respectively), ensuring that it can be detected by both radars at the same time. A motion trajectory is generated in the coordinate system of each radar, denoted as τ A and τ BAt this time, only the rotation angle and the translation matrix between the two trajectories need to be solved to realize the space alignment of the two trajectories, that is, to complete the calibration of the two radars.

[0085] The trajectory is represented by sequence discrete points, and A i represents the i-th point of the trajectory detected by radar A, B i represents the i-th point of the trajectory detected by radar B, and α AB represents the rotation angle from the radar A coordinate system to the radar B coordinate system. When α AB satisfies the requirements, the variance of the distance between the trajectories τ A and τ B should be minimum, so the solution of α AB can be converted into the following optimization problem:

[0086]

[0087] where N is the number of trajectory points, R AB is the rotation matrix, and satisfies

[0088]

[0089] Solve the above optimization problem:

[0090]

[0091] where R′ AB is the derivative of R AB with respect to α AB .

[0092] Let

[0093] Let the above partial derivative be zero, then:

[0094]

[0095] By calculating α AB , the alignment of the poses of the two radars can be realized, and the corresponding translation matrix can be calculated in combination with the trajectory distance.

[0096] In step S220, an image of a monitoring area collected by a visual sensor is obtained, and the monitoring area includes an activity area of a worker when the worker works on a working platform.

[0097] Specifically, the visual sensor (top visual sensor) takes a picture of the working platform from above, and can capture an image of the monitoring area including the working platform and a certain range around the working platform.

[0098] In step S230, it is determined whether the worker inside the image recognition work platform is working according to the worker.

[0099] Specifically, the image can be input to a target detection model, and the target detection model outputs a recognition result. The output result can include a detected hand of the worker or a work tool category, and a position relationship of a bounding box of the hand or the work tool relative to the work platform. According to the recognition result, it is determined whether the worker is working.

[0100] More specifically, when the target detection model identifies the hand of the worker or the work tool in the image, the identified hand or work tool can be labeled with a bounding box (detection frame). According to the position relationship of the bounding box relative to the work platform, it can be determined whether the worker is working at this time. For example, the worker may need to hold the work tool with the hand and stretch it to the outside of the work platform guardrail when working. At this time, the target detection model can identify the hand and the work tool and label them with a bounding box, and can determine that the worker is working according to the position relationship of the bounding box at this time.

[0101] In the embodiments of the present application, the target detection model can be a deep learning model. Examples thereof can include but are not limited to a one-stage detection model such as a model of the YOLO series, an SSD model, etc., and a two-stage detection model such as a model of the RCCN series, etc.

[0102] In step S240, in the case where it is identified that the worker is working, it is determined whether the non-obstacle information caused by the worker working is included in the obstacle information.

[0103] In step S250, in the case where it is determined that the non-obstacle information is included in the obstacle information, the non-obstacle information is removed.

[0104] Specifically, if it is identified that the worker inside the working platform is working, for example, the processor can determine the obstacle positions according to the detected obstacle information uploaded by the radar sensor, compare the obstacle positions with the position relationship of the bounding box, if it is determined that there is an obstacle position in the determined obstacle positions that matches the position relationship (for example, the position relationship of the bounding box is basically the same as the obstacle position), it can be determined that the obstacle at the obstacle position detected by the radar sensor is caused by the working action of the worker, and therefore the corresponding obstacle information is non-obstacle information, and the non-obstacle information is contained in the obstacle information uploaded by the radar sensor. The processor can remove the non-obstacle information from the obstacle information. In one example, the radar sensor is arranged on three sides of the working platform, i.e., the left side, the right side and the rear side, and if the non-obstacle information is determined to be from one of the three side radar sensors, for example, the left side radar sensor, the processor can directly shield (disable) the detection signal uploaded by the left side radar sensor.

[0105] In the embodiments of the present application, it can only be interested in whether there is a collision risk in some specified direction. For example, the working platform moves in a first direction (for example, moves towards the left side of the working platform), at this time, it can only be interested in whether there is a collision risk in front of the left side of the working platform. Therefore, for example, the processor can obtain the motion information (for example, the motion direction) of the working platform, and obtain the obstacle information of the detected obstacles in the target direction. For example, the processor can obtain the detection signal uploaded by the left side radar sensor from the left side radar sensor, and determine the obstacle information according to the detection signal. If at this time the worker faces a second direction to work (for example, faces the rear side of the working platform to work), whether the processor (target detection model) identifies whether the worker is working or not can be ignored for subsequent anti-collision strategies. That is, for the detection signal uploaded by the radar sensor, the processor only needs to associate the detection signal of the radar sensor in the target direction (for example, the motion direction of the working platform). For example, the working platform moves towards the left side, and then the processor only needs to obtain the detection signal from the left side radar sensor, and can shield (disable or discard) the detection signal uploaded from other radar sensors (for example, the right side radar sensor and the rear side radar sensor). In this case, only the position relationship of the bounding box output by the target detection model is associated with the target direction (for example, the worker's hand and the working tool are moving on the left side of the working platform), there is a possibility of matching the obstacle position detected by the radar sensor.

[0106] In step S260, an anti-collision strategy is performed according to the real obstacle information after the non-obstacle information is removed from the obstacle information.

[0107] Specifically, after removing the non-obstacle information, the remaining is the real obstacle information. The obstacle distance and the relative speed of the detected obstacle with the working platform can be determined according to the real obstacle information;

[0108] Calculate a risk level according to the obstacle distance and the relative speed; and

[0109] Execute a collision avoidance strategy according to the risk level.

[0110] The risk level is calculated according to the obstacle distance and the relative speed, including calculating the risk level according to the following formula:

[0111] R = min(m, f(D) + v flag )

[0112] f(D) = m - 2 * (D - 1)

[0113]

[0114] wherein R is the risk level, m is the highest risk level, D is the obstacle distance, v flag is the speed threshold.

[0115] After determining the risk level, a corresponding collision avoidance strategy can be formulated for different risk levels. For example, the collision avoidance strategy can include alarm (such as sound / light alarm), deceleration movement, emergency braking, etc., and when the risk level is higher, a higher degree of strategy (such as deceleration movement or even emergency braking) can be adopted.

[0116] The embodiment of the present application provides a processor 1 configured to execute the collision avoidance control method of any of the above embodiments.

[0117] Specifically, the processor 1 can be configured to:

[0118] Obtain obstacle information of an obstacle detected by a radar sensor 3;

[0119] Obtain an image of a monitoring area collected by a vision sensor 2, the monitoring area containing an activity area of a working personnel when working on a working platform;

[0120] Identify whether the working personnel is working according to the image;

[0121] In a case where it is identified that the working personnel is working, determine whether the obstacle information contains non-obstacle information caused by the working of the working personnel;

[0122] In a case where it is determined that the obstacle information contains the non-obstacle information, remove the non-obstacle information;

[0123] The anti-collision strategy is executed according to the real obstacle information in the obstacle information after the non-obstacle information is removed.

[0124] In the embodiment of the application, whether the worker is working is identified according to the image, including:

[0125] inputting the image into a target detection model;

[0126] The target detection model outputs the identification result, and the output result includes the detected hand of the worker or the category of the work tool and the position relationship of the bounding box of the hand or the work tool relative to the work platform;

[0127] According to the identification result, whether the worker is working is identified.

[0128] In the embodiment of the application, in the case where it is identified that the worker is working, whether the non-obstacle information caused by the worker working is included in the obstacle information is determined, including:

[0129] According to the obstacle information, the obstacle position of the detected obstacle is determined;

[0130] It is determined whether there is an obstacle position with a matching position relationship in the obstacle position;

[0131] In the case where it is determined that there is an obstacle position with a matching position relationship, it is determined that the non-obstacle information is included in the obstacle information.

[0132] In the embodiment of the application, the processor 1 can also be configured to:

[0133] respectively acquire first object information and second object information detected by the first radar sensor and the second radar sensor for movement of a target object in the overlapping detection area;

[0134] According to the first object information and the second object information, a first motion trajectory in a first coordinate system of the first radar sensor and a second motion trajectory in a second coordinate system of the second radar sensor are respectively generated;

[0135] A rotation angle and a translation matrix between the first motion trajectory and the second motion trajectory are determined to realize spatial alignment of the first motion trajectory and the second motion trajectory, so as to complete calibration of the first radar sensor and the second radar sensor.

[0136] In the embodiment of the application, the anti-collision strategy is executed according to the real obstacle information in the obstacle information after the non-obstacle information is removed, including:

[0137] According to the real obstacle information, the obstacle distance of the detected obstacle and the relative speed of the obstacle and the work platform are determined;

[0138] calculate a risk level according to the obstacle distance and the relative speed; and

[0139] execute a collision avoidance strategy according to the risk level.

[0140] In the embodiments of the present application, calculating the risk level according to the obstacle distance and the relative speed comprises calculating the risk level according to the following formula:

[0141] R = min(m, f(D) + v flag )

[0142] f(D) = m - 2 * (D - 1)

[0143]

[0144] wherein R is the risk level, m is the highest risk level, D is the obstacle distance, v flag is the speed threshold.

[0145] In the embodiments of the present application, obtaining the obstacle information of the obstacle detected by the radar sensor comprises:

[0146] obtaining the obstacle information of the obstacle in the target direction detected by the radar sensor.

[0147] In the embodiments of the present application, the target direction is the motion direction of the working platform.

[0148] Figure 3 A functional block diagram of the collision avoidance control system for the aerial work platform according to the embodiments of the present application is schematically shown. As Figure 3 shown, the processor 1 of the collision avoidance control system can functionally comprise:

[0149] a worker identification module 4 containing a target detection model, configured to: obtain the image of the monitoring area collected from the visual sensor 2, the monitoring area containing the activity area of the worker when working on the working platform; identify whether the worker is working according to the image, and output the identification result;

[0150] an obstacle detection module 5, configured to: obtain the detection signal of the detected obstacle from the radar sensor 3, and obtain the obstacle information according to the detection signal, the obstacle information can include the position information and the speed information of the obstacle;

[0151] The multi-sensor data fusion module 6 is configured to: receive the identification result and the obstacle information from the worker identification module 4 and the obstacle detection module 5 respectively, determine whether the non-obstacle information caused by the worker operation is contained in the obstacle information according to the identification result and the obstacle information according to the identification result that the worker is operating; and remove the non-obstacle information in a case where it is determined that the non-obstacle information is contained in the obstacle information.

[0152] The anti-collision control module 7 is configured to: obtain the real obstacle information after the non-obstacle information is removed from the multi-sensor data fusion module 6, and formulate and execute the anti-collision strategy according to the real obstacle information.

[0153] The embodiment of the application provides a kind of aerial work platform, comprising:

[0154] Work platform;And

[0155] The anti-collision control system of any of the above embodiments.

[0156] The embodiment of the application provides a kind of machine readable storage medium, which stores instructions, and the execution in the processor is executed when the processor is executed to make the processor realize the anti-collision control method of any of the above embodiments.

[0157] The scheme provided by the embodiment of the application can have at least one of the following beneficial effects:

[0158] (1) the semantic information in the image is obtained by using the visual sensor to remove the non-obstacle information, effectively reduces the false alarm rate of the anti-collision control system, and ensures the correctness of the anti-collision control action.

[0159] (2) the detection range of the environment around the work platform is improved by arranging the multi-directional radar sensor. At the same time, a design of visual sensor is provided, which facilitates the camera to obtain the working condition in the work platform and can be used to identify the hand or working tool of the worker to avoid false alarm.

[0160] (3) the two-by-two calibration method is used for the plurality of radar sensors, which can support the joint calibration of the radar sensors in the radar sensor arrangement scheme in the embodiment of the application, effectively improve the ranging accuracy of the radar sensor, and also can avoid that the same obstacle is detected as different obstacles by two radar sensors.

[0161] (4) the anti-collision control method includes the calculation of risk level, not only considers the spatial absolute position relationship between the obstacle and the work platform, but also considers the relative motion relationship, which can more accurately describe the collision risk situation.

[0162] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk memory, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0163] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor 1 of a general purpose computer, special purpose computer, embedded processor or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor 1 of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams block or blocks.

[0164] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart illustrations and / or block diagrams block or blocks.

[0165] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart illustrations and / or block diagrams block or blocks.

[0166] In one typical arrangement, the computing device includes one or more processors 1 (CPU), input / output interfaces, network interfaces, and memory.

[0167] The memory can include non-persistent memory and / or volatile memory, such as a random access memory (RAM) including a cache area for the temporary storage of data. The memory can also include non-volatile memory, such as read only memory (ROM) for storing structural information and / or instruction code. Note that the memory can also include other volatile memory that is not a RAM, such as static non-volatile memory (e.g., flash memory) or forms of non-volatile memory that are not a ROM. The memory is an example of computer readable storage media.

[0168] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0169] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not only include those elements, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0170] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A collision avoidance control method characterized by, The application is applied to a high-altitude operation platform, the high-altitude operation platform comprises a work platform and an anti-collision control system, the anti-collision control system comprises a visual sensor for collecting images of a monitoring area and a radar sensor for detecting obstacles, and the control method comprises the following steps: obtaining obstacle information of obstacles detected by the radar sensor; obtaining images of the monitoring area collected by the visual sensor, the monitoring area containing an activity area of a worker when the worker is working on the work platform; identifying whether the worker is working according to the images; in the case that it is identified that the worker is working, determining whether non-obstacle information caused by worker work is contained in the obstacle information; in the case that it is determined that the non-obstacle information is contained in the obstacle information, removing the non-obstacle information; executing an anti-collision strategy according to real obstacle information obtained by removing the non-obstacle information from the obstacle information; wherein the step of identifying whether the worker is working according to the images comprises: inputting the images into a target detection model; the target detection model outputs an identification result, the identification result comprising a detected hand or work tool category of the worker and a position relationship of the hand or work tool relative to the work platform; identifying whether the worker is working according to the identification result; the step of, in the case that it is identified that the worker is working, determining whether non-obstacle information caused by worker work is contained in the obstacle information comprises: determining an obstacle position of a detected obstacle according to the obstacle information; determining whether there is an obstacle position matching the position relationship in the obstacle position; in the case that it is determined that there is an obstacle position matching the position relationship, determining that the non-obstacle information is contained in the obstacle information.

2. The collision avoidance control method according to claim 1, characterized by, the radar sensor comprises a first radar sensor and a second radar sensor, a first detection area of the first radar sensor and a second detection area of the second radar sensor have an overlapping detection area, and the anti-collision control method further comprises: respectively obtaining first object information and second object information detected by the first radar sensor and the second radar sensor for a target object moving in the overlapping detection area; generating a first motion trajectory in a first coordinate system of the first radar sensor and a second motion trajectory in a second coordinate system of the second radar sensor according to the first object information and the second object information respectively; determining a rotation angle and a translation matrix between the first motion trajectory and the second motion trajectory to realize spatial alignment of the first motion trajectory and the second motion trajectory, thereby completing calibration of the first radar sensor and the second radar sensor.

3. The collision avoidance control method according to claim 1, characterized by, the step of executing an anti-collision strategy according to real obstacle information obtained by removing the non-obstacle information from the obstacle information comprises: determining an obstacle distance of a detected obstacle and a relative speed of the obstacle and the work platform according to the real obstacle information; calculating a risk level according to the obstacle distance and the relative speed; and executing a collision avoidance strategy according to the risk level.

4. The collision avoidance control method according to claim 3, characterized by, The calculating a risk level according to the obstacle distance and the relative speed comprises calculating a risk level according to the following formula: R = min( m , f ( D ) + v flag ) f ( D ) = m －2*(D－1) wherein, R is the risk level, m is the highest risk level, D is the obstacle distance, is the speed threshold.

5. The collision avoidance control method according to any one of claims 1 to 4, characterized by, The obtaining the obstacle information of the obstacle detected by the radar sensor comprises: Obtaining the obstacle information of the obstacle in the target direction detected by the radar sensor.

6. The collision avoidance control method according to claim 5, characterized by, The target direction is the moving direction of the working platform.

7. A processor, comprising: A processor configured to implement the collision avoidance control method according to any one of claims 1 to 6.

8. An anti-collision control system characterized by, The application is applied to an aerial work platform, and the collision avoidance control system comprises: a visual sensor configured to collect images of a monitoring area, the monitoring area containing an activity area of a worker when the worker is working on the working platform; a radar sensor configured to detect obstacles; and The processor according to claim 7.

9. The collision avoidance control system according to claim 8, characterized by characterized in that The visual sensor is installed on the working platform through a support rod.

10. The collision avoidance control system according to claim 9, characterized by The support rod is an electrically retractable support rod.

11. The collision avoidance control system of claim 8, wherein, The radar sensor comprises a left radar sensor, a rear radar sensor and a right radar sensor respectively located at the left side, the rear side and the right side of the working platform, wherein the rear radar sensor and the left radar sensor have a first overlapping detection area, and the rear radar sensor and the right radar sensor have a second overlapping detection area.

12. An aerial work platform, characterized by comprise: a working platform; and the collision avoidance control system according to any one of claims 8 to 11.

13. A machine-readable storage medium, characterized in that, The machine-readable storage medium has instructions stored thereon, and the instructions, when executed by a processor, cause the processor to implement the collision avoidance control method according to any one of claims 1 to 6.

Images