Garbage bottle picking robot and control method

The garbage bottle picking robot uses neural network chips and visual systems to identify and clean garbage bottles, solving the problem that existing sweeping robots have difficulty in dealing with large garbage, and improving recycling rate and cleaning effect.

CN111823212BInactive Publication Date: 2024-11-15SILING ROBOT TECH (HARBIN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010698942.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-07-20
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing sweeping robots are difficult to efficiently clean large garbage such as plastic bottles, and the cleaning effect is poor in complex environments, resulting in low recycling rates of recyclable materials and increased pollution.

Method used

A garbage bottle cleaning robot is designed, using neural network chips combined with lidar and binocular vision cameras, and is controlled through the robot system framework ROS to realize the identification and efficient cleaning of garbage bottles, and enhance the cleaning ability in complex environments.

Benefits of technology

It improves the recycling rate of recyclable materials, reduces the difficulty of daily garbage cleaning, and provides better cleaning results in special environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111823212B_ABST
    Figure CN111823212B_ABST
Patent Text Reader

Abstract

The present invention discloses a garbage bottle picking robot and a control method, comprising a bottom plate, an intermediate layer partition, a shell and a mechanical arm, and also comprising a micro host, a development board, a binocular vision camera, and a laser radar. The robot is controlled by the micro host as a high-level machine and the development board as a low-level machine, and the garbage bottles are identified by the neural network chip included in the micro host, and then the hand-eye system and other devices are controlled to pick up the garbage bottles. The robot generates maps, automatically inspects and avoids obstacles through devices such as laser radar and GPS modules, and can also be remotely operated and obtain various types of data through a remote controller. The robot provided by the present invention can be used for automatic picking of garbage bottles, and has the advantages of realizing automatic picking and classification of garbage bottles, reducing the difficulty of garbage bottle recycling, beautifying the urban environment, and saving labor and cleaning costs.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of intelligent robots, and in particular to a garbage bottle picking robot and a control method thereof. Background Art

[0002] As the phenomenon of aging in Chinese society becomes increasingly prominent, domestic labor costs continue to rise, and the demographic dividend gradually disappears, the current method of manual maintenance and cleaning has problems such as artificial blind spots, low cleaning efficiency, and untimely cleaning. In addition, cleaners are not only very hard-working, with mechanical and repetitive work and high labor intensity, but the working environment leads to a higher incidence of chronic respiratory diseases, and even the complex and changeable traffic and road conditions can put their lives at risk.

[0003] However, existing sweeping robots still have many disadvantages: 1. Sweeping robots generally use a rotating brush structure, which makes internal cleaning and maintenance difficult; 2. Most sweeping robots have limited application scenarios and are only suitable for dry and flat indoor floors; 3. The application range is also small, and they are only suitable for cleaning small garbage such as dust and debris, and are powerless against slightly larger garbage such as plastic bottles. Garbage bottles are recyclable garbage. Placing plastic bottles or glass bottles together with other garbage will increase the difficulty of recycling, reduce the recovery rate of recyclable materials, and increase pollution. Summary of the invention

[0004] In view of this, the present invention provides a garbage bottle picking robot.

[0005] The present invention provides a garbage bottle picking robot, comprising: a bottom plate, an intermediate layer partition plate, a shell, a mechanical arm and a micro host, wherein a driving wheel, a stepping motor and a universal driven wheel are arranged on the lower surface of the bottom plate, and the micro host, a development board and a stepping motor driver are arranged on the upper surface of the bottom plate, wherein the stepping motor driver is used to drive the stepping motor, thereby driving the driving wheel to control the movement of the robot, and the micro host is electrically connected to the development board, and serves as a host computer to control the development board as a slave computer;

[0006] The middle layer partition is fixedly connected to the bottom plate through a connecting column, and the mechanical arm is fixedly connected to one end of the upper surface of the middle layer partition;

[0007] The outer shell is fixed on the bottom plate and wraps the middle layer partition. A reserved opening is provided at one end of the upper surface of the outer shell so that the mechanical arm can extend from the reserved opening and put the cleaned garbage bottles into a garbage basket fixed in the middle of the upper surface of the outer shell. A laser radar is fixed at the other end of the upper surface.

[0008] The end of the robotic arm is equipped with a binocular vision camera, which together with the robotic arm forms a hand-eye system for picking up garbage bottles;

[0009] The micro host includes a neural network chip, which is used to determine whether the target to be picked up is a garbage bottle based on feedback information obtained by the laser radar and the binocular vision camera.

[0010] The present invention designs a garbage bottle picking robot, which learns through a neural network chip, determines whether the target to be picked up is a garbage bottle according to the feedback information obtained by the laser radar and the binocular vision camera, and performs efficient picking processing. This can increase the recovery rate of recyclable materials, reduce the difficulty of daily garbage cleaning, and can also achieve better cleaning effects than traditional garbage cleaning robots in special environments such as environments with a large number of tourists.

[0011] Preferably, the development board is an Arduino Uno development board.

[0012] Preferably, there are 6 connecting columns, so that the overall structure is more stable.

[0013] Preferably, there are two driving wheels, two universal driven wheels, two stepping motors and two stepping motor drivers.

[0014] Preferably, a step-down module, a 4G communication module, a GPS module and a battery are also provided on the bottom plate;

[0015] The micro host is electrically connected to the development board, laser radar, robotic arm, binocular vision camera, 4G communication module and GPS module, and is used to communicate as a host computer and control all components electrically connected to the micro host;

[0016] Preferably, the binocular vision camera and the mechanical arm are connected to the host computer, and the camera is used to identify the garbage bottle, and the mechanical arm is used to grab it. The camera is fixedly set on the mechanical arm gripper, so that the shooting angle of the camera can be controlled by judging the posture of the mechanical arm.

[0017] Optionally, the laser radar is connected to a host computer to assist in determining the size, shape and material of the garbage bottles.

[0018] A 4G antenna is also fixed on the shell and is electrically connected to the 4G communication module.

[0019] Optionally, the robot is installed with other communication modules such as a Bluetooth module for remote communication.

[0020] Preferably, the stepper motor driver is electrically connected to the development board and the stepper motor at the same time, and receives pulse information from the development board, thereby driving the stepper motor according to the pulse information, and the stepper motor drives the driving wheel to move.

[0021] Preferably, the battery is electrically connected to the stepper motor driver, micro host, laser radar and robotic arm through a step-down module to provide power.

[0022] Preferably, the garbage bottle picking robot also includes a WIFI module.

[0023] The WIFI module is electrically connected to the micro host for remote communication.

[0024] Optionally, determine whether the WIFI module is connected to the network. If connected, use the WIFI module for remote communication. If not connected, use the 4G communication module for remote communication.

[0025] Optionally, the robot has two stepper motors, and an acceleration sensor is installed near each stepper motor and electrically connected to the development board to correct the speed of the robot.

[0026] A laser ranging sensor and a temperature and humidity sensor are also installed at the end of the robotic arm and are electrically connected to the development board to correct the accuracy of the robot.

[0027] The present invention also provides a control method for a garbage bottle picking robot, which is designed based on the robot system framework ROS and controls any of the above-mentioned garbage bottle picking robots. The method includes:

[0028] S1. The robot's micro host obtains a map of the working area through a laser radar;

[0029] S2, the micro host automatically plans a route inspection in the working area;

[0030] S3, the neural network chip controls the laser radar and the binocular vision camera to identify garbage bottles during the inspection process;

[0031] S4, the laser radar combines with the GPS module to calculate the coordinates of the garbage bottle relative to the robot on the map;

[0032] S5. The robot drives to the location of the garbage bottle;

[0033] S6, using a binocular vision camera to measure relative position information between the garbage bottle and the robotic arm;

[0034] S7. The micro host controls the robotic arm to pick up the garbage bottle and put it into the garbage basket, and continues the inspection.

[0035] Preferably, the map of the working area is constructed by SLAM.

[0036] Optionally, the map of the working area may also be obtained through a WIFI module or other remote communication means.

[0037] Preferably, the control method of the garbage bottle picking robot further includes:

[0038] The micro host communicates with the remote controller through the WIFI module or 4G module electrically connected to the micro host, and the communication includes: receiving a remote control signal sent by the remote controller, or sending surrounding environment obstacle information, robot positioning information and robot field of view information to the remote controller.

[0039] Preferably, step S3 comprises:

[0040] The neural network chip constructs a neural network for training based on the feedback information from the laser radar and the binocular vision camera, and recognizes the shape and material of the garbage bottle.

[0041] Optionally, the laser radar and binocular vision camera respectively obtain target information to construct corresponding training sets and verification sets, and construct a neural network for training, and use artificial intelligence image recognition technology to identify the shape of the garbage bottle and determine whether the target is a garbage bottle. The laser radar and binocular vision camera can also be used together to train the neural network to further improve the accuracy of judging the material of the garbage bottle and achieve the effect of accurately picking up plastic bottles or glass bottles.

[0042] The present invention also provides a control system for a garbage bottle picking robot, which is designed based on the robot system framework ROS and controls any of the above-mentioned garbage bottle picking robots, including:

[0043] Map service node, used to obtain the map of the working area through the lidar and GPS modules;

[0044] A path planning node, used to automatically plan a route inspection in a working area through the micro host;

[0045] A neural network node, used to control the laser radar and binocular vision camera to identify garbage bottles during inspection through the neural network chip;

[0046] A GPS node, used to calculate the coordinates of the garbage bottle relative to the robot on the map through the GPS module and the laser radar;

[0047] A motor control node, used for driving to the location of the garbage bottle;

[0048] A camera node, used to measure the relative position information between the garbage bottle and the robotic arm through the binocular vision camera;

[0049] The robot arm node is used to control the robot arm to pick up the garbage bottle and put it into the garbage basket through the micro host, and continue to inspect.

[0050] Preferably, the system further comprises:

[0051] The laser radar node is used to publish the obstacle information scanned by the laser radar to the path planning node, or publish the scanned garbage bottle information to the neural network node.

[0052] The robot posture publishing node is used to publish the coordinate transformation of each robot sensor and component to each node, such as sending the value of the laser ranging sensor to the micro host for processing.

[0053] Preferably, the system further comprises:

[0054] A communication node is used to communicate with a remote controller through the micro host via the WIFI module or 4G module electrically connected to the micro host, wherein the communication includes: receiving a remote control signal sent by the remote controller, or sending surrounding environment obstacle information, robot positioning information and robot field of view information to the remote controller.

[0055] Preferably, the remote controller can be an industrial robot controller or a mobile phone APP, which can communicate with the robot's communication node through a WIFI module or a 4G module connected to the micro host, can manually control the robot's travel, receive information released by the robot's laser radar, GPS module, and binocular vision camera module, and can remotely control the robot both indoors and outdoors.

[0056] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements part or all of the steps of any one of the control methods for a garbage bottle picking robot recorded in the above method embodiments when executing the computer program.

[0057] The present invention also provides a computer-readable storage medium, which stores a computer program, and is characterized in that when the computer program is executed by a processor, it implements part or all of the steps of any one of the control methods for garbage bottle picking robots recorded in the above method embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0059] Figure 1 This is a schematic diagram of the structure of a garbage bottle picking robot in an embodiment of the present invention;

[0060] Figure 2 This is a flow chart of a control method of a garbage bottle picking robot in an embodiment of the present invention;

[0061] Figure 3 2 is a schematic diagram of the right-side internal structure of the robot in an embodiment of the present invention;

[0062] Figure 4 Schematic diagram of the left internal structure of the robot in an embodiment of the present invention;

[0063] Figure 5 It is a schematic diagram of actual communication of each node in an embodiment of the present invention.

[0064] In the figure:

[0065] 1. Base plate 2. Driving wheel 3. Stepper motor 4. Driven wheel 5. Micro host 6. Development board 7. Stepper motor driver 8. Buck module 9. 4G communication module 10. GPS module 11. Battery 12. Middle layer partition 13. Connecting column 14. Robotic arm 15. Shell 16. Trash basket 17. 4G antenna 18. LiDAR 19. Binocular vision camera. DETAILED DESCRIPTION

[0066] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0067] An embodiment of the present invention provides a garbage bottle picking robot, such as Figure 1 As shown, it includes: a bottom plate 1, an intermediate layer partition plate 12, a shell 15, a mechanical arm 14 and a micro host 5. The lower surface of the bottom plate 1 is provided with a driving wheel 2, a stepper motor 3 and a universal driven wheel 4. The upper surface of the bottom plate 1 is provided with the micro host 5, a development board 6 and a stepper motor driver 7. The stepper motor driver 7 is used to drive the stepper motor 3, thereby driving the driving wheel 2 to control the movement of the robot. The micro host 5 is electrically connected to the development board 6, and acts as a host computer to control the development board 6 as a slave computer.

[0068] The middle layer partition plate 12 is fixedly connected to the bottom plate 1 through a connecting column 13, and the mechanical arm 14 is fixedly connected to one end of the upper surface of the middle layer partition plate 12;

[0069] The outer shell 15 is fixed on the bottom plate 1 and wraps the middle layer partition 12. A reserved opening is provided at one end of the upper surface of the outer shell 15 so that the mechanical arm 14 can extend from the reserved opening and put the cleaned garbage bottles into the garbage basket 16 fixed in the middle of the upper surface of the outer shell 15. A laser radar 18 is fixed at the other end of the upper surface.

[0070] A binocular vision camera 19 is mounted at the end of the robotic arm 14, which together with the robotic arm 14 forms a hand-eye system for cleaning and picking up garbage bottles;

[0071] The micro host 5 includes a neural network chip, which is used to determine whether the target to be picked up is a garbage bottle based on the feedback information obtained by the laser radar 18 and the binocular vision camera 19.

[0072] Preferably, the bottom plate 1 is also provided with a step-down module 8, a 4G communication module 9, a GPS module 10 and a battery 11. Figure 3 and Figure 4 As shown;

[0073] The micro host 5 is electrically connected to the development board 6, the laser radar 18, the mechanical arm 14, the binocular vision camera 19, the 4G communication module 9 and the GPS module 10, and is used to communicate as a host computer and control all components electrically connected to the micro host 5;

[0074] A 4G antenna 17 is also fixed on the housing 15 and is electrically connected to the 4G communication module 9 .

[0075] Preferably, the stepper motor driver 7 is electrically connected to the development board 6 and the stepper motor 3 at the same time, and receives pulse information from the development board 6, thereby driving the stepper motor 3 according to the pulse information, and the stepper motor 3 drives the driving wheel 2 to move.

[0076] Preferably, the battery 11 is electrically connected to the stepper motor driver 7, the micro host 5, the laser radar 18 and the robotic arm 14 through the step-down module 8 to provide power.

[0077] Preferably, the garbage bottle picking robot also includes a WIFI module.

[0078] The WIFI module is electrically connected to the micro host 5 for remote communication.

[0079] The embodiment of the present invention also provides a control method for a garbage bottle picking robot, which is designed based on the robot system framework ROS and controls any of the above-mentioned garbage bottle picking robots, such as Figure 2 As shown, the method includes:

[0080] S1. The micro host of the robot obtains a map of the working area through the laser radar;

[0081] S2, the micro host automatically plans a route inspection in the working area;

[0082] S3, the neural network chip controls the laser radar and the binocular vision camera to identify garbage bottles during the inspection process;

[0083] S4, the laser radar combines with the GPS module to measure the coordinates of the garbage bottle relative to the robot on the map;

[0084] S5, the robot drives to the location of the garbage bottle;

[0085] S6, the binocular vision camera measures the relative position information between the garbage bottle and the robotic arm;

[0086] S7. The micro host controls the robotic arm to pick up the garbage bottle and put it into the garbage basket, and continue the inspection.

[0087] Preferably, the control method of the garbage bottle picking robot further includes:

[0088] The micro host communicates with the remote controller through the WIFI module or 4G module electrically connected to the micro host, and the communication includes: receiving a remote control signal sent by the remote controller, or sending surrounding environment obstacle information, robot positioning information and robot field of view information to the remote controller.

[0089] Preferably, step S3 comprises:

[0090] The neural network chip constructs a neural network for training based on the feedback information from the laser radar and the binocular vision camera, and recognizes the shape and material of the garbage bottle.

[0091] Optionally, the laser radar and binocular vision camera respectively construct corresponding training sets and verification sets through feedback information, and construct a neural network for training, and use artificial intelligence image recognition technology to identify the shape of the garbage bottle and determine whether the target is a garbage bottle. The laser radar and binocular vision camera can also be used together to train the neural network to further improve the accuracy of judging the material of the garbage bottle and achieve the effect of accurately picking up plastic bottles or glass bottles.

[0092] The cellular neural network or convolutional neural network (CNN) chip involved in the above embodiments may also be replaced by other machine vision recognition chips in actual implementation, such as machine vision recognition chips based on shallow learning, machine vision recognition chips based on deep learning algorithms such as autoencoder, sparse coding, restricted Boltzmann machine (RBM), and deep belief network.

[0093] An embodiment of the present invention further provides a control system for a garbage bottle picking robot, which is designed based on the robot system framework ROS and controls any of the above-mentioned garbage bottle picking robots, including:

[0094] A map service node, used for obtaining a map of a working area through the laser radar;

[0095] A path planning node, used to automatically plan a route inspection in a working area through the micro host;

[0096] A neural network node, used to control the laser radar and binocular vision camera to identify garbage bottles during inspection through the neural network chip;

[0097] A GPS node, used to calculate the coordinates of the garbage bottle relative to the robot on the map through the GPS module and the laser radar;

[0098] A motor control node, used for driving to the location of the garbage bottle;

[0099] A camera node, used to measure the relative position information between the garbage bottle and the robotic arm through the binocular vision camera;

[0100] The robot arm node is used to control the robot arm to pick up the garbage bottle and put it into the garbage basket through the micro host, and continue to inspect.

[0101] In a specific embodiment, each node is designed based on the robot system framework ROS, and the specific node communication diagram is as follows: Figure 5 As shown:

[0102] The motor control node processes the speed command and outputs a pulse signal to the lower computer according to the speed command. This node will subscribe to the cmd_vel topic published by the path planning node or the remote control node. The specific message format of the topic is mainly:

[0103] Vector3 linear

[0104] Vector3 angular

[0105] Vector3 is defined in the ROS system framework as:

[0106] float32 x

[0107] float32 y

[0108] float32 z

[0109] This topic contains the linear velocity and angular velocity of the robot in the three directions of x, y, and z in the spatial coordinate system. After receiving the topic, the motor control node resolves the velocity into the rotation speed of the two driving wheels, and then further converts it into a pulse frequency to control the pulse signal of the lower computer pin, thereby driving the motor.

[0110] In a specific embodiment, the map service node publishes the map obtained by SLAM in the map topic. The main message format of the map topic is as follows:

[0111] MapMetaData info

[0112] int8[]data

[0113] This topic contains the map's metadata and occupancy probability.

[0114] In a specific embodiment, the GPS node publishes an odom topic, and the odom topic message format is mainly:

[0115] geometry_msgs / PoseWithCovariance pose

[0116] This topic contains the location information of the robot. The GPS node calibrates and converts the longitude and latitude coordinates provided by the GPS on the map obtained by SLAM, and then publishes the odom topic.

[0117] In a specific embodiment, the camera node publishes the captured image information through the compressed_image topic, identifies plastic bottles through a trained network, and monitors in real time whether there is plastic bottle garbage in the field of view. If a target is detected, the distance of the target is measured using the binocular parallax, the coordinates of the target on the SLAM map are calculated, and a move_goal action request is issued. After receiving feedback information that the request is completed, the relative position of the target and the robot is further measured, and a grip_goal action request is issued. When the target is not detected, a reachable target point is randomly generated and a move_goal request is sent.

[0118] The main message format of compressed_image is:

[0119] string format

[0120] uint8[]data

[0121] This topic includes the format of image data and the buffer data of the image.

[0122] The main message format of move_goal is:

[0123] float64 x

[0124] float64 y

[0125] float64 z

[0126] The action message indicates the location coordinates of the target point on the SLAM map.

[0127] The message format of grip_goal is:

[0128] float64 x

[0129] float64 y

[0130] float64 z

[0131] float64 w

[0132] Contains the coordinates of the target relative to the robot, as well as the target's posture.

[0133] In a specific embodiment, after receiving the grip_goal action request, the robot arm node learns the target posture of the robot arm through inverse kinematics, controls the movement of the robot arm, grabs the target, puts it into the trash can, and then feeds back the completion information to the camera node.

[0134] In a specific embodiment, the path planning node is a path planning node provided in ROS, including global planning and local planning nodes. The path planning node subscribes to tf, odom, scan, and map topics. After receiving the move_goal action request, the overall path to the target point is planned through global planning, and the driving speed of each small section is planned through local planning to achieve real-time obstacle avoidance. The speed is published through the cmd_vel topic. After reaching the vicinity of the target point, the completion information is fed back to the camera node. The path planning node will also update the SLAM map.

[0135] Preferably, the system further comprises:

[0136] The laser radar node is used to publish the obstacle information scanned by the laser radar to the path planning node, or publish the scanned garbage bottle information to the neural network node.

[0137] The robot posture publishing node is used to publish the coordinate transformation of each robot sensor and component to each node, such as sending the value of the laser ranging sensor to the micro host for processing.

[0138] In a specific embodiment, the laser radar node is responsible for publishing the obstacle information scanned by the radar to the path planning node, publishing the scan topic, and the message format of the topic is mainly:

[0139] Header header

[0140] float32 angel_min

[0141] float32 angle_max

[0142] float32 angle_increment

[0143] float32 time_increment

[0144] float32 scan_time

[0145] float32 range_min

[0146] float32 range_max

[0147] float32[]ranges

[0148] float32[]intensities

[0149] This topic contains the ID sequence that increases in the scanning order, the timestamp of the laser data, the name of the scan data, the angle at which the scan starts, the angle at which the scan ends, the angle increased for each scan, the measurement time interval, the scanning time interval, the minimum value of the distance measurement, the maximum value of the distance measurement, the obstacle distance in a week, and the light intensity. These topic messages will first be used for SLAM to build a map of the robot's surroundings. SLAM has a function package in ROS that can be called directly. The map is represented in the form of a raster map and is used for global planning in navigation. This topic will also be subscribed by the path planning node for local planning in navigation and updating the map.

[0150] In a specific embodiment, the robot posture publishing node publishes the tf topic, and the message format of the topic is mainly:

[0151] Transform tf::Transform

[0152] This topic contains the coordinate transformation data of the robot's components and sensors.

[0153] Preferably, the system further comprises:

[0154] A communication node is used to communicate with a remote controller through the micro host via the WIFI module or 4G module electrically connected to the micro host, wherein the communication includes: receiving a remote control signal sent by the remote controller, or sending surrounding environment obstacle information, robot positioning information and robot field of view information to the remote controller.

[0155] Preferably, the remote controller can be an industrial robot controller or a mobile phone APP, which can communicate with the robot's communication node through a WIFI module or a 4G module connected to the micro host, can manually control the robot's travel, receive information released by the robot's laser radar, GPS module, and binocular vision camera module, and can remotely control the robot both indoors and outdoors.

[0156] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements part or all of the steps of any one of the control methods for a garbage bottle picking robot recorded in the above method embodiments when executing the computer program.

[0157] The present invention also provides a computer-readable storage medium, which stores a computer program, and is characterized in that when the computer program is executed by a processor, it implements part or all of the steps of any one of the control methods for garbage bottle picking robots recorded in the above method embodiments.

[0158] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present disclosure, and will not be described in detail here.

[0159] Those of ordinary skill in the art will appreciate that the method steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0160] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A garbage bottle picking robot, characterized in that: include: A base plate (1), an intermediate layer partition plate (12), a housing (15), a mechanical arm (14) and a micro-host (5); a driving wheel (2), a stepping motor (3) and a universal driven wheel (4) are arranged on the lower surface of the base plate (1); the micro-host (5), a development board (6) and a stepping motor driver (7) are arranged on the upper surface of the base plate (1); the stepping motor driver (7) is used to drive the stepping motor (3), thereby driving the driving wheel (2) to control the movement of the robot; the micro-host (5) is electrically connected to the development board (6) and serves as a host computer to control the development board (6) as a slave computer; The intermediate layer partition plate (12) is fixedly connected to the bottom plate (1) via a connecting column (13), and the mechanical arm (14) is fixedly connected to one end of the upper surface of the intermediate layer partition plate (12); The shell (15) is fixed on the bottom plate (1) and wraps the middle layer partition (12). A reserved opening is provided at one end of the upper surface of the shell (15) so that the mechanical arm (14) can extend from the reserved opening and put the cleaned garbage bottles into a garbage basket (16) fixed in the middle of the upper surface of the shell (15). A laser radar (18) is fixed at the other end of the upper surface. A binocular vision camera (19) is mounted at the end of the robotic arm (14), and together with the robotic arm (14) forms a hand-eye system for cleaning and picking up garbage bottles; The micro host (5) includes a neural network chip, and the neural network chip is used to determine whether the target to be cleaned is a garbage bottle based on feedback information obtained by the laser radar (18) and the binocular vision camera (19); The bottom plate (1) is also provided with a voltage reduction module (8), a 4G communication module (9), a GPS module (10) and a battery (11); The micro host (5) is electrically connected to the development board (6), the laser radar (18), the robotic arm (14), the binocular vision camera (19), the 4G communication module (9) and the GPS module (10), and is used to communicate as a host computer and control all components electrically connected to the micro host (5); A 4G antenna (17) is also fixed on the housing (15) and is electrically connected to the 4G communication module (9); The micro host obtains a map of the working area through the laser radar; Also includes: camera node and path planning node; The camera node publishes the captured image information through a compressed image topic, identifies trash bottles through a trained network, and monitors in real time whether there are trash bottles and trash in the field of view. If a target is detected, the distance of the target is measured using the parallax of the binocular vision camera, the coordinates of the target on the map are calculated, and a mobile target action request is issued. After receiving feedback information on the completion of the request, the relative position of the target and the robot is further measured, and a control target action request is issued; when the target is not detected, a reachable target point is randomly generated, and a mobile target request is sent; The path planning node updates the map; The micro host obtains a map of the working area through the laser radar, including: the laser radar node is responsible for publishing the obstacle information scanned by the radar to the path planning node, publishing a scanning topic to build a map corresponding to the robot's surrounding environment; wherein the scanning topic contains an ID sequence that increases in scanning order, a timestamp of the laser data, a name of the scanning data, an angle at which the scan starts, an angle at which the scan ends, an angle increased for each scan, a measurement time interval, a scanning time interval, a minimum value of the distance measurement, a maximum value of the distance measurement, a distance of obstacles in one week, and light intensity; the path planning node also subscribes to the scanning topic for local planning in navigation and updating the map; The battery (11) is electrically connected to the stepper motor driver (7), the microcomputer (5), the laser radar (18) and the mechanical arm (14) through the step-down module (8) to provide power.

2. A garbage bottle picking robot according to claim 1, characterized in that: The stepper motor driver (7) is electrically connected to the development board (6) and the stepper motor (3) at the same time, and receives pulse information from the development board (6), thereby driving the stepper motor (3) according to the pulse information, and the stepper motor (3) drives the driving wheel (2) to move.

3. The garbage bottle picking robot according to claim 1, characterized in that: Also includes WIFI module, The WIFI module is electrically connected to the micro host (5) for remote communication.

4. A control method for a garbage bottle picking robot, designed based on a robot system framework ROS, to control the garbage bottle picking robot as claimed in any one of claims 1 to 3, comprising: S1. The robot's micro host obtains a map of the working area through a laser radar; S2, the micro host automatically plans a route inspection in the working area; S3, the neural network chip controls the laser radar and binocular vision camera to identify garbage bottles during the inspection process; S4, the laser radar combines with the GPS module to measure the coordinates of the garbage bottle relative to the robot on the map; S5, the robot drives to the location of the garbage bottle; S6, the binocular vision camera measures the relative position information between the garbage bottle and the robotic arm; S7, the micro host controls the mechanical arm to pick up the garbage bottle and put it into the garbage basket, and continue the inspection; S8, when the camera node fails to detect the garbage bottle, randomly generates a reachable target point; S9, the path planning node updates the map; Among them, the robot's micro host obtains a map of the working area through a laser radar, including: the laser radar node is responsible for publishing the obstacle information scanned by the radar to the path planning node, and publishing a scanning topic to build a map corresponding to the robot's surrounding environment; wherein the scanning topic contains an ID sequence that increases in scanning order, a timestamp of the laser data, a name of the scanning data, an angle at which the scan starts, an angle at which the scan ends, an angle increased for each scan, a measurement time interval, a scanning time interval, a minimum distance measurement, a maximum distance measurement, a distance of obstacles in a week, and light intensity; the path planning node also subscribes to the scanning topic for local planning in navigation and updating the map.

5. The control method of the garbage bottle picking robot according to claim 4, further comprising: The micro host communicates with the remote controller via a WIFI module or a 4G module electrically connected to the micro host, and the communication includes: receiving a remote control signal sent by the remote controller, or sending surrounding environment obstacle information, robot positioning information and robot field of view information to the remote controller.

6. According to the control method of the garbage bottle picking robot of claim 4, the step S3 comprises: The neural network chip constructs a neural network for training based on the feedback information from the laser radar and the binocular vision camera, and recognizes the shape and material of the garbage bottle.

7. A control system for a garbage bottle picking robot, designed based on the robot system framework ROS, to control the garbage bottle picking robot as claimed in any one of claims 1 to 3, comprising: A map service node, used for obtaining a map of a working area through the laser radar; A path planning node, used to automatically plan a route inspection in a working area through the micro host; A neural network node, used to control the laser radar and binocular vision camera to identify garbage bottles during inspection through the neural network chip; A GPS node, used to calculate the coordinates of the garbage bottle relative to the robot on the map through the GPS module and the laser radar; A motor control node, used for driving to the location of the garbage bottle; A camera node, used to measure the relative position information between the garbage bottle and the robotic arm through the binocular vision camera; A robot arm node, used to control the robot arm to pick up the trash bottle and put it into the trash basket through the micro host, and continue the inspection; The camera node is also used to randomly generate a reachable target point when no garbage bottle is detected; The path planning node is also used to update the map; A map service node, used to obtain a map of the working area through the laser radar, including: a map service node, used to publish the obstacle information scanned by the laser radar to the path planning node, and publish a scanning topic to build a map corresponding to the environment around the robot; wherein the scanning topic contains an ID sequence increasing in scanning order, a timestamp of the laser data, a name of the scanning data, an angle at which the scan starts, an angle at which the scan ends, an angle increased for each scan, a measurement time interval, a scanning time interval, a minimum value of the distance measured, a maximum value of the distance measured, a distance of obstacles in a week, and light intensity; the path planning node also subscribes to the scanning topic for local planning in navigation and updating the map.

8. The control system of the garbage bottle picking robot according to claim 7, further comprising: A communication node is used to communicate with a remote controller through the micro host via a WIFI module or a 4G module electrically connected to the micro host, and the communication includes: receiving a remote control signal sent by the remote controller, or sending surrounding environment obstacle information, robot positioning information and robot field of view information to the remote controller.