WO2019201346A1 - 自移动设备、服务器及其自动工作系统 - Google Patents

自移动设备、服务器及其自动工作系统 Download PDF

Info

Publication number
WO2019201346A1
WO2019201346A1 PCT/CN2019/083539 CN2019083539W WO2019201346A1 WO 2019201346 A1 WO2019201346 A1 WO 2019201346A1 CN 2019083539 W CN2019083539 W CN 2019083539W WO 2019201346 A1 WO2019201346 A1 WO 2019201346A1
Authority
WO
WIPO (PCT)
Prior art keywords
identification signal
module
self
mobile device
image
Prior art date
Application number
PCT/CN2019/083539
Other languages
English (en)
French (fr)
Inventor
何明明
Original Assignee
苏州宝时得电动工具有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 苏州宝时得电动工具有限公司 filed Critical 苏州宝时得电动工具有限公司
Priority to EP19788531.2A priority Critical patent/EP3783534A4/en
Priority to US17/048,566 priority patent/US20210157331A1/en
Priority to CN201980019807.0A priority patent/CN111868743A/zh
Publication of WO2019201346A1 publication Critical patent/WO2019201346A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
    • G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0219—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory ensuring the processing of the whole working surface
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0225—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving docking at a fixed facility, e.g. base station or loading bay
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/94—Hardware or software architectures specially adapted for image or video understanding

Definitions

  • the present invention relates to self-mobile devices, servers and their automated working systems.
  • Self-mobile devices are a type of robot that understands and judges the complex environment through sensors, and makes decisions and decisions on the complex environment, and implements goal-oriented movement to complete certain tasks. It can accept both user-entered instructions and automatically run according to the program being run. Self-mobile devices can be used indoors or outdoors, can be used in industry or home, can be used to replace security inspections, replace people to clean the ground, can also be used for family companion, auxiliary office, etc., such as sweeping robots, automatic lawn mowers and so on.
  • the self-moving device can recognize the object based on the image taken by the camera device disposed thereon, and then perform an action based on the recognized object, such as obstacle avoidance, along the line, and the like.
  • an action based on the recognized object such as obstacle avoidance, along the line, and the like.
  • any recognition technology cannot ensure that the percentage is correct.
  • mobile robots may malfunction due to misidentification, leak recognition, and the like.
  • the problem to be solved by the present invention is to provide an automated working system that accurately and timely identifies the working environment of the mobile device.
  • An automated working system comprising:
  • the self-mobile device includes:
  • An image detecting module detects the environment of the self-mobile device and generates an environment image
  • a first identification module configured to identify a specific object in the image based on the environment image, to generate a first identification signal
  • a first communication module communicatively coupled to the server
  • control module optionally controlling the first communication module to send the environment image and/or the first identification signal to the server;
  • the server includes:
  • a second identification module configured to identify a specific object in the image based on the environment image, to generate a second identification signal
  • a second communication module in communication with the first communication module, receiving the environment image and/or the first identification signal, and transmitting the second identification signal;
  • the control module controls an action of the self-mobile device based on the first identification signal and/or the second identification signal.
  • the controller controls the first communication module to send the environment image to the server, and receives the The second identification signal.
  • the preset condition includes generating the first identification signal within a first preset time.
  • the preset condition includes a confidence that the first identification signal is greater than a first preset value.
  • control module when the control module determines that the first identification signal does not meet the preset condition, the control module controls the self-mobile device to enter a safe working mode, and changes the action of the self-mobile device. .
  • the control module when the first communication module is connected to the second communication module, the control module sends the environment image and/or the first identification signal to the server, and receives the second identification. signal.
  • control module controls the movement mode of the self-mobile device based on the second identification signal.
  • control module controls the manner of movement of the self-mobile device based on the first identification signal.
  • the self-mobile device includes a battery module, and the control module controls the local communication module to disconnect from the server based on a power of the battery module being less than a preset power.
  • control module sends an authorization request to the user equipment, and receives an authorization signal of the user equipment, and controls the local communication module to send the detection signal and/or the locality based on the authorization signal. Identify the signal to the server.
  • the particular object includes an obstacle
  • the control module controls the self-mobile device to retreat or turn based on the obstacle.
  • the particular object includes a charging station
  • the control module controls the movement from the mobile device toward the charging station based on the charging station.
  • the specific object includes a work area boundary
  • the control module controls the self-mobile device to move within the work area or along the work area boundary based on the work area boundary.
  • the first identification module invokes a preset first depth learning model
  • the second recognition module invokes a preset second depth learning model, the number of model parameters of the second depth learning model Greater than the first recognition model
  • the server includes a software update module that generates an update data packet based on the environmental image and/or the first identification signal, the communication module transmitting the update data packet to the self-mobile device.
  • control module updates the first identification module based on the update data packet.
  • the first communication module comprises a fifth generation mobile communication module or a mobile communication module having a maximum transmission speed greater than 1 Gbps.
  • the environmental image comprises an original image or a processed image.
  • a self-mobile device that moves and works within the work area including:
  • An image detecting module detects the environment of the self-mobile device and generates an environment image
  • a first identification module configured to identify a specific object in the image based on the environment image, to generate a first identification signal
  • the first communication module selectively transmits the environment image and/or the first identification signal to the server and receives a second identification signal corresponding to the environment image;
  • control module configured to control an action of the self-mobile device according to the first identification signal and/or the second identification signal.
  • the controller controls the first communication module to send the environment image to the server, and receives the The second identification signal.
  • the preset condition includes generating the first identification signal within a first preset time.
  • the preset condition includes a confidence that the first identification signal is greater than a first preset value.
  • control module when the control module determines that the first identification signal does not meet the preset condition, the control module controls the self-mobile device to enter a safe working mode, and changes the action of the self-mobile device. .
  • control module transmits the environmental image and/or the first identification signal to the server and receives the second identification signal when the first communication module is operating normally.
  • control module controls the movement mode of the self-mobile device based on the second identification signal.
  • control module controls the manner of movement of the self-mobile device based on the first identification signal.
  • the particular object includes an obstacle
  • the control module controls the self-mobile device to retreat or turn based on the obstacle.
  • the specific object includes a charging station
  • the control module controls the self-mobile device to move toward the charging station based on the charging station.
  • the specific object includes a work area boundary
  • the control module controls the self-mobile device to move within the work area or along the work area boundary based on the work area boundary.
  • the first communication module comprises a fifth generation mobile communication module or a mobile communication module having a maximum transmission speed greater than 1 Gbps.
  • the environmental image comprises an original image or a processed image.
  • a server that includes:
  • a second communication module in communication with the self-mobile device, receiving the environment image sent by the mobile device
  • a second identification module configured to identify a specific object based on the environment image, to generate a second identification signal
  • the second communication module sends the second identification signal to the self-mobile device.
  • the server includes a software update module that generates an update data packet based on the environment image, the communication module transmitting the update data packet to the self-mobile device.
  • the beneficial effects of the present invention are: using a relatively simplified image recognition algorithm from the hardware and software of the mobile device itself to realize the recognition of basic features in the image; using the powerful hardware and software of the server to operate relative to The complex image recognition algorithm compensates for the defects of the image recognition algorithm from the mobile device itself; realizes low-latency data transmission through the high-speed communication module, enabling efficient data transmission between the mobile device and the server, ensuring that the mobile device can be timely Get accurate recognition results and act accordingly.
  • one of the objects of the present invention is to provide a mobile robot including a satellite signal receiving device to receive satellite signals, a communication module to receive information of a 5G mobile base station, and a mobile robot to utilize receiving satellite signals and
  • the signal transmitted by the parsed 5G mobile base station is used as a reference signal (differential correction number) to achieve accurate position fix.
  • Another object of the present invention is to provide a mobile robot comprising a satellite signal receiving device for receiving a satellite signal; a communication module real-time connecting with a 5G mobile base station group to acquire a reference signal, and the mobile robot utilizing the receiving satellite signal and the analyzed 5G mobile
  • the signal transmitted by the base station group is used as a reference signal (differential correction number) to achieve accurate position fix.
  • the present invention adopts the following scheme:
  • a mobile robot comprising: a processing unit, a satellite signal receiving device, and a communication module;
  • the processing unit is disposed inside the mobile robot, and is electrically connected to the satellite signal receiving device and the communication module;
  • the satellite signal receiving device is configured to receive a satellite signal
  • the communication module is connected to a 5G base station to receive information sent by the base station;
  • the processing unit parses the position reference signal according to the information transmitted by the received communication module and calculates the position coordinates of the mobile robot in combination with the received satellite signal.
  • the communication module is electrically connected to the processing unit.
  • the communication module is connected to the 5G base station to receive information transmitted by the 5G base station and fed back to the processing unit.
  • the communication module is integrated in the processing unit and electrically connected to the processing unit.
  • the satellite signal receiving device comprises an antenna for receiving satellite signals; and a data processing module disposed on the mobile robot, wherein the data processing module is electrically connected to the processing unit to receive and process the satellite signals.
  • the antenna is externally placed on the surface of the mobile robot or attached to the mobile robot housing.
  • the embodiment of the invention further provides a positioning system for a mobile robot, comprising the above mobile robot, a 5G base station;
  • the mobile robot calculates a position coordinate of the mobile robot based on the satellite signal received by the satellite signal receiving device and based on the information transmitted by the 5G base station received by the communication module and parsing the reference signal.
  • the embodiment of the invention further provides a positioning method for a positioning system of a mobile robot, comprising the above mobile robot, a 5G base station; the method comprises the following steps:
  • the information is exchanged with the base station through the communication module, and the information sent by the accepted base station is transmitted to the processing unit or the data processing module;
  • the processing unit or the data processing module parses the reference signal in the information based on the information transmitted by the base station;
  • the processing unit or the data processing module calculates the position coordinates of the mobile robot according to the signal of the satellite and the parsed reference signal.
  • the reference signal in step S3 is used as a differential correction number.
  • the processing unit of the mobile robot further controls the movement of the mobile robot according to the position coordinate.
  • the embodiment of the invention further provides a positioning method for a positioning system of a mobile robot, comprising the above mobile robot, further comprising a plurality of 5G base stations; the method, the following steps:
  • the data processing module parses the reference signal in the information based on the information transmitted by the data processing center.
  • the data processing module calculates the position coordinates of the mobile robot according to the satellite signal and the parsed reference signal.
  • the reference signal in step S13 is used as a differential correction number.
  • the processing unit of the mobile robot controls the movement of the mobile robot according to the position coordinate.
  • the mobile robot uses the combination of the 5G communication base station and the satellite signal receiving device to position and position during operation; specifically, the information transmitted by the 5G communication base station received in real time is used as the reference signal, and the real-time receiving from the satellite is received.
  • Signal, fast and accurate position and positioning; the implementation is simple, high positioning accuracy, cost saving, convenient installation, improved work efficiency, reduced manual interference, and improved user experience.
  • Figure 1 is a schematic illustration of an automated working system in accordance with one embodiment of the present invention.
  • FIG. 2 is a schematic structural view of a self-moving device according to an embodiment of the present invention.
  • FIG. 3 is a schematic illustration of a working system in accordance with one embodiment of the present invention.
  • FIG. 4 is a schematic structural diagram of a server according to an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of a server according to an embodiment of the present invention.
  • the automated work system includes self-mobile devices and servers.
  • the self-moving device is an automatic lawn mower 1.
  • the self-mobile device may also be an unattended device such as an automatic cleaning device, an automatic watering device, an automatic snow sweeper, and the like.
  • the automated working system also includes a server 10 that can communicate with the automatic mower 1 and provide functions such as storage, calculation, and the like.
  • the automated working system 100 also includes a charging station 2 for refueling the automatic mower 1.
  • the automatic lawn mower 1 includes a casing 3, and the moving module 5 is mounted on the casing 3.
  • the automatic mower 1 is driven to move;
  • the task execution module 7 is installed at the bottom of the casing 3, and includes a cutting assembly to perform mowing work.
  • the automatic mower 1 also includes an energy module that provides energy for the movement and operation of the automatic mower 1.
  • the energy module is the battery module 9.
  • the automatic mower 1 further includes a control module 11 electrically connected to the mobile module 5, the task execution module 7 and the energy module, and controls the mobile module 5 to drive the automatic mower 1 to move, and controls the task execution module 7 to perform work tasks.
  • the automatic lawn mower 1 further includes an image detecting module 13 that is mounted to the casing 3, detects an image in the vicinity of the automatic lawn mower 1, and outputs an environmental image.
  • the automatic mower 1 further includes a first identification module 15 that receives an environment image output by the image detection module 13, and the first recognition module 15 identifies a specific object in the image to generate a first identification signal.
  • the control module 11 controls the mobile module 5 based on the first identification signal to control the action of the self-mobile device 1 according to different working scenarios.
  • the automatic mower 1 also includes a first communication module 17 that is communicably coupled to the server 10.
  • FIG. 3 is a schematic illustration of a working system in accordance with one embodiment of the present invention.
  • the server is a cloud-based server.
  • the server 10 may also be a physical server such as a single server, a server cluster, or a distributed server.
  • the automatic mower 1 can communicate with the user equipment 20 through the first communication module 17 or other communication means to implement the second monitoring of the automatic lawn mower 1 by the user.
  • the server 10 can also communicate with the user device 20 to directly transfer the operation result of the server 10 to the user.
  • FIG. 4 is a schematic structural diagram of a server 10 according to an embodiment of the present invention.
  • the server 10 includes a second communication module 19, which is in communication with the first communication module 17, and receives an environment image or other signals sent by the first communication module 17.
  • the server 10 further includes a second identification module 21, the role of the second identification module 21 being substantially the same as that of the first identification module 15, generating a second identification signal based on the specific image in the environment image recognition image, through the second communication module 19 Send to automatic mower 1.
  • the difference is that the operation speed of the second identification module 21 is greater than that of the first identification module 15, and therefore, the second identification module 21 can invoke a more complex image recognition algorithm to provide a faster or more accurate recognition result.
  • the image detection module 13 includes an imaging device that performs related operations from the image captured by the camera device in accordance with the image captured by the camera device.
  • the camera device may be multiple.
  • the camera device may capture still images at different times at specific time intervals.
  • the camera device can capture video. Since the video is composed of image frames, the image frames in the acquired video can be acquired continuously or discontinuously and one frame image can be selected as one image.
  • the image detecting module 13 can directly output the original image.
  • the image detecting module 13 further includes a compression device or a trimming device, etc., and performs preliminary processing on the original image to output the processed image.
  • the first communication module 17 includes a 5G communication module.
  • the user bandwidth of the 5G design standard will reach 1Gbps (1024Mbps), which is 30 times that of current 4G, with more uniform data transmission rate, lower latency and lower unit cost.
  • the current state-of-the-art 4G LTE average latency is about 90 milliseconds
  • the average wired network is about 50-120 milliseconds
  • the 5G target delay is less than 10 milliseconds. Its design goal is about 1/50 of 4G, and it is much lower than the current one. Ordinary wired network.
  • the first communication module 17 includes a large-scale input/output antenna unit, and the number of input antennas of the large-scale input/output unit is greater than or equal to 2, and the number of output antennas is greater than or equal to 2.
  • the number of input antennas of the large-scale input/output unit is 4, and the number of output antennas is 4.
  • the base station 2 receives 2 transmissions and the 2 reception antenna devices support 2*2 MIMO.
  • the base station 4 sends 4 and receives 4 transmissions from the mobile device 4, and 4 data streams are concurrently, that is, 4 *4 MIMO, which doubles the rate compared to 2*2 MIMO.
  • the arrangement of the antennas in the large-scale input/output unit is an antenna array, which independently transmits and receives signals and ensures a sufficiently low correlation with each other, and each antenna has a completely isolated data stream, thereby realizing efficient manual data.
  • the first identification module 15 includes a first storage unit and a first operation unit.
  • the first storage unit stores a first depth learning model.
  • the first operation unit invokes the first A deep learning model performs an operation to output a first identification signal.
  • the success rate of image recognition is affected by two main factors: first, the design of the recognition algorithm model; and second, the size of the training set of the recognition algorithm (the number of pictures used to train the recognition model).
  • the complexity of the recognition algorithm model increases, such as an increase in the number of convolution layers and an increase in the number of neural network nodes, which contributes to the improvement of recognition accuracy.
  • the increased complexity of the algorithm model leads to an increase in computing power requirements.
  • the training set size in the first identification module 15 of the automatic mower 1 is relatively small.
  • the first depth model needs to be optimized, such as intentionally controlling the number of convolution layers and the number of neuron nodes. Since the amount of calculation required for image recognition is large, when the device does not have strong server support, the hardware and software resources of the automatic mower 1 are limited, and the first identification module 15 may have an identification error, a slow recognition speed, or even a system jam. Case.
  • the second identification module 21 receives the environment image, calls the second depth learning model stored in the second identification module 21 to perform an operation, and outputs a second identification signal. It is understood that the operation scale of the second depth learning model is larger than the first depth learning model, and the algorithm model complexity is greater than the first depth learning model.
  • the first identification module 15 includes a central processing unit (CPU), and further includes an image processing unit (GPU) or a digital signal processing unit (DSP) to provide sufficient floating point.
  • the first identification module 15 includes a central processing unit (CPU) and a dedicated neural network processing unit (NPU) that provides optimized computational support directly for neural network operations.
  • the first recognition model 15 includes a software program of an image recognition method based on a neural network, an image recognition method based on a wavelet moment, and the like, to process, analyze, and recognize the captured image.
  • the object area of the image corresponding to the object type.
  • the object area may be characterized by features such as gray scale of the object, contour of the object, and the like.
  • the manner in which the object area is represented by the outline of the object includes obtaining the identified object area by the contour extraction method.
  • the contour extraction method includes, but is not limited to, a binary, grayscale, canny operator, and the like.
  • the first identification module 15 includes a neural network model (such as CNN) obtained through pre-training, and identifies an object region corresponding to each object type from the image by executing a neural network model.
  • a neural network model such as CNN
  • the particular object identified by the first identification module 15 includes an obstacle.
  • the types of obstacles include, but are not limited to, charging stations, flower beds, trees, other garden tools, pets, and the like.
  • the control module 11 can perform obstacle avoidance according to the first identification signal, and specifically, can pause, retreat or turn.
  • the first identification module 15 can identify different types of obstacles, and the control module 11 performs different obstacle avoidance operations according to different types of obstacles, such as obstacles for flower beds, trees, etc., in order to be able to cut obstacles.
  • the control module 11 reduces the execution criteria of the obstacle avoidance operation, such as controlling only the deceleration of the mobile module 5.
  • the control module 11 can increase the execution standard of the obstacle avoidance operation, such as controlling the movement module 5 to retreat, and simultaneously controlling the task execution module 7 to stop working, ensuring that the pet and the human body are not damaged, and improving the safety of the automatic mower 1 operation. Sex.
  • the particular object identified by the first identification module 15 includes a work area boundary.
  • the working area of the automatic mower 1 is grassland, and the boundary of the working area is non-grass, and the working area is separated from the non-working area by grass and non-grass.
  • the first identification module 15 identifies the grass/non-grass in the image
  • the control module 11 controls the automatic mower 1 to move within the grass, detecting the occurrence of non-grass in the image or a certain position in the image.
  • the control module 11 controls the movement module 5 to perform a reverse or steering action.
  • the installation position of the image detecting module 13 determines the image recognized by the first identifying module 15, and accordingly affects the control mode of the control module 11. For example, the image detecting module 13 detects the ground in front of the automatic mower 1, and if the ground position detected by the image detecting module 13 is relatively close to the automatic mower 1, the control module 11 needs to respond to the first output of the first identifying module 15 more quickly.
  • the identification signal controls the movement module 5 to retreat or turn; if the ground position detected by the image detection module 13 is relatively far from the automatic lawn mower 1, the control module 11 can make a judgment after receiving the first identification signal, and control the movement module 5 to continue to Move forward or change the direction of movement.
  • the first identification module 15 identifies the grass/non-grass in the image
  • the control module 11 determines the position of the automatic mower 1 relative to the boundary of the work area by the first identification signal, and can control the movement module 5 to work along the way.
  • the area boundaries move so that the automatic mower 1 cuts to the boundary of the work area or returns along the boundary of the work area.
  • the control module 11 can control the automatic mower 1 to move along the inner boundary of the working area, or control the part of the automatic mower 1 to be located outside the working area in the working area, and can also control the automatic mower 1 to work along the working area.
  • the outer boundary of the area moves.
  • the particular object identified by the first identification module 15 includes the charging station 2.
  • the control module 11 controls the moving direction of the mobile module 5 according to the charging station 2 identified by the first identification module 15, so that the self-mobile device 1 can move toward the charging station 2, thereby realizing the self-mobile device 1 Return to the charging station 2.
  • the first identification module 15 is capable of not only identifying the charging station 2, but also identifying the marker on the charging station 2 or the docking terminal of the charging station 2.
  • the control module 11 controls the moving direction of the mobile module 5 according to the docking terminal of the charging station 2 identified by the first identifying module 15, so that the mobile device 1 is docked with the charging station 2.
  • the first communication module 17 and the second communication module 19 remain in normal communication under normal conditions.
  • the image detecting module 13 collects an image and outputs an environment image.
  • the first identifying module 15 identifies and outputs a first identification signal based on the environment image, and the control module 11 performs control based on the first identification signal.
  • a preset condition is set to the first identification module 15.
  • the control module 11 performs control based on the first identification signal, if the first identification signal does not satisfy the preset. Under the condition, the environmental image is transmitted to the server 10 and recognized by the second identification module 21.
  • the first identification signal output by the first identification module 15 is subjected to a confidence calculation. If the confidence is greater than the first preset value, the control module 11 performs control according to the first identification signal; if the confidence is less than the first a predetermined value, the control module 11 controls the first communication module 17 to send the environment image to the second communication module 19, and the second identification module 21 identifies and outputs the second identification signal, and is sent by the second communication module 19 to the first A communication module 17, the control module 11 performs control according to the second identification signal.
  • the computation time of the first identification module 15 is counted.
  • the control module 11 includes a timer.
  • the image detecting module 13 sends the environment image to the first identifying module 15, the timer starts timing. If the time of the timer has exceeded the first preset time, the image detecting module 13 still has no output.
  • the first identification signal the control module 11 controls the first communication module 17 to transmit the environmental image to the second communication module 19, and the second identification module 21 identifies and outputs the second identification signal, and is sent by the second communication module 19 to the first A communication module 17, the control module 11 performs control according to the second identification signal.
  • the first identification module 15 is prevented from being unresponsive due to hardware or software limitations.
  • the control module 11 controls the automatic mower 1 to enter the safe working mode.
  • the working mode of the automatic mower 1 can be set according to actual needs, and the moving module 5 can be controlled to reduce the moving speed or stop the movement, and the moving module 5 can be controlled to retreat or turn, or the task execution module 7 can be controlled. stop working.
  • the image detecting module 13 collects an image and outputs an environment image
  • the first identifying module 15 identifies and outputs a first identification signal based on the environment image
  • the first The communication module 17 transmits the environmental image to the second communication module 19 and receives the second identification signal output by the second identification module 21.
  • the control module 11 can preferentially control the automatic mower 1 by using the second identification signal, thereby improving the recognition accuracy of the environmental image. Since the automatic mower 1 and the server 10 are connected by wireless communication, there is a possibility of disconnection. In one embodiment, the control module 11 counts from the first communication module 17 transmitting the environmental image.
  • the control module 11 performs control based on the first identification signal.
  • the first identification signal output by the first identification module 15 is subjected to a confidence calculation. If the confidence is greater than the second predetermined value, the control module 11 performs control according to the first identification signal.
  • the receiving time of the second identification signal, the confidence of the first identification signal, the operation time of the first identification signal, and the like may be comprehensively considered, thereby On the basis of ensuring that the control module 11 can respond to the environmental image in time, the accuracy of the recognition is ensured.
  • the first communication module 17 when the amount of power of the battery module 9 is less than the preset amount of power, the first communication module 17 is controlled to stop operating, thereby reducing power consumption. Since the transmission rate of the first communication module 17 is high, the power consumption thereof is higher than that of the conventional communication module, and the first communication module 17 stops working, which can improve the success rate of the automatic lawn mower 1 returning to the charging station 3, and can also protect the battery module. 9, to avoid the exhaustion of the battery module 9 life.
  • some users may want to avoid sending image information to the server for privacy reasons. Based on this, before the environment image of the control module 11 is sent to the server 10, it is necessary to acquire the user authorization. If the environment image is authorized by the user, the control module 11 can control the first communication module 17 to send to the server 10. If the environment image is not authorized, the control module 11 cannot control the first communication module 17 to send to the server 10.
  • the first communication module 17 is in communication with the user equipment 20, and when the automatic lawn mower 1 starts working, the control module 11 sends an authorization request to the user equipment 20, and when the control module 11 receives the authorization issued by the user equipment 20. After the signal, the first communication module 17 can send an environmental image.
  • the authorization signal may include authorization for a specific time or a specific scenario, and the control module 11 transmits the environment image according to the authorization range of the authorization signal.
  • FIG. 5 is a schematic structural diagram of a server 10 according to an embodiment of the present invention.
  • the server 10 includes a software update module 23 that trains the software program of the first identification module based on the environment image received by the second communication module 19 and the first identification signal.
  • the training methods include, but are not limited to, adjusting internal parameters of the software program, configuration information of the software program, and the like.
  • the first identification module 15 invokes a deep learning model.
  • the deep learning training process requires massive data support and maintains high flexibility, while the server 10 has powerful computing resources and can effectively extract corresponding training parameters.
  • the software program includes a network structure and a connection mode of the neural network model, and the parameters in the neural network model are trained by a back propagation algorithm to improve the accuracy of the neural network model, and the software update module 23 includes The parameters in the trained neural network model are encapsulated in an update packet.
  • the update package may include a patch package applied to the software, a data package required to update the software, and the like.
  • the first software program or the second software program includes a network structure and a connection manner of the neural network model, and correspondingly, the update data packet includes parameters in the corresponding neural network.
  • the update data packet obtained after the training update of the first software program or the second software program includes related parameters such as weight parameters in the corresponding CNN. , offset parameters, etc.
  • the level of the first recognition algorithm model can be relatively objectively evaluated by comparing the first identification signal and the second identification signal corresponding to the same environmental image, particularly for an object, or for a particular scene. The accuracy of certain object recognition is low. Such comparison can be completed at the end of the automatic mower or at the server 10, and the server 10 can train the comparison result or remind the manufacturer to optimize, thereby perfecting the first recognition algorithm model.
  • the self-mobile device includes a processing unit, a positioning module, and a communication module.
  • the processing unit is disposed inside the mobile device and electrically connected to the positioning module and the communication module.
  • the positioning module includes a satellite signal receiving device. Receiving a satellite signal; the communication module is connected to the base station to receive information of the base station; the processing unit calculates the position coordinate of the mobile device according to the information transmitted by the received communication module and parses out the position reference signal and the received satellite signal (ie, Positioning from mobile devices).
  • the satellite signal receiving apparatus includes an antenna (preferably disposed externally from the mobile device), and the data processing module (which is disposed on the high-precision positioning board) is disposed in the self-mobile device, and is processed.
  • the unit is electrically connected.
  • the communication module is compatible with a mobile network (eg, 4G mobile network, 5G mobile network, etc.) and is electrically connected to the processing unit.
  • the communication module is configured to be pluggable from the mobile device, and is connected to a mobile 5G base station in the vicinity thereof during operation, and receives information transmitted by the base station and feeds back to the processing unit.
  • the communication module has built-in communication cards such as a standard sim card, a Micro sim card, and a Nano sim card.
  • the communication module is pluggably mounted to the self-mobile device.
  • the communication module is integrated in the processing unit, and no additional standard SIM card, Micro sim card, Nano sim card, etc., but a certain rule is set in the processing unit, thereby reducing the volume of the processing unit and improving stability.
  • the processing unit parses out the reference signal in the information according to the information received by the receiving communication module according to the agreed reading algorithm, and uses the reference signal as a differential signal, and then combines the signal transmitted by the satellite signal receiving device to calculate the self-mobile device ( Current) position coordinates (ie, positional positioning from the mobile device). Very good to meet the needs of mobile devices.
  • the base station is a 5G-based mobile base station.
  • the base station In addition to the function of a radio transceiver station for transmitting information between the mobile communication switching center and the mobile phone terminal, the base station also has a (GPS-based) positioning and timing function, and the base station usually also It includes a high-precision satellite antenna that uses dual-frequency mode to receive satellite information (distinguishing the traditional single-frequency mode).
  • the self-mobile device includes a body, a satellite signal, a receiving device, and a communication module.
  • a satellite signal receiving device comprising an antenna for receiving a satellite signal, the data processing module being disposed in the body to receive the satellite signal and processed;
  • the communication module is electrically connected to the data processing module to receive the information transmitted by the base station (5G base station) and transmit the information to the data processing module; the data processing module receives the information transmitted by the base station and parses the corrected number and calculates the satellite signal received by the antenna. From the mobile device location coordinates (ie current location information). The processing unit of the mobile device controls the movement from the mobile device according to the position coordinates.
  • the antenna of the satellite signal receiving device is pluggably mounted to the self-mobile device. In this way, the antenna can be removed during idle time to facilitate storage from the mobile device.
  • the data processing module is disposed on the high-precision positioning board and electrically connected to the processing unit, and the components of the processing unit are disposed through the layout and the circuit board.
  • the high precision positioning board is integrated in the processing unit, and the components of the processing unit are configured by the layout and the circuit board.
  • the self-mobile device has an external port for installing the communication module; if the communication module needs to be replaced, the user can purchase the matching module and install it by himself.
  • the data processing module (which is disposed on the high-precision positioning board) supports dual antenna input, and supports dual-band signals of BDS B1/B2, GPS L1/L2, and GLONASS G1/G2.
  • the data processing module supports the standard NMEA-0183GPGGA, GPGGARTK, GPGSV, GPGLL, GPGSA, GPGST, GHDHD, GPRMC, GPVTG, GPZDAetc; CMR (GPS) CMROBS, CMRREF, RTCM2.X RTCM1, RTCM3, RTCM9, RTCM1819, RTCM31, RTCM59; RTCM3.0 1004, 1005, 1006, 1007, 1008, 1011, 1012, 1104, 1033 RTCM 3.2 MSM4 & MSM5 1074, 1084, 1124, 1075, 1085, 1125 and other format data.
  • the description is made by taking a connection between a mobile device and a single 5G base station as an example.
  • the mobile device uses the satellite signal received by its own satellite signal receiving device and the communication module to connect with the base station and exchange information to obtain the location information of the mobile base station and calculate the position coordinate (current position information) from the mobile device as the differential correction number.
  • the position coordinate current position information
  • a signal network wider than the CORS coverage is a cellular communication base station network.
  • a plurality of 5G mobile base stations are configured in a certain area, and are composed of a series of cellular base stations, which divide the entire communication area into a single cell, and adopt a cellular wireless networking mode at the terminal. Connected to the network device via a wireless channel.
  • a plurality of mobile base stations are established in a certain area, for example, the number of the plurality of base stations is generally three or more, the first base station, the second base station, the third base station, and the second base station.
  • Data processing center the base station is separated from each other by a certain distance (for example, 50 to 1005 km), and the data processing center combines the network RTK algorithm to integrate the data of the entire base station network to perform calculations to simulate the virtual base station VRS ( ⁇ 1m) from the vicinity of the mobile device. And solve the more accurate reference signal (differential correction number) to achieve high-precision positioning from mobile devices.
  • the data processing center performs information interaction with the mobile device 0.
  • the transmission mode between the base station and the mobile device in the mobile base station group is based on a 5G mobile network.
  • the configuration of the self-mobile device is the same as that described in FIGS. 1 to 3.
  • the first base station, the second base station, the third base station, and the second base station (all are 5G base stations).
  • a positioning method of the self-mobile device includes the following steps:
  • the processing unit or the data processing module parses the reference signal in the information based on the information transmitted by the base station;
  • the processing unit or the data processing module calculates the position coordinates of the mobile device according to the signal of the satellite and the parsed reference signal.
  • step S1 further comprises performing noise reduction and filtering on the received satellite signal.
  • step S2 further includes the communication module interacting with the data processing center through the mobile network.
  • the step S3 data processing module reads the reference signal in the information based on the agreed read algorithm.
  • the step S3 data processing module reads the reference signal in the information based on the agreed read algorithm and uses the reference signal as a differential correction number.
  • step S4 the processing unit that is further included in the self-mobile device controls the movement from the mobile device according to the position coordinate.
  • the satellite signal may be a navigation signal such as a GPS signal, a Beidou navigation signal, a European Galileo signal, or a Russian Glonass signal.
  • the communication module exchanges information with the 5G mobile base station. The information is transmitted to a processing unit or a data processing module, and the processing unit or the data processing module reads the position information in the information according to an agreed algorithm and uses it as a differential signal.
  • the processing unit of the mobile device controls the movement from the mobile device according to the position coordinates.
  • a positioning method from a mobile device in which a positioning method using a plurality of mobile base station groups to connect from a mobile device, includes the following steps:
  • the data processing module parses the reference signal in the information based on the information transmitted by the data processing center.
  • the data processing module calculates the position coordinates of the mobile device according to the satellite signal and the parsed reference signal.
  • step S11 further comprises performing noise reduction and filtering on the received satellite signal.
  • step S12 further includes the communication module interacting with the data processing center through the mobile network.
  • the data processing module reads the reference signal in the information based on the agreed reading algorithm in step S13.
  • the data processing module reads the reference signal in the information based on the agreed reading algorithm in step S13 and uses the reference signal as a differential correction number.
  • step S14 the processing unit that is further included in the self-mobile device controls the movement from the mobile device according to the position coordinate.
  • the self-mobile device is internally configured with a communication module, and the communication module is integrated with a 5G communication module (the module is adopted, and is compatible with 3G, 4G, etc.); the communication module is working.
  • the signal transmitted by the receiving base station parses the position information in the signal as a reference signal.
  • satellite signal receiving device for a self-mobile device, "satellite signal receiving device", “communication module”, “5G mobile base station”, “5G mobile base station group” may be included therein only.
  • the content of the "satellite signal receiving device” may be selected from one or a combination thereof including the related technical features in the embodiment, wherein the content regarding the "communication module” may be selected from the embodiment included.
  • One or a combination of the related technical features wherein the content regarding the "mobile base station” may be selected from one or a combination thereof including the related technical features in the embodiment, wherein regarding the "5G mobile base station group”
  • the content may be selected from one or a combination of the related technical features of the embodiments.
  • At least one battery pack is carried from the mobile device. It can be implemented in a self-mobile device. For example, a battery pack is used, and the battery pack is placed as far as possible from the center of gravity of the mobile device to improve stability during work. Two battery packs (which can be electrically connected in series or electrically connected in parallel) are used. The battery pack (the tool is projected from the top to the ground) is placed as far as possible from the central area of the mobile device to improve stability during operation.
  • the satellite signal receiving device includes an antenna for receiving the satellite signal, and the data processing module (which is disposed on the high precision positioning board) is disposed in the self-mobile device to receive the receiving of the antenna.
  • the signal, as well as the reference signal transmitted in conjunction with the communication module, is located from the current position coordinates of the mobile device.
  • the antenna is externally mounted on the mobile device in a manner that is externally attached to the mobile device.
  • the surface of the self-moving device housing is attached in the form of a patch.
  • the communication module is compatible with the mobile network (eg, 4G network, 5G network), and connects to the nearby base station to receive the information transmitted by the base station and feed back to the processing unit.
  • the communication module has a built-in communication card such as a standard sim card, a Micro sim card, or a Nano sim card.
  • it is installed in a pluggable manner from a mobile device or integrated in a processing unit.
  • the 5G base station in addition to the functions of the radio transceiver station for transmitting information between the mobile communication switching center and the mobile telephone terminal, the 5G base station also has a (GPS-based) positioning and timing function, in which case the base station usually includes High-precision satellite antenna that uses dual-frequency mode to receive satellite information (distinguishing the traditional single-frequency mode).
  • a certain area contains a plurality of 5G mobile base stations, and the base station network constitutes a continuous operation reference station.
  • the MAX voltage of the battery pack can be 12V, 16V, 20V, 24V, 40V, 60V, etc., and the specific voltage can be seen from the application of the mobile device, which is not limited herein.
  • the battery chip inside the battery pack can be selected from a lithium-based battery, a fuel battery, and the like.
  • the self-moving device may be a lawn mower, a sweeping machine, a snow sweeper, etc.; preferably, the lawn mower, the sweeping machine, and the snow sweeper also have an autonomous path planning function.

Abstract

本发明涉及一种自移动设备，在工作区域内移动和工作，包括：图像检测模块，检测自移动设备的环境，生成环境图像；第一识别模块，基于环境图像识别图像中的特定对象，生成第一识别信号；第一通信模块，可选择地发送环境图像和/或第一识别信号至服务器并接收与环境图像相对应的第二识别信号；控制模块，根据第一识别信号和/或第二识别信号控制自移动设备的动作。本发明的有益效果是：保证自移动设备能够及时地获取准确的识别结果，并据此进行动作。

Description

自移动设备、服务器及其自动工作系统 技术领域
本发明涉及自移动设备、服务器及其自动工作系统。
背景技术
自移动设备是一类通过传感器感知周围环境和自身的状态，对复杂环境进行理解和判断，在此基础上进行决策和规划，实现面向目标的移动，从而完成一定工作任务的机器人。它既可以接受用户输入的指令运行，又可以根据所运行的程序自动运行。自移动设备可用在室内或室外，可用于工业或家庭，可用于取代保安巡视、取代人们清洁地面，还可用于家庭陪伴、辅助办公等，如扫地机器人、自动割草机等等。
自移动设备可以基于设置于其上的摄像装置所摄取的图像来识别物体，进而基于所识别的物体进行动作，如避障、沿行等。然而，在实际应用中，任何识别技术不能确保百分正确，面对复杂的真实环境，移动机器人可能因错误识别、漏识别等出现错误动作。
发明内容
为克服现有技术的缺陷，本发明所要解决的问题是提供一种准确且及时地识别自移动设备工作环境的自动工作系统。
本发明解决现有技术问题所采用的技术方案是：
一种自动工作系统，包括：
自移动设备，在工作区域内移动和工作，
服务器，与所述自移动设备通信；
所述自移动设备包括：
图像检测模块，检测所述自移动设备的环境，生成环境图像；
第一识别模块，基于所述环境图像识别所述图像中的特定对象，生成第一识别信号；
第一通信模块，与所述服务器通信连接；
控制模块，可选择地控制所述第一通信模块发送所述环境图像和/或第一识别信号至所述 服务器；
所述服务器包括：
第二识别模块，基于所述环境图像识别所述图像中的特定对象，生成第二识别信号；
第二通信模块，与所述第一通信模块通信连接，接收所述环境图像和/或第一识别信号，发送所述第二识别信号；
所述控制模块基于所述第一识别信号和/或所述第二识别信号控制所述自移动设备的动作。
在其中一个实施例中，所述控制模块判断所述第一识别信号不满足预设条件时，所述控制器控制所述第一通信模块发送所述环境图像至所述服务器，并接收所述第二识别信号。
在其中一个实施例中，所述预设条件包括第一预设时间内生成所述第一识别信号。
在其中一个实施例中，所述预设条件包括所述第一识别信号的置信度大于第一预设值。
在其中一个实施例中，所述控制模块判断所述第一识别信号不满足所述预设条件时，所述控制模块控制所述自移动设备进入安全工作模式，改变所述自移动设备的动作。
在其中一个实施例中，当所述第一通信模块与所述第二通信模块连接时，控制模块发送所述环境图像和/或第一识别信号至所述服务器，并接收所述第二识别信号。
在其中一个实施例中，若第二预设时间内接收到所述第二识别信号，所述控制模块基于所述第二识别信号控制所述自移动设备的移动方式。
在其中一个实施例中，若所述第一识别信号的置信度大于第二预设值，所述控制模块基于所述第一识别信号控制所述自移动设备的移动方式。
在其中一个实施例中，所述自移动设备包括电池模块，控制模块基于所述电池模块的电量小于预设电量，控制所述本地通信模块与所述服务器断开连接。
在其中一个实施例中，所述控制模块发送授权请求至用户设备，并接收所述用户设备的授权信号，基于所述授权信号控制所述本地通信模块发送所述检测信号和/或所述本地识别信号至所述服务器。
在其中一个实施例中，所述特定对象包括障碍物，所述控制模块基于所述障碍物控制所述自移动设备后退或转向。
在其中一个实施例中，所述特定对象包括充电站，所述控制模块基于所述充电站控制所 述自移动设备朝向所述充电站移动。
在其中一个实施例中，所述特定对象包括工作区域边界，所述控制模块基于所述工作区域边界控制所述自移动设备在所述工作区域内移动或沿所述工作区域边界移动。
在其中一个实施例中，所述第一识别模块调用预先设置的第一深度学习模型，所述第二识别模块调用预先设置的第二深度学习模型，所述第二深度学习模型的模型参数数量大于所述第一识别模型。
在其中一个实施例中，所述服务器包括软件更新模块，基于所述环境图像和/或第一识别信号生成更新数据包，所述通信模块将所述更新数据包发送至所述自移动设备。
在其中一个实施例中，所述控制模块基于所述更新数据包更新所述第一识别模块。
在其中一个实施例中，所述第一通信模块包括第五代移动通信模块或最大传输速度大于1Gbps的移动通信模块。
在其中一个实施例中，所述环境图像包括原始图像或经过处理的图像。
本发明解决现有技术问题所采用的一种技术方案是：
一种自移动设备，在工作区域内移动和工作，包括：
图像检测模块，检测所述自移动设备的环境，生成环境图像；
第一识别模块，基于所述环境图像识别所述图像中的特定对象，生成第一识别信号；
第一通信模块，可选择地发送所述环境图像和/或第一识别信号至服务器并接收与所述环境图像相对应的第二识别信号；
控制模块，根据所述第一识别信号和/或第二识别信号控制所述自移动设备的动作。
在其中一个实施例中，所述控制模块判断所述第一识别信号不满足预设条件时，所述控制器控制所述第一通信模块发送所述环境图像至所述服务器，并接收所述第二识别信号。
在其中一个实施例中，所述预设条件包括第一预设时间内生成所述第一识别信号。
在其中一个实施例中，所述预设条件包括所述第一识别信号的置信度大于第一预设值。
在其中一个实施例中，所述控制模块判断所述第一识别信号不满足所述预设条件时，所述控制模块控制所述自移动设备进入安全工作模式，改变所述自移动设备的动作。
在其中一个实施例中，当所述第一通信模块正常工作时，控制模块发送所述环境图像和/或第一识别信号至所述服务器，并接收所述第二识别信号。
在其中一个实施例中，若第二预设时间内接收到所述第二识别信号，所述控制模块基于所述第二识别信号控制所述自移动设备的移动方式。
在其中一个实施例中，若所述第一识别信号的置信度大于第二预设值，所述控制模块基于所述第一识别信号控制所述自移动设备的移动方式。
在其中一个实施例中，所述特定对象包括障碍物，所述控制模块基于所述障碍物控制所述自移动设备后退或转向。
在其中一个实施例中，所述特定对象包括充电站，所述控制模块基于所述充电站控制所述自移动设备朝向所述充电站移动。
在其中一个实施例中，所述特定对象包括工作区域边界，所述控制模块基于所述工作区域边界控制所述自移动设备在所述工作区域内移动或沿所述工作区域边界移动。
在其中一个实施例中，所述第一通信模块包括第五代移动通信模块或最大传输速度大于1Gbps的移动通信模块。
在其中一个实施例中，所述环境图像包括原始图像或经过处理的图像。
本发明解决现有技术问题所采用的一种技术方案是：
一种服务器，包括：
第二通信模块，与自移动设备通信连接，接收所述自移动设备发送的环境图像；
第二识别模块，基于所述环境图像识别特定对象，生成第二识别信号；
所述第二通信模块将所述第二识别信号发送至所述自移动设备。
在其中一个实施例中，所述服务器包括软件更新模块，基于所述环境图像生成更新数据包，所述通信模块将所述更新数据包发送至所述自移动设备。
与现有技术相比，本发明的有益效果是：利用自移动设备本身的硬件和软件运行相对简化的图像识别算法，实现对图像中的基本特征的识别；利用服务器强大的硬件和软件运行相对复杂的图像识别算法，弥补自移动设备本身图像识别算法的缺陷；通过高速的通信模块实现地低延时的数据传输，使得自移动设备和服务器之间实现高效数据传输，保证自移动设备能够及时地获取准确的识别结果，并据此进行动作。
为克服现有技术的缺陷，本发明的目的之一在于：提出一种移动机器人，其包含卫星信号接收装置以接收卫星信号；通讯模块以接收5G移动基站的信息；移动机器人利用接收卫 星信号以及解析的5G移动基站发送的信号作为参考信号(差分改正数)以实现精确的位置定位。
本发明的另一目的在于，提出一种移动机器人，其包含卫星信号接收装置以接收卫星信号；通讯模块实时的与5G移动基站群连接获取参考信号，移动机器人利用接收卫星信号以及解析的5G移动基站群发送的信号作为参考信号(差分改正数)以实现精确的位置定位。
为实现上述目的，本发明采用如下方案：
一种移动机器人，其特征在于，包含，处理单元，卫星信号接收装置，通讯模块；
所述处理单元配置于移动机器人的内部，其电性连接所述卫星信号接收装置及所述通讯模块；
所述卫星信号接收装置，用以接收卫星信号；
所述通讯模块，与5G基站连接，以接收所述基站发送的信息；
其中，所述处理单元，依据接收的通讯模块传输的信息解析出位置参考信号并结合接收的所述卫星信号计算出移动机器人的位置坐标。
优选的，该通讯模块与处理单元电性相连，运行时所述通讯模块与所述5G基站连接接收所述5G基站传输的信息并反馈至所述处理单元。
优选的，该通讯模块集成于所述处理单元并与所述处理单元电性相连。
优选的，该卫星信号接收装置包含天线，用以接收卫星信号；配置于移动机器人的数据处理模块，所述数据处理模块与所述处理单元电性连接接收并处理所述卫星信号。
优选的，该天线，外置于所述移动机器人或贴附于移动机器人壳体的表面。
本发明实施例还提出一种移动机器人的定位系统，包含上述的移动机器人，5G基站；
所述移动机器人基于所述卫星信号接收装置接收的卫星信号以及基于所述通讯模块接收的所述5G基站传输的信息并解析出参考信号计算出移动机器人的位置坐标。
本发明实施例还提出一种移动机器人的定位系统的定位方法，包含上述移动机器人，5G基站；所述方法包括如下步骤：
S1，通过卫星信号接收装置接收卫星信号，并将接收信号传输至处理单元或数据处理模块；
S2，通过通讯模块与基站连接进行信息交互，并将接受的基站发送的信息传输至处理单 元或数据处理模块；
S3，处理单元或数据处理模块基于基站传输的信息解析出信息中的参考信号；
S4，处理单元或数据处理模块根据卫星的信号和解析的参考信号计算出移动机器人的位置坐标。
优选的，步骤S3中所述参考信号作为差分改正数。
优选的，步骤S4后，还包括，所述移动机器人的处理单元依据该位置坐标控制移动机器人移动。
本发明实施例还提出一种移动机器人的定位系统的定位方法，包含上述移动机器人，还包含多个5G基站；所述方法，如下步骤：
S11，通过定位模块接收装置接收卫星信号并将接收信号传输至数据处理模块；
S12，通过通讯模块接收数据处理中心传输信息并将接受的信息传输至数据处理模块；
S13，数据处理模块基于数据处理中心传输的信息解析出信息中的参考信号；
S14，数据处理模块根据的卫星信号和解析的参考信号计算出移动机器人的位置坐标。
优选的，步骤S13中所述参考信号作为差分改正数。
优选的，步骤S14后还包括，所述移动机器人的处理单元依据该位置坐标控制移动机器人移动。
与现有技术相比，本发明的有益效果：
移动机器人在运行时，利用5G通讯基站与卫星信号接收装置的组合来位置定位；具体的，利用实时接收的5G通讯基站传输的信息并以此位置信息作为参考信号，与实时的接收来自卫星的信号，进行快速准确的位置定位；该实施方案简单，定位精度高，节省成本，安装方便，提高工作效率，降低了人工干涉，提高了用户体验。
附图说明
以上所述的本发明的目的、技术方案以及有益效果可以通过下面附图实现：
图1是本发明的一个实施例的自动工作系统示意图。
图2是本发明的一个实施例的自移动设备结构示意图。
图3是本发明的一个实施例的工作系统示意图。
图4是本发明的一个实施例的服务器结构示意图。
图5是本发明的一个实施例的服务器结构示意图。
具体实施方式
图1为本发明的一个实施例的自动工作系统100示意图。如图1所示，自动工作系统包括自移动设备和服务器。本实施例中，自移动设备为自动割草机1，在其他实施例中，自移动设备也可以为自动清洁设备、自动浇灌设备、自动扫雪机等适合无人值守的设备。自动工作系统还包括服务器10，可与自动割草机1进行通信，并提供存储、计算等功能。自动工作系统100还包括充电站2，用于为自动割草机1补给电能。
图2为本发明的一个实施例的自动割草机1的结构示意图，如图2所示，本实施例中，自动割草机1包括壳体3；移动模块5，安装于壳体3，带动自动割草机1移动；任务执行模块7，安装于壳体3底部，包括切割组件，执行割草工作。自动割草机1还包括能源模块，为自动割草机1的移动和工作提供能量。本实施例中，能源模块为电池模块9。自动割草机1还包括控制模块11，与移动模块5、任务执行模块7以及能源模块电连接，控制移动模块5带动自动割草机1移动，并控制任务执行模块7执行工作任务。自动割草机1还包括图像检测模块13，安装于壳体3，检测自动割草机1附近区域内的图像，输出环境图像。自动割草机1还包括第一识别模块15，接收图像检测模块13输出的环境图像，第一识别模块15识别图像中的特定对象，生成第一识别信号。控制模块11基于第一识别信号控制移动模块5，根据不同的工作场景控制自移动设备1的动作。自动割草机1还包括第一通信模块17，可与服务器10通信连接。
图3为本发明的一个实施例的工作系统示意图。如图3所示，本实施例中，服务器为基于云架构的服务器。在其他实施例中，服务器10也可以是单台服务器、服务器集群、分布式服务器等物理服务器。自动割草机1可以通过第一通信模块17或其他通信方式与用户设备20进行通信，以实现用户对自动割草机1的第二监控。服务器10也可以与用户设备20进行通信，将服务器10的运算结果直接传输给用户。
图4为本发明的一个实施例的服务器10的结构示意图。如图4所示，本实施例中，服务器10包括第二通信模块19，与第一通信模块17通信连接，接收第一通信模块17发送的 环境图像或其他信号。服务器10还包括第二识别模块21，第二识别模块21的作用与第一识别模块15的作用基本相同，基于环境图像识别图像中的特定对象，生成第二识别信号，通过第二通信模块19发送至自动割草机1。区别在于，第二识别模块21的运算速度大于第一识别模块15，因此，第二识别模块21可调用更为复杂的图像识别算法，提供速度更快或精度更高的识别结果。
在一个实施例中，图像检测模块13包括一个摄像装置，自移动设备1依据该摄像装置所摄取的图像执行相关操作，在其他实施例中，摄像装置可以为多个。在其他实施例中，摄像装置可以以特定时间间隔拍摄的不同时刻下的静态图像。在其他实施例中，摄像装置可以拍摄视频，由于视频是由图像帧构成的，因而可以通过连续或不连续地采集所获取视频中的图像帧并选用一帧图像作为一幅图像。本实施例中，图像检测模块13可直接输出原始图像。在其他实施例中，图像检测模块13还包括压缩装置或剪裁装置等，对原始图像进行初步处理，输出处理后的图像。
在一个实施例中，第一通信模块17包括5G通信模块。相对于4G技术而言，5G设计标准中用户带宽最高将达到1Gbps(1024Mbps)，是目前4G的30倍，拥有更统一的数据传输速率、更低的延迟和更低的单位成本。目前最先进的4G LTE平均延迟大概在90毫秒,普通的有线网络约为50-120毫秒，而5G的目标延迟是10毫秒以内，其设计目标约为4G的1/50，并且远低于目前的普通有线网络。
在一个实施例中，第一通信模块17包括大规模输入输出天线单元，大规模输入输出单元的输入天线数量大于等于2，输出天线数量大于等于2。在其中一个实施例中，大规模输入输出单元的输入天线数量为4，输出天线数量为4。一般的，2收2发的基站与2收2发的天线设备支持2*2 MIMO，本实施例中，基站4发4收自移动设备4发4收，有4个数据流并发，即4*4 MIMO，与2*2 MIMO相比速率翻倍。大规模输入输出单元中天线的布置形式为天线阵列，独立地收发信号并保证彼此间足够低的相关性，且每副天线有完全隔离开来的数据流，从而实现高效手法数据。
在一个实施例中，第一识别模块15包括第一存储单元和第一运算单元，第一存储单元存储有第一深度学习模型，基于图像检测模块13输出的环境图像，第一运算单元调用第一深度学习模型进行运算，输出第一识别信号。图像识别的成功率受到两个主要因素的影响：其 一，识别算法模型的设计；其二：识别算法训练集的规模(用以训练识别模型的图片的数量)。识别算法模型的复杂度增加，如卷积层数增加、神经网络节点数增加，有助于识别准确度的提升。但是，算法模型复杂度提升会导致对运算能力要求的提升。训练集规模越大，训练效果则越好，训练完毕的识别算法的成功率越高。为了提升成功率，实际工程应用中要求训练集越大越好。但是，考虑到成本、时间，自动割草机1的第一识别模块15中训练集规模相对较小。为了支撑深度算法的需求，第一深度模型需要进行优化，比如有意控制卷积层数、神经元节点数。由于图像识别需要的计算量大，当设备没有强大的服务器支撑时，自动割草机1自带的硬件和软件资源有限，第一识别模块15可能会出现识别错误、识别速度慢甚至系统卡死的情况。本实施例中，第二识别模块21接收环境图像，调用存储在第二识别模块21内的第二深度学习模型进行运算，输出第二识别信号。以理解的是，第二深度学习模型的运算规模大于第一深度学习模型，算法模型复杂度大于第一深度学习模型。
在一个实施例中，为了保证足够的识别成功率，第一识别模块15包括中央处理单元(CPU)，还包括图像处理单元(GPU)或数字信号处理单元(DSP)，以提供足够的浮点运算能力。在其他实施例中，第一识别模块15包括中央处理单元(CPU)，还包括专用的神经网络处理单元(NPU)，直接为神经网络运算提供优化的运算支持。
在一个实施方式中，第一识别模型15中包括基于神经网络的图像识别算法、基于小波矩的图像识别算法等图像识别方法的软件程序，以对所摄取的图像进行处理、分析和识别并得到图像中对应物体种类的对象区域。其中，所述对象区域可由物体的灰度、物体的轮廓等特征表征。例如，所述对象区域由物体的轮廓表示的方式包括通过轮廓线提取方法获得所识别对象区域。所述轮廓线提取方法包括但不限于：二值、灰度、canny算子等方法。然后，针对预先标记有物体种类标签的物体与图像中的对象区域，通过对物体和图像的内容、特征、结构、关系、纹理及灰度等的对应关系，相似性和一致性的分析来寻求相似图像目标，以使得图像中的对象区域与预先标记的物体种类相对应。在一种实施方式中，第一识别模块15包含经由预先训练而得的神经网络模型(如CNN)，并藉由执行神经网络模型从图像识别出对应各物体种类的对象区域。
在一个实施例中，第一识别模块15识别的特定对象包括障碍物。对于自动割草机1而言，障碍物的种类包括但不限于：充电站、花坛、树木、其他花园工具、宠物等。在自动割 草机1正常工作过程中，当第一识别模块15识别到障碍物，控制模块11可以根据第一识别信号进行避障，具体的，可以暂停、后退或转向。在一种实施方式中，第一识别模块15可以识别不同障碍物的种类，控制模块11根据障碍物的不同种类执行不同的避障操作，如对于花坛、树木等障碍物，为了能够切割到障碍物边缘，控制模块11减少避障操作的执行标准，比如仅控制移动模块5减速。对于宠物和人体，控制模块11可以增加避障操作的执行标准，比如控制移动模块5后退，同时控制任务执行模块7停止工作，确保宠物和人体不受伤害，提高自动割草机1操作的安全性。
在一个实施例中，第一识别模块15识别的特定对象包括工作区域边界。自动割草机1的工作区域为草地，工作区域边界为非草地，通过草地和非草地将工作区域与非工作区域区分开。
在一个实施例中，第一识别模块15识别图像中的草地/非草地，控制模块11控制自动割草机1在草地内移动，检测到图像中出现非草地或图像中的某个位置出现非草地时，控制模块11控制移动模块5执行后退或转向动作。具体的，图像检测模块13的安装位置决定了第一识别模块15识别的图像，相应地影响了控制模块11的控制方式。如，图像检测模块13检测自动割草机1前方的地面，若图像检测模块13检测的地面位置相对靠近自动割草机1，控制模块11需要更快地响应第一识别模块15输出的第一识别信号，控制移动模块5后退或转向；若图像检测模块13检测的地面位置相对远离自动割草机1，控制模块11可以在接收到第一识别信号后做出判断，控制移动模块5继续向前移动或者改变移动方向。
在一个实施例中，第一识别模块15识别图像中的草地/非草地，控制模块11通过第一识别信号判断自动割草机1相对于工作区域边界的位置，可以控制移动模块5沿着工作区域边界移动，使得自动割草机1切割到工作区域边界，或者沿着工作区域边界回归。具体的，控制模块11可以控制自动割草机1沿着工作区域内边界移动，或者控制自动割草机1部分位于工作区域内部分位于工作区域外，也可以控制自动割草机1沿着工作区域外边界移动。
在一个实施例中，第一识别模块15识别的特定对象包括充电站2。当自动割草机1在回归模式中，控制模块11根据第一识别模块15识别的充电站2控制移动模块5的移动方向，使得自移动设备1能够向着充电站2移动，从而实现自移动设备1回归充电站2。在其他实施例中，第一识别模块15不仅能够识别充电站2，还能够识别充电站2上的标志物或充电站 2的对接端子。控制模块11根据第一识别模块15识别的充电站2的对接端子，控制移动模块5的移动方向，使得自移动设备1与充电站2对接。
在一个实施例中，正常情况下，第一通信模块17与第二通信模块19保持正常通信。自动割草机1在工作过程中，图像检测模块13采集图像并输出环境图像，第一识别模块15基于环境图像进行识别并输出第一识别信号，控制模块11基于第一识别信号进行控制。但是，由于第一识别模块15的运算能力和识别精度有限，可能会出现识别错误较多或识别速度过慢的情况。对于某些物体，由于视角、环境、光线、形变等因素的差异，可能会出现无法成功识别的情况。因此，本实施例中，对第一识别模块15设置一个预设条件，若第一识别信号满足预设条件，则控制模块11基于第一识别信号进行控制，若第一识别信号不满足预设条件，则将环境图像发送至服务器10，由第二识别模块21进行识别。
在一个实施例中，对第一识别模块15输出的第一识别信号进行置信度计算，若置信度大于第一预设值，则控制模块11根据第一识别信号进行控制；若置信度小于第一预设值，则控制模块11控制第一通信模块17将环境图像发送至第二通信模块19，由第二识别模块21识别并输出第二识别信号，并由第二通信模块19发送至第一通信模块17，控制模块11根据第二识别信号进行控制。
在一个实施例中，对第一识别模块15的运算时间进行统计。控制模块11包括定时器，当图像检测模块13将环境图像发送至第一识别模块15，则定时器开始定时，若定时器的时间已经超过第一预设时间，而图像检测模块13仍然没有输出第一识别信号，则控制模块11控制第一通信模块17将环境图像发送至第二通信模块19，由第二识别模块21识别并输出第二识别信号，并由第二通信模块19发送至第一通信模块17，控制模块11根据第二识别信号进行控制。从而避免第一识别模块15由于硬件或软件限制造成响应不及时。
在一个实施例中，为了避免自动割草机1由于没有及时识别到目标物体而发生危险，若第一识别信号不满足预设条件，控制模块11控制自动割草机1进入安全工作模式。在安全工作模式中，自动割草机1的工作方式可以根据实际需要进行设置，可以控制移动模块5降低移动速度或停止移动，也可以控制移动模块5后退或转向，也可以控制任务执行模块7停止工作。
在一个实施例中，自动割草机1在工作过程中，图像检测模块13采集图像并输出环境 图像，第一识别模块15基于环境图像进行识别并输出第一识别信号，与此同时，第一通信模块17将环境图像发送至第二通信模块19，并接收第二识别模块21输出的第二识别信号。控制模块11可优先选用第二识别信号控制自动割草机1，从而提高对环境图像的识别准确度。由于自动割草机1与服务器10通过无线通信的方式连接，因此存在断开连接的可能性。在一个实施例中，控制模块11从第一通信模块17发送环境图像起计时，若在第二预设时间内没有接收到第二识别信号，则控制模块11基于第一识别信号进行控制。在其他实施例中，对第一识别模块15输出的第一识别信号进行置信度计算，若置信度大于第二预设值，则控制模块11根据第一识别信号进行控制。在其他实施例中，根据自动割草机1的硬件能力和软件复杂程度，可综合考量第二识别信号的接收时间、第一识别信号的置信度、第一识别信号的运算时间等参数，从而在保证控制模块11能够及时对环境图像作出响应的基础上，保证识别的准确性。
在一个实施例中，当电池模块9的电量小于预设电量时，控制第一通信模块17停止工作，从而减小电量消耗。由于第一通信模块17的传输速率高，其功耗比传统的通信模块高，第一通信模块17停止工作能够提高自动割草机1回到充电站3充电的成功率，也能够保护电池模块9，避免电量耗尽影响电池模块9的寿命。
在一个实施例中，如图3所示，在一些场景下，一些用户可能出于隐私原因而想要避免向服务器发送图像信息。基于此，控制模块11环境图像发送至服务器10之前，需要获取用户授权。若环境图像经过用户授权，则控制模块11可以控制第一通信模块17发送至服务器10，若环境图像未经授权范围，则控制模块11不能控制第一通信模块17发送至服务器10。本实施例中，第一通信模块17与用户设备20通信连接，在自动割草机1开始工作时，控制模块11向用户设备20发送授权请求，当控制模块11接收到用户设备20发出的授权信号后，第一通信模块17才能发送环境图像。在其他实施例中，授权信号中可以包括针对特定时间或特定场景下的授权，控制模块11根据授权信号的授权范围发送环境图像。
图5为本发明的一个实施例的服务器10结构示意图。如图5所示，在一个实施例中，服务器10包括软件更新模块23，基于第二通信模块19接收的环境图像、第一识别信号对第一识别模块的软件程序进行训练。训练的方式包括但不限于调整软件程序的内部参数、软件程序的配置信息等。本实施例中，第一识别模块15调用深度学习模型，深度学习训练过程需 要海量数据支撑并保持较高灵活性，而服务器10计算资源强大，能够有效地提取出相应的训练参数。例如，所述软件程序中包括神经网络模型的网络结构及连接方式，利用通过反向传播算法训练所述神经网络模型中的参数，以提高神经网络模型的正确率，则软件更新模块23将包含训练后的神经网络模型中的参数封装在更新数据包中。
更新数据包可以包括应用于软件的补丁包、更新软件所需的数据包等。例如，在一个实施例中，第一软件程序或第二软件程序中包括神经网络模型的网络结构及连接方式，对应地，更新数据包中包含相应神经网络中的参数。例如，在第一软件程序或第二软件程序采用CNN执行的情况下，对第一软件程序或第二软件程序进行训练更新后获得的更新数据包中包括对应的CNN中的相关参数如权重参数、偏置参数等。
在一个实施例中，通过比对对应于同一环境图像的第一识别信号和第二识别信号，可以相对客观地评价第一识别算法模型的水平，特别是针对某种物体、或特定场景下的某种物体识别准确率偏低。这样的比对可以在自动割草机1端完成，也可以在服务器10端完成，服务器10可针对比对结果进行训练，也可提醒厂家进行优化，从而完善第一识别算法模型。
在一个实施例中，自移动设备包含，处理单元，定位模块，通讯模块；该处理单元配置于自移动设备内部并电性连接定位模块及通讯模块；该定位模块包含有卫星信号接收装置，以接收卫星信号；该通讯模块与基站连接，以接收基站的信息；该处理单元依据接收的通讯模块传输的信息并解析出位置参考信号以及结合接收的卫星信号计算出自移动设备的位置坐标(即，对自移动设备进行位置定位)。
上述实施方式中，卫星信号接收装置包含，天线(较佳的采用外置的方式配置于自移动设备)，数据处理模块(其设置于高精度定位板卡)配置于自移动设备内，与处理单元电性连接。
通讯模块兼容移动网络(如，4G移动网络，5G移动网络等)，与处理单元电性相连。较佳的，通讯模块采用可插拔的配置于自移动设备，运行时与其附近的移动5G基站连接，接收基站传输的信息并反馈至处理单元。通讯模块内置有标准sim卡、Micro sim卡、Nano sim卡等可通讯卡。在一实施方式中，通讯模块采用可插拔的方式安装于自移动设备。在一实施方式中，通讯模块集成于处理单元，这时无须额外标准sim卡、Micro sim卡、Nano sim卡等而是通过一定的规则设置于处理单元，这样可减小处理单元的体积，提高稳定性。
处理单元依据接收通讯模块传递的信息，按照约定的读取算法，解析出信息中的参考信号，并将此参考信号作为差分信号，再结合卫星信号接收装置传输的信号，计算出自移动设备的(当前)位置坐标(即，对自移动设备位置定位)。很好的满足自移动设备的需求。
基站为基于5G的移动基站，该基站除了通过移动通信交换中心与移动电话终端之间进行信息传递的无线电收发信电台的功能外，还具有(基于GPS)定位及授时功能，这时基站通常还包括高精度卫星天线，该天线采用双频模式接收卫星信息(区别传统的单频模式)。
在一个实施例中，自移动设备包含，本体，卫星信号，接收装置，通讯模块。
卫星信号接收装置，包含天线用于接收卫星信号，数据处理模块配置于本体内，以接收卫星信号并处理；
通讯模块电性连接数据处理模块，以接收基站(5G基站)传输的信息并将该信息传输至数据处理模块；数据处理模块接收基站传输的信息并解析出改正数并结合天线接收的卫星信号计算出自移动设备位置坐标(即当前位置信息)。自移动设备的处理单元依据该位置坐标控制自移动设备移动。
在一实施方式中，卫星信号接收装置的天线采用可插拔的方式安装于自移动设备。这样在闲时可将天线取下，以方便自移动设备的存放。
在一实施方式中，数据处理模块设置于高精度定位板卡与处理单元电性连接，处理单元的元器件通过布局配置与电路板。
在一实施方式中，高精度定位板卡集成于处理单元中，处理单元的元器件通过布局配置与电路板。
在一个实施方式中，自移动设备具有外置端口，用以安装通讯模块；若需更换通讯模块，用户可自行购买匹配的模块自行安装。
在一实施方式中，数据处理模块(其设置于高精度定位板卡)支持双天线输入，支持BDS B1/B2、GPS L1/L2、GLONASS G1/G2三系统双频信号。数据交互时，其支持标准NMEA-0183GPGGA,GPGGARTK,GPGSV,GPGLL,GPGSA,GPGST,GPHDT,GPRMC,GPVTG,GPZDAetc；CMR(GPS)CMROBS,CMRREF，RTCM2.X RTCM1,RTCM3,RTCM9,RTCM1819,RTCM31,RTCM59；RTCM3.0 1004,1005,1006,1007,1008,1011,1012,1104,1033RTCM3.2 MSM4&MSM5 1074,1084,1124,1075,1085,1125等格式数据。
以自移动设备与单个5G基站连接为例进行描述。自移动设备利用自身的卫星信号接收装置接收的卫星信号以及通讯模块与基站连接并进行信息交互获取该移动基站的位置信息并将该消息作为差分改正数计算出自移动设备位置坐标(当前位置信息)。这样本实施方案，无需额外自建基站作参照站，极大的降低了系统成本。
为了解决上述问题，人们提出了比CORS覆盖范围更广泛的信号网络是蜂窝通讯基站网络，
在实际的应用中，在某一区域内配置有多个5G移动基站，由一系列的蜂窝基站构成，这些蜂窝基站把整个通信区域划分成一个个蜂窝小区，采用蜂窝无线组网方式，在终端和网络设备之间通过无线通道连接起来。
在一个实施例中，在某区域内建立多个移动基站，例如多个基站的数量一般3个或3个以上，第一基准站，第二基准站，第三基准站，第二基准站，数据处理中心；基准站相互距离一定的距离(如，50～1005公里)，数据处理中心结合网络RTK算法整合全基准站网络的数据进行运算可以在自移动设备附近模拟出虚拟基站VRS(<1m)并解算出更准确的参考信号(差分修正数)，实现自移动设备的高精度定位。本实施方式中，数据处理中心与自移动设备0进行信息交互。移动基站群中的基准站到自移动设备间的传输方式采用基于5G移动网络。本实施方式中，自移动设备的配置同图1-图3中描述的方案。第一基准站，第二基准站，第三基准站，第二基准站(皆是5G基站)。
接下来描述一种自移动设备的定位方法，以自移动设备与单个5G移动基站连接的定位方法，包括如下步骤：
S1，通过定位模块接收卫星信号并将该卫星信号传输至处理单元或数据处理模块；
S2，通过通讯模块与5G基站连接接受基站发送的信息传输至处理单元或数据处理模块；
S3，处理单元或数据处理模块基于基站传输的信息解析出信息中的参考信号；
S4，处理单元或数据处理模块根据卫星的信号和解析的参考信号计算出自移动设备的位置坐标。
在一实施方式中，步骤S1还包含对接收的卫星信号进行降噪，滤波。
在一实施方式中，步骤S2还包含通讯模块与数据处理中心通过移动网络进行信息交互。
在一实施方式中，步骤S3数据处理模块基于约定的读取算法读取信息中的参考信号。
在一实施方式中，步骤S3数据处理模块基于约定的读取算法读取信息中的参考信号并将该参考信号作为差分改正数。
在一实施方式中，步骤S4后，还包含自移动设备的处理单元依据该位置坐标控制自移动设备移动。
上述的实施方式中，卫星信号可以为GPS信号、北斗导航信号、欧洲的Galileo信号、俄罗斯的Glonass信号等导航定位信号。通讯模块与5G移动基站进行信息交互。该信息传输至处理单元或数据处理模块，处理单元或数据处理模块依据约定的算法读取出信息中的位置信息并将其作为差分信号。自移动设备的处理单元依据该位置坐标控制自移动设备移动。
接下来描述一自移动设备的定位方法，以自移动设备利用多个移动基站群连接的定位方法，包括如下步骤：
S11，通过定位模块接收装置接收卫星信号并将接收信号传输至数据处理模块；
S12，通过通讯模块接收数据处理中心传输信息并将接受的信息传输至数据处理模块；
S13，数据处理模块基于数据处理中心传输的信息解析出信息中的参考信号；
S14，数据处理模块根据的卫星信号和解析的参考信号计算出自移动设备的位置坐标。
在一实施方式中，步骤S11还包含对接收的卫星信号进行降噪，滤波。
在一实施方式中，步骤S12还包含通讯模块与数据处理中心通过移动网络进行信息交互。
在一实施方式中，步骤S13中数据处理模块基于约定的读取算法读取信息中的参考信号。
在一实施方式中，步骤S13中数据处理模块基于约定的读取算法读取信息中的参考信号并将该参考信号作为差分改正数。
在一实施方式中，步骤S14后，还包含自移动设备的处理单元依据该位置坐标控制自移动设备移动。
在一个实施方式中，该自移动设备的内部配置有通讯模块，该通讯模块集成有5G通讯模块(该模块采用，采用向下兼容的方式，以兼容3G，4G等)；工作时该通讯模块与基站建立连接，进行信息交互。接收基站传递的信号解析出信号中的位置信息以此作为参考信号。
需要提醒注意的是，在一实施方式中，对于一自移动设备而言，在“卫星信号接收装置”、“通讯模块”、“5G移动基站”、“5G移动基站群”、可以只包括其中的一个或多个技术特征。其中关于“卫星信号接收装置”的内容系可选自于包含有实施方式中其相关技术特征的其中之 一或其组合，其中关于“通讯模块”的内容系可选自于包含有实施方式中其相关技术特征的其中之一或其组合，其中关于“移动基站”的内容系可选自于包含有实施方式中其相关技术特征的其中之一或其组合，其中关于“5G移动基站群”的内容系可选自于包含有实施方式中其相关技术特征的其中之一或其组合。
在自移动设备的设计中，自移动设备搭载至少一个电池包。其可采用内置于自移动设备的方案。如，采用1个电池包，该电池包尽量设置于自移动设备的重心位置，以提高工作过程中的稳定性。采用2个电池包(其可采用电性串联或电性并联)，该电池包(工具从上投影至地面看)尽量设置于自移动设备的中央区域，以提高工作过程中的稳定性。
在卫星信号接收装置的设计中，卫星信号接收装置，包含天线，用于接收卫星信号，数据处理模块(其设置在高精度定位板卡)配置于自移动设备内，以接收该天线的接收的信号，以及结合通讯模块传输的参考信号，定位自移动设备当前的位置坐标。
天线采用外置于自移动设备的方式，固定在自移动设备。在卫星天线的设计中，采用贴片的形式贴附于自移动设备壳体的表面。
在通讯模块的设计中，通讯模块兼容移动网络(如，4G网络，5G网络)，运行时与其附近的基站连接接收基站传输的信息并反馈至处理单元。较佳的，通讯模块，内置有标准sim卡、Micro sim卡、Nano sim卡等可通信卡。较佳的，采用可插拔的方式安装于自移动设备或者集成于处理单元中。
在移动基站的设计中，5G基站除了通过移动通信交换中心，与移动电话终端之间进行信息传递的无线电收发信电台的功能外，还具有(基于GPS)定位及授时功能，这时基站通常包括高精度卫星天线，该天线采用双频模式接收卫星信息(区别传统的单频模式)。
在移动基站群的设计中，某一区域内的包含多个5G移动基站，该基站组网构成连续运行参考站。
在电池包的设计中，电池包的MAX电压可为12V，16V，20V，24V，40V，60V等，具体的电压可视自移动设备的应用场合，在此不作限定。电池包内部的电池芯片可选用锂基电池，燃料电池等。
上述的实施方式中，自移动设备可为割草机，清扫机，扫雪机等；较佳的，割草机，清扫机，扫雪机其还具有自主路径规划功能。
也就是说，可以将上述特征作任意的排列组合，并用于自移动设备的改进。
以上描述的实施方式仅表达了本发明的几种实施方式，由于文字表达得有限性，而在客观上存在无限的可能结构，对于本领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以作出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。因此，本发明专利的保护范围应以所附权利要求为准。

Claims (22)

  1. 一种自动工作系统，包括：
    自移动设备，在工作区域内移动和工作，
    服务器，与所述自移动设备通信；
    其特征在于：
    所述自移动设备包括：
    图像检测模块，检测所述自移动设备的环境，生成环境图像；
    第一识别模块，基于所述环境图像识别所述图像中的特定对象，生成第一识别信号；
    第一通信模块，与所述服务器通信连接；
    控制模块，可选择地控制所述第一通信模块发送所述环境图像和/或所述第一识别信号至所述服务器；
    所述服务器包括：
    第二识别模块，基于所述环境图像识别所述图像中的特定对象，生成第二识别信号；
    第二通信模块，与所述第一通信模块通信连接，接收所述环境图像和/或第一识别信号，发送所述第二识别信号；
    所述控制模块基于所述第一识别信号和/或所述第二识别信号控制所述自移动设备的动作。
  2. 根据权利要求1所述的自动工作系统，其特征在于，所述控制模块判断所述第一识别信号不满足预设条件时，所述控制器控制所述第一通信模块发送所述环境图像至所述服务器，并接收所述第二识别信号。
  3. 根据权利要求2所述的自动工作系统，其特征在于，所述预设条件包括第一预设时间内生成所述第一识别信号或所述第一识别信号的置信度大于第一预设值。
  4. 根据权利要求2所述的自动工作系统，其特征在于，所述控制模块判断所述第一识别信号不满足所述预设条件时，所述控制模块控制所述自移动设备进入安全工作模式，改变所述自移动设备的动作。
  5. 根据权利要求1所述的自动工作系统，其特征在于，当所述第一通信模块与所述第二通信模块连接时，控制模块发送所述环境图像和/或第一识别信号至所述服务器，并接收所述 第二识别信号。
  6. 根据权利要求5所述的自动工作系统，其特征在于，若第二预设时间内接收到所述第二识别信号，所述控制模块基于所述第二识别信号控制所述自移动设备的移动方式。
  7. 根据权利要求5所述的自动工作系统，其特征在于，若所述第一识别信号的置信度大于第二预设值，所述控制模块基于所述第一识别信号控制所述自移动设备的移动方式。
  8. 根据权利要求1所述的自动工作系统，其特征在于，所述第一识别模块调用预先设置的第一深度学习模型，所述第二识别模块调用预先设置的第二深度学习模型，所述第二深度学习模型的模型参数数量大于所述第一识别模型。
  9. 根据权利要求8所述的自动工作系统，其特征在于，所述服务器包括软件更新模块，基于所述环境图像和/或第一识别信号生成更新数据包，所述通信模块将所述更新数据包发送至所述自移动设备；所述控制模块基于所述更新数据包更新所述第一识别模块。
  10. 根据权利要求1所述的自动工作系统，其特征在于，所述第一通信模块包括第五代移动通信模块或最大传输速度大于1Gbps的移动通信模块。
  11. 根据权利要求1所述的自动工作系统，其特征在于，所述环境图像包括原始图像或经过处理的图像。
  12. 一种自移动设备，在工作区域内移动和工作，其特征在于，包括：
    图像检测模块，检测所述自移动设备的环境，生成环境图像；
    第一识别模块，基于所述环境图像识别所述图像中的特定对象，生成第一识别信号；
    第一通信模块，可选择地发送所述环境图像和/或所述第一识别信号至服务器并接收与所述环境图像相对应的第二识别信号；
    控制模块，基于所述第一识别信号和/或第二识别信号控制所述自移动设备的动作。
  13. 根据权利要求12所述的自移动设备，其特征在于，所述控制模块判断所述第一识别信号不满足预设条件时，所述控制器控制所述第一通信模块发送所述环境图像至所述服务器，并接收所述第二识别信号。
  14. 根据权利要求13所述的自移动设备，其特征在于，所述预设条件包括第一预设时间内生成所述第一识别信号或所述第一识别信号的置信度大于第一预设值。
  15. 根据权利要求13所述的自移动设备，其特征在于，所述控制模块判断所述第一识别 信号不满足所述预设条件时，所述控制模块控制所述自移动设备进入安全工作模式，改变所述自移动设备的动作。
  16. 根据权利要求12所述的自移动设备，其特征在于，当所述第一通信模块正常工作时，控制模块发送所述环境图像和/或第一识别信号至所述服务器，并接收所述第二识别信号。
  17. 根据权利要求16所述的自移动设备，其特征在于，若第二预设时间内接收到所述第二识别信号，所述控制模块基于所述第二识别信号控制所述自移动设备的移动方式。
  18. 根据权利要求16所述的自移动设备，其特征在于，若所述第一识别信号的置信度大于第二预设值，所述控制模块基于所述第一识别信号控制所述自移动设备的移动方式。
  19. 根据权利要求1所述的自移动设备，其特征在于，所述第一通信模块包括第五代移动通信模块或最大传输速度大于1Gbps的移动通信模块。
  20. 根据权利要求1所述的自移动设备，其特征在于，所述环境图像包括原始图像或经过处理的图像。
  21. 一种服务器，其特征在于，包括：
    第二通信模块，与自移动设备通信连接，接收所述自移动设备发送的环境图像；
    第二识别模块，基于所述环境图像识别特定对象，生成第二识别信号；
    所述第二通信模块将所述第二识别信号发送至所述自移动设备。
  22. 根据权利要求21所述的服务器，其特征在于，所述服务器包括软件更新模块，基于所述环境图像生成更新数据包，所述通信模块将所述更新数据包发送至所述自移动设备。
PCT/CN2019/083539 2018-04-19 2019-04-19 自移动设备、服务器及其自动工作系统 WO2019201346A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP19788531.2A EP3783534A4 (en) 2018-04-19 2019-04-19 SELF-MOVING DEVICE, SERVER AND AUTOMATIC WORKING SYSTEM FOR IT
US17/048,566 US20210157331A1 (en) 2018-04-19 2019-04-19 Self-moving device, server, and automatic working system thereof
CN201980019807.0A CN111868743A (zh) 2018-04-19 2019-04-19 自移动设备、服务器及其自动工作系统

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810354642.X 2018-04-19
CN201810354642 2018-04-19

Publications (1)

Publication Number Publication Date
WO2019201346A1 true WO2019201346A1 (zh) 2019-10-24

Family

ID=68240568

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/083539 WO2019201346A1 (zh) 2018-04-19 2019-04-19 自移动设备、服务器及其自动工作系统

Country Status (4)

Country Link
US (1) US20210157331A1 (zh)
EP (1) EP3783534A4 (zh)
CN (1) CN111868743A (zh)
WO (1) WO2019201346A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110989689A (zh) * 2019-12-16 2020-04-10 兰州理工大学 一种基于自动追踪的机器人系统

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190332848A1 (en) * 2018-04-27 2019-10-31 Honeywell International Inc. Facial enrollment and recognition system
USD988372S1 (en) * 2020-03-24 2023-06-06 Husqvarna Ab Reference station for robotic mower
CN115685997A (zh) * 2022-10-14 2023-02-03 深圳市正浩创新科技股份有限公司 自移动设备的沿边控制方法、装置、介质及自移动设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1381339A (zh) * 2001-04-18 2002-11-27 三星光州电子株式会社 使用移动通信网络的机器人清洁系统
CN104575489A (zh) * 2014-12-03 2015-04-29 嘉兴市德宝威微电子有限公司 机器人识别系统
CN106407993A (zh) * 2016-09-13 2017-02-15 成都创慧科达科技有限公司 一种基于图像识别技术的智能语音机器人系统及方法
CN106940562A (zh) * 2017-03-09 2017-07-11 华南理工大学 一种移动机器人无线集群系统及神经网络视觉导航方法

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10115185B2 (en) * 2014-12-05 2018-10-30 At&T Intellectual Property I, L.P. Dynamic image recognition model updates
CN105095919A (zh) * 2015-09-08 2015-11-25 北京百度网讯科技有限公司 图像识别方法和装置
WO2017177444A1 (zh) * 2016-04-15 2017-10-19 深圳前海达闼云端智能科技有限公司 一种云机器人系统、机器人及机器人云平台
US10575696B2 (en) * 2016-07-13 2020-03-03 Irobot Corporation Autonomous robot auto-docking and energy management systems and methods
US10025984B2 (en) * 2016-07-18 2018-07-17 X Development Llc Delegation of object and pose detection
US10261513B2 (en) * 2016-12-19 2019-04-16 drive.ai Inc. Methods for communicating state, intent, and context of an autonomous vehicle
US10671925B2 (en) * 2016-12-28 2020-06-02 Intel Corporation Cloud-assisted perceptual computing analytics

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1381339A (zh) * 2001-04-18 2002-11-27 三星光州电子株式会社 使用移动通信网络的机器人清洁系统
CN104575489A (zh) * 2014-12-03 2015-04-29 嘉兴市德宝威微电子有限公司 机器人识别系统
CN106407993A (zh) * 2016-09-13 2017-02-15 成都创慧科达科技有限公司 一种基于图像识别技术的智能语音机器人系统及方法
CN106940562A (zh) * 2017-03-09 2017-07-11 华南理工大学 一种移动机器人无线集群系统及神经网络视觉导航方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110989689A (zh) * 2019-12-16 2020-04-10 兰州理工大学 一种基于自动追踪的机器人系统

Also Published As

Publication number Publication date
EP3783534A4 (en) 2022-01-05
CN111868743A (zh) 2020-10-30
US20210157331A1 (en) 2021-05-27
EP3783534A1 (en) 2021-02-24

Similar Documents

Publication Publication Date Title
WO2019201346A1 (zh) 自移动设备、服务器及其自动工作系统
CN109258601B (zh) 一种智能化的除草无人机自动对靶喷雾流量控制装置与方法
JP2022523836A (ja) 農薬散布の制御方法、デバイス及び記憶媒体
US20190049984A1 (en) Automatic Working System and Control Method Thereof
US20200042009A1 (en) Unmanned lawn mower with autonomous driving
CN116931602A (zh) 高安全性的变电站巡检无人机系统和方法
WO2019137136A1 (zh) 植保无人机作业质量检测装置及其检测方法
CN107390699B (zh) 一种甘蔗种植机的路线规划系统及其路线规划方法
US10219422B2 (en) Machine-to-machine sharing of wayline deviation information
WO2016097891A1 (en) Robotic vehicle for detecting gps shadow zones
CN113115621B (zh) 智能割草系统及其自主建图方法
US11882787B1 (en) Automatic sensitivity adjustment for an autonomous mower
WO2020206946A1 (zh) 自动收割机的倒伏行驶规划系统及其方法
CN107069859A (zh) 一种基于无人机基站的无线充电系统及方法
CN106808608A (zh) 一种智能切割机
CN109001756A (zh) 基于嵌入式设备的多线激光雷达障碍物检测系统及方法
US20200356093A1 (en) Management apparatus, management system, moving object, and computer readable storage medium
CN114044139A (zh) 一种无人机及基于无人机的智慧农场管理方法
CN110162102A (zh) 基于云平台和机器视觉的无人机自动识别跟踪方法及系统
CN108445877A (zh) 基于移动3g/4g网络环境下的仿生运动控制系统与控制方法
US20230042867A1 (en) Autonomous electric mower system and related methods
WO2021139413A1 (zh) 一种智能割草机
CN113358100B (zh) 嵌入式与yolo4改进算法的无人机实时目标识别系统
CN115047870A (zh) 一种玉米籽粒收获机无人驾驶系统及其运行方法
CN208795220U (zh) 一种近红外植保测绘无人机系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19788531

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2019788531

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2019788531

Country of ref document: EP

Effective date: 20201119