WO2022161245A1 - 一种提高机器人关节转矩检测精度的方法及多关节机器人 - Google Patents

一种提高机器人关节转矩检测精度的方法及多关节机器人 Download PDF

Info

Publication number
WO2022161245A1
WO2022161245A1 PCT/CN2022/072969 CN2022072969W WO2022161245A1 WO 2022161245 A1 WO2022161245 A1 WO 2022161245A1 CN 2022072969 W CN2022072969 W CN 2022072969W WO 2022161245 A1 WO2022161245 A1 WO 2022161245A1
Authority
WO
WIPO (PCT)
Prior art keywords
joint
information
torque
robot
torque sensor
Prior art date
Application number
PCT/CN2022/072969
Other languages
English (en)
French (fr)
Inventor
张小川
孙恺
Original Assignee
苏州艾利特机器人有限公司
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
Priority claimed from CN202110126502.9A external-priority patent/CN114800465B/zh
Application filed by 苏州艾利特机器人有限公司 filed Critical 苏州艾利特机器人有限公司
Publication of WO2022161245A1 publication Critical patent/WO2022161245A1/zh

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls

Definitions

  • the invention relates to the field of industrial robots, in particular to a method for improving the detection accuracy of robot joint torque and a multi-joint robot.
  • Industrial robots are a class of robots that are widely used in industrial environments. Industrial robots can usually be divided into traditional industrial robots and collaborative robots. Collaborative robots are a new type of industrial robot, which is small in size and easy to install, and can be used in commercial , service industry, industry and other scenarios. The diverse application scenarios of collaborative robots put forward higher requirements for safety and precision.
  • the precision of the end of the industrial robot needs to be good enough to ensure the effect of the work.
  • some robot manufacturers have begun to try to install torque sensors in the joints of industrial robots to improve the accuracy of the robot end.
  • most robot manufacturers have their own robot body design.
  • the optimization of the robot mainly lies in the optimization of hardware and software.
  • the robot itself has a relatively stable structure.
  • To add a torque sensor to the existing robot structure it is necessary to adjust the robot body. Major changes have been made to the design.
  • the joint itself may be affected by external force, and the torque sensor is coupled, which affects the detection accuracy of the torque sensor, which in turn affects the force control accuracy of the industrial robot, making the end accuracy of the industrial robot insufficient to meet the accuracy requirements.
  • the existing technology is improved by changing the mechanical structure. This method is difficult to design, complicated to implement, and also increases the weight and volume of the joint. At the same time, the increased mechanical structure increases the robot The resulting manufacturing cost is not conducive to widespread use.
  • the purpose of the present invention is to provide a method and a multi-joint robot for improving the accuracy of robot joint torque detection, which is easy to implement and has good accuracy.
  • the present invention can adopt the following technical solutions: a method for improving the detection accuracy of robot joint torque, the robot includes a plurality of joints, and each joint includes a torque sensor, and the method includes: preset information step: preset the torque of each joint of the robot The calibration information of the sensor and the establishment of the robot mechanical model, the calibration information represents the coupling effect of the joint force on the joint torque information of the torque sensor; the obtaining step: obtaining the joint torque information of each joint; the calculating step: according to the robot mechanics Model and joint torque information, calculate the joint force information of each joint of the robot; decoupling step: decoupling the joint torque information of the torque sensor according to the calibration information and the joint force information, and output decoupling After the joint torque information.
  • the obtaining step includes: obtaining joint torque information of each joint through a torque sensor; and/or, detecting the current of each joint, and obtaining joint torque information of each joint accordingly.
  • the method includes: repeating the calculation step and the decoupling step, iteratively processing the joint torque information, and outputting the final joint torque information.
  • the method includes: when the number of iterations is greater than or equal to a preset number of times, outputting final joint torque information.
  • the method includes: outputting final joint torque information when the difference between the joint torque information output in the previous iteration period and the joint torque information output in the current iteration period is less than or equal to a preset difference.
  • the torque sensor includes an input connection flange and an output connection flange, and the torque sensor is used as an output connection flange of the robot joint.
  • the torque sensor is a torque sensor.
  • a multi-joint robot includes several joints and connecting rods, the joints include a motor, a reducer, and a torque sensor, and the torque sensor includes an input flange connection plate and an output connection connection plate
  • the multi-joint robot includes: a preset information unit for preset calibration information of each joint of the robot and establishing a mechanical model of the robot, the calibration information representing the coupling influence of the joint force on the joint torque information of the torque sensor; obtaining a unit for acquiring joint torque information of each joint; a computing unit, electrically connected to the preset information unit and the acquiring unit, for acquiring joint force information of each joint according to the mechanical model and the joint torque information;
  • the decoupling unit is electrically connected to the torque sensor and the calculation unit, and is used for decoupling the joint torque information of the torque sensor according to the calibration information and the joint force information, and outputting the decoupled joint torque information .
  • the acquisition unit is a torque sensor, and/or the acquisition unit includes a current detection unit, and the joint torque information of each joint is acquired according to the current detection unit.
  • the multi-joint robot is a six-joint collaborative robot.
  • the beneficial effects of the specific embodiment of the present invention are: by presetting calibration information and establishing a robot mechanical model, according to the detection of the torque sensors of each joint of the robot, the influence of external force coupling of the joints is removed, and the joint torque information is improved. accuracy. Further, by iteratively processing the joint torque information, the accuracy of the joint torque information is further improved. In this way, the influence of external force coupling of the joints is removed, and the structural design of the robot itself does not need to be changed, which is easy to implement and has a low cost.
  • FIG. 1 is a schematic diagram of a method for improving the detection accuracy of robot joint torque according to an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a method for improving the detection accuracy of robot joint torque according to another embodiment of the present invention
  • FIG. 3 is a schematic diagram of a method for improving the detection accuracy of machine joint torque according to another embodiment of the present invention
  • FIG. 4 is a block diagram of a multi-joint robot according to an embodiment of the present invention
  • FIG. 5 is a schematic diagram of a multi-joint robot according to an embodiment of the present invention
  • FIG. 6 is a schematic diagram of a joint of a multi-joint robot according to an embodiment of the present invention
  • the present invention protects a method for improving the detection accuracy of robot joint torque.
  • the robot includes a plurality of joints, and each joint includes a torque sensor.
  • the method includes: S1 preset information step, preset torque of each joint of the robot
  • the calibration information of the sensor and the establishment of the mechanical model of the robot, the predetermined information represents the coupling effect of the joint force on the joint torque information detected by the torque sensor;
  • the acquisition step S2 is to acquire the joint torque information of each joint;
  • the calculation step S3, according to the The robot mechanical model and joint torque information are described, and the joint force information of each joint of the robot is estimated;
  • the decoupling step S4 is to decouple the joint torque information of the torque sensor according to the calibration information and the joint force information.
  • the calibration information of the torque sensor of the joint represents the coupling influence of the force and torque of the robot in various directions.
  • the calibration information represents the influence of the force in each direction on the accuracy of the joint torque information output by the torque sensor of the joint.
  • the influence of the external force on different robots is different.
  • the coupling effect of the force and torque in each direction on the joint torque information detected by the torque sensor is confirmed.
  • the joint force information of each joint can be obtained according to the mechanical model of the robot and joint torque information.
  • various methods of establishing a mechanical model can be used in this scheme, as long as the purpose of establishing a mechanical model of the robot can be achieved.
  • the joint torque information can also be acquired in various ways.
  • the magnitude of the external force of each joint is confirmed, the influence of the external force coupling of each joint is determined by comparing the calibration information, and the joint torque information is decoupled, and the accuracy of the joint torque information after decoupling is improved. promote.
  • the motion of the robot is realized by the rotational motion of each joint, and each joint rotates together to achieve various desired poses.
  • the influence of the external force on the motion of the robot is also mainly reflected in the joint torque information.
  • the joint torque information and robot mechanical model are obtained, and the main external forces affected by the robot are obtained, and decoupling processing is performed, thereby improving the detection accuracy of the joint torque sensor, thereby improving the force control accuracy of the robot.
  • the obtaining step S2 is used to obtain joint torque information of each joint.
  • the obtaining step S2 includes: obtaining the joint torque information of each joint through a torque sensor, that is, each joint
  • the joints include torque sensors, and the torque sensors of each joint are detected respectively to obtain joint torque information of each joint; and/or, the S2 obtaining step includes: detecting the current of each joint, and obtaining joint torque information of each joint accordingly, That is, the joint torque information of each joint can be calculated according to the current of each joint, combined with the necessary parameters of the joint and general knowledge in the field.
  • the way of obtaining joint torque information according to the current of each joint can avoid that when the torque sensor has a low initial detection accuracy, the joint force calculation based on the joint torque information detected by the torque sensor may be inaccurate, so that better performance cannot be achieved.
  • Decoupling of joint torque information from torque sensors At the same time, the joint torque information is calculated by the joint current, the accuracy of the calculated joint torque information cannot reach a high level, and the joint force information is calculated by using it, so as to decouple the joint torque information of the torque sensor, which can effectively improve the torque
  • the accuracy of the joint torque information of the sensor is obtained, and then the joint torque information with better accuracy is obtained.
  • the method includes: repeating the calculation step and the decoupling step, iteratively processing the joint torque information, and outputting the final joint torque information.
  • the final joint torque information is the basis for subsequent processing of the robot, and the force control operation of the robot is performed according to the final joint torque information. That is, after obtaining the decoupled joint torque information in the above manner, repeat the above calculation steps, and calculate the joint force information of each joint of the robot according to the joint torque information of each joint after decoupling and the mechanical model, and Perform decoupling processing to obtain the decoupled joint torque information again. That is, by repeating the calculation steps, the joint torque information is iteratively processed, the joint torque information is continuously updated, and the final joint torque information is output, and the final joint torque information is the last time The joint torque information output by the calculation step is repeated.
  • the method includes: when the number of iterations is greater than or equal to a preset number of times, outputting final joint torque information.
  • the joint torque information of each joint is set as T N
  • the joint torque information after decoupling is T N ′
  • N represents the number of times the calculation steps are repeated.
  • N 1.
  • the obtaining step obtains the joint torque information T 1 of each joint
  • the calculating step calculates the joint force information based on the joint torque information T 1 obtained in the obtaining step
  • the decoupling step applies the torque sensor to the torque sensor according to the calibration information and the joint force information.
  • the method includes: when the difference between the joint torque information output in the previous iteration period and the joint torque information output in the current iteration period is less than a preset difference, outputting the final joint rotation moment information.
  • the accuracy of joint torque information is gradually improved.
  • the number of iterations is greater than or equal to the preset number of times, or the difference between the joint torque information output in the previous iteration cycle and the joint torque information output in the current iteration cycle
  • the difference is smaller than the preset difference, it indicates that the accuracy of the joint torque information is relatively high, and at this time, higher-precision joint torque information can be obtained, thereby making the robot force control accuracy better.
  • the torque information of each joint is set as T N
  • the joint torque information after decoupling is T N ′
  • N represents the number of times to repeat the calculation steps.
  • the acquiring step acquires the joint torque information T 1 of the torque sensor of each joint
  • the calculating step calculates the joint force information based on the joint torque information T 1 acquired in the acquiring step
  • the decoupling step is based on the calibration information and the joint force information.
  • the torque sensor of the robot joint includes an input connection flange and an output connection flange, and the joint torque sensor is used as the output connection flange of the robot joint. That is, the torque sensor is not only used for detecting the output torque of the joint, but also used for the output connection flange of the joint, so that the function of detecting the joint torque information can be realized without changing the structure design of the robot joint.
  • the provided method enables higher accuracy of joint torque information and better robot force control accuracy.
  • the above-mentioned torque sensor is a torque sensor, and the torque sensor is used to output joint torque information of the robot, and the calibration information is used to represent the coupled influence of force and torque on torque.
  • the torque sensor may also be a three-dimensional torque sensor, etc.
  • the joint torque information is obtained through the torque sensor, and the decoupling processing is performed by the above method, which can effectively reduce or even eliminate the external force on the joint torque. The influence of information detection, thereby improving the force control accuracy of the robot.
  • the beneficial effect of the above preferred embodiment is that the method obtains the influence of external force on the joint torque information by presetting the calibration information of the joint torque sensor and performs decoupling processing, so that the detection accuracy of the joint torque information is high, and the force control of the robot is improved. Accuracy is better.
  • the present invention is also used to provide a multi-joint robot.
  • the multi-joint robot 100 includes several joints and connecting rods.
  • FIG. 6 a schematic diagram of the joint 20 according to an embodiment of the present invention is shown.
  • the joint includes a motor, a reducer, and a torque sensor, and the torque sensor includes an input flange connection plate and an output connection flange connection plate.
  • the multi-joint robot 100 includes: a preset information unit 30 for preset each joint of the robot The calibration information and the mechanical model of the robot are established, the calibration information represents the coupling effect of the joint force on the joint torque information of the torque sensor; the acquisition unit 40 is used to acquire the joint torque information of each joint; the calculation unit 50, electrical is connected to the preset information unit 30 and the obtaining unit 40, and is used to obtain the joint force information of each joint according to the mechanical model and joint torque information; the decoupling unit 60 is electrically connected to the torque sensor 22 and the calculation The unit 50 is configured to perform decoupling processing on the joint torque information of the torque sensor according to the calibration information and the joint force information, and output the decoupled joint torque information.
  • the motor is connected to the input end of the reducer, the input flange connection plate is connected to the output end of the reducer, and the output connection flange plate is connected to the shell 21 of the joint 20 .
  • the torque sensor 22 is connected to the output end of the reducer and the shell 21 of the joint.
  • the torque sensor is not only used to detect the torque information of the joint, but also used as the output connection flange of the joint 20.
  • the joint's torque sensor and output are connected to the flange, and the joint has a compact structure.
  • the calculation unit 50 of the robot 100 determines the force information of the joint 20, and the decoupling unit outputs the decoupled joint torque information to the joint torque information of the torque sensor 22, thereby improving the accuracy of the joint torque information.
  • the multi-joint robot 100 includes an acquisition unit 40 to acquire joint torque information of each joint 20 .
  • the acquisition unit is a torque sensor, and the joint torque information of each joint is acquired through the torque sensor of each joint, and/or the acquisition unit includes a current detection unit, according to the current detection unit The unit acquires the joint torque information of each joint.
  • the current detection unit acquires the joint torque information of each joint, the necessary parameters of the joint and general knowledge in the field, and then acquires the joint torque information of each joint.
  • the calculation unit calculates the joint force information based on the joint torque information obtained by the current detection unit, and the decoupling unit further determines the joint torque information of the torque sensor based on the joint force information. Decoupling obtains high-precision joint torque information.
  • each joint 20 of the multi-joint robot 100 includes a torque sensor
  • the torque sensor 22 includes a support beam and a measurement beam
  • the measurement beam is pasted with strain gauges
  • the support beam and the measurement beam Connected between the input flange connection plate and the output connection flange plate, the support beam is used to enhance the rigidity of the torque sensor.
  • the torque sensor is suitable for detecting joint torque information and ensuring rigidity so that it is suitable for use as an output connection flange of the joint.
  • the multi-joint robot provided by the present invention has a compact structure and good joint torque information accuracy. The motion of the multi-joint robot is completed by the rotation of each joint, and detecting the joint torque information and decoupling it can improve the force control accuracy of the robot.
  • the torque sensor is a torque sensor.
  • the multi-joint robot 100 is a six-joint collaborative robot.

Landscapes

  • Engineering & Computer Science (AREA)
  • Robotics (AREA)
  • Mechanical Engineering (AREA)
  • Manipulator (AREA)

Abstract

一种提高机器人关节转矩检测精度的方法及多关节机器人，机器人包括多个关节，各关节包括力矩传感器，高机器人关节转矩检测精度的方法包括：预设信息步骤：预设机器人各关节力矩传感器的标定信息以及建立机器人力学模型，标定信息表示关节受力对力矩传感器的关节转矩信息的耦合影响；获取步骤：获取各关节的关节转矩信息；计算步骤：根据机器人力学模型和关节转矩信息，推算机器人各关节的关节受力信息；去耦步骤：根据标定信息和关节受力信息，对力矩传感器的关节转矩信息进行去耦处理，输出去耦后的关节转矩信息，能够使关节转矩信息精度高、机器人力控精度较好。

Description

一种提高机器人关节转矩检测精度的方法及多关节机器人 技术领域
本发明涉及工业机器人领域，特别是涉及一种提高机器人关节转矩检测精度的方法及多关节机器人。
背景技术
工业机器人是在工业环境中广泛运用的一类机器人，工业机器人通常可分为传统的工业机器人和协作机器人，协作机器人是一种新型的工业机器人，其体积小、易安装，能够被用于商业、服务业、工业等多种场景下。协作机器人多元化的应用场景使得其对安全性、精度提出了更高的要求。
在一些使用场景下，例如焊接，需要工业机器人的末端精度足够好，方能够保证工作的效果。现有技术中，一些机器人生产商已经开始尝试在工业机器人的关节里安装力矩传感器，以提升机器人末端精度。目前，大部分机器人厂商都有了自身的机器人本体设计，对于机器人的优化主要在于对硬件、软件的优化，机器人自身结构相对稳定，要给现有的机器人结构增加力矩传感器，需要对机器人的本体设计做出较大改变。
同时，当工业机器人的各关节安装力矩传感器时，关节自身可能受外力影响，力矩传感器存在耦合，影响力矩传感器的检测精度，进而影响工业机器人的力控精度，使得工业机器人末端精度不足以满足精度要求高的场景。为了解决关节力矩传感器耦合的问题，现有技术中通过改变机械结构的方式予以改善，这种方式设计难度较大、实现复杂、还会增加关节的重量和体积，同时增加的机械结构增加了机器人的生成制造成本，不利于广泛推广使用。
因此，有必要设计一种易于实现、精度较好的提高机器人关节转矩检测精度的方法及多关节机器人。
发明内容
鉴于此，本发明的目的在于提供一种易于实现、精度较好的提高机器人关节转矩检测精度的方法及多关节机器人。
本发明可采用如下技术方案：一种提高机器人关节转矩检测精度的方法，所述机器人包括多个关节，各关节包括力矩传感器，所述方法包括：预设信息步骤：预设机器人各关节力矩传感器的标定信息以及建立机器人力学模型， 所述标定信息表示关节受力对力矩传感器的关节转矩信息的耦合影响；获取步骤：获取各关节的关节转矩信息；计算步骤：根据所述机器人力学模型和关节转矩信息，推算机器人各关节的关节受力信息；去耦步骤：根据所述标定信息和所述关节受力信息，对力矩传感器的关节转矩信息进行去耦处理，输出去耦后的关节转矩信息。
进一步的，所述获取步骤包括：通过力矩传感器获取各关节的关节转矩信息；和/或，检测各关节的电流，据此获取各关节的关节转矩信息。
进一步的，所述方法包括：重复所述计算步骤和去耦步骤，迭代处理所述关节转矩信息，输出最终的关节转矩信息。
进一步的，所述方法包括：当所述迭代次数大于或等于预设次数时，输出最终的关节转矩信息。
进一步的，所述方法包括：当上一迭代周期输出的关节转矩信息和当前迭代周期输出的关节转矩信息差值小于或等于预设差值时，输出最终的关节转矩信息。
进一步的，所述力矩传感器包括输入连接法兰盘和输出连接法兰盘，所述力矩传感器用作所述机器人关节的输出连接法兰盘。
进一步的，所述力矩传感器为转矩传感器。
本发明还可采用如下技术方案：一种多关节机器人，包括若干关节和连杆，所述关节包括电机、减速器、力矩传感器，所述力矩传感器包括输入法兰连接盘和输出连接连接盘，所述多关节机器人包括：预设信息单元，用于预设机器人各关节的标定信息和建立机器人的力学模型，所述标定信息表示关节受力对力矩传感器的关节转矩信息的耦合影响；获取单元，用于获取各关节的关节转矩信息；计算单元，电连接于所述预设信息单元和获取单元，用于根据所述力学模型和关节转矩信息获取各关节的关节受力信息；去耦单元，电连接于力矩传感器和计算单元，用于根据所述标定信息和所述关节受力信息，对力矩传感器的关节转矩信息进行去耦处理，输出去耦后的关节转矩信息。
进一步的，所述获取单元为力矩传感器，和/或，所述获取单元包括电流检测单元，根据所述电流检测单元获取各关节的关节转矩信息。
进一步的，所述多关节机器人为六关节协作机器人。
与现有技术相比，本发明具体实施方式的有益效果为：通过预设标定信 息和建立机器人力学模型，根据机器人各关节的力矩传感器的检测，去掉关节外力耦合影响，进而提升关节转矩信息的精度。进一步的，通过迭代处理关节转矩信息，从而进一步提升关节转矩信息的精度。此种方式去除关节外力耦合影响，无需改变机器人自身结构设计，易于实现、成本较低。
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是本发明一个实施例的提高机器人关节转矩检测精度的方法示意图
图2是本发明另一实施例的提高机器人关节转矩检测精度的方法示意图
图3是本发明又一实施例的提高机器关节转矩检测精度的方法示意图
图4是本发明一个实施例的多关节机器人的模块图
图5是本发明一个实施例的多关节机器人的示意图
图6是本发明一个实施例的多关节机器人的关节示意图
具体实施方式
为使本发明的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的方案进行清楚、完整的描述，显然，所描述的实施例是本发明的一部分实施例而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
本发明保护一种提高机器人关节转矩检测精度的方法，所述机器人包括多个关节，各关节包括力矩传感器，参图1，所述方法包括：S1预设信息步骤，预设机器人各关节力矩传感器的标定信息以及建立机器人力学模型，所述预定信息表示关节受力对力矩传感器检测的关节转矩信息的耦合影响；S2获取步骤，获取各关节的关节转矩信息；S3计算步骤，根据所述机器人力学模型和关节转矩信息，推算机器人各关节的关节受力信息；S4去耦步骤，根据所述标定信息和所述关节受力信息，对力矩传感器的关节转矩信息进行去耦处理，输出去耦后的关节转矩信息。具体的，关节的力矩传感器的标定信息表示机器人各个方向的力、力矩的耦合影响，具体的，所述标定信息表示各个方向的力对关节的力矩传感器输出的关节转矩信息的精度的影响，外力大小对于不同的机器人的影响不同，通过预设标定数据，确认各向力、力矩对力矩传感器检测的关节转矩信息的耦合影响。通过建立力学模型，根据机 器人的力学模型和关节转矩信息获知各关节的关节受力信息，其中，本方案中可采用各种建立力学模型的方式，只要能实现建立机器人力学模型的目的即可，其中，本方案中对关节转矩信息的获取也可采用多种方式。通过各关节的关节受力信息，确认各关节的外力大小，对比标定信息确定各关节的外力耦合的影响，对所述关节转矩信息进行去耦处理，去耦后的关节转矩信息精度得以提升。机器人的运动通过各关节的旋转运动实现，各关节配合旋转以达到各种期望的位姿，当机器人受到外力时，外力对机器人的运动的影响也主要体现于关节转矩信息，通过基于各关节的关节转矩信息和机器人力学模型，获知机器人受影响的主要外力作用，并去耦处理，进而提升关节力矩传感器的检测精度，从而提高了机器人的力控精度。
进一步的，所述S2获取步骤，用于获取各关节的关节转矩信息，在本发明的一个实施例中，所述S2获取步骤包括：通过力矩传感器获取各关节的关节转矩信息，即各关节包括力矩传感器，各关节的力矩传感器分别检测以获取各关节的关节转矩信息；和/或，所述S2获取步骤包括：检测各关节的电流，据此获取各关节的关节转矩信息，即根据各关节的电流，结合关节必要参数和本领域通用知识，可以计算出各关节的关节转矩信息。根据各关节的电流获取关节转矩信息的方式，可以避免力矩传感器在初始检测准确度低的情况下，根据力矩传感器检测的关节转矩信息计算关节受力可能不准，进而不能实现较好的对力矩传感器的关节转矩信息的去耦作用。同时，通过关节电流计算关节转矩信息，所计算的关节转矩信息精度不能达到较高水平，而借助其计算关节受力信息，从而对力矩传感器的关节转矩信息去耦，能够有效提升力矩传感器的关节转矩信息的精度，进而得到精度较好的关节转矩信息。
进一步的，所述方法包括：重复所述计算步骤和去耦步骤，迭代处理所述关节转矩信息，输出最终的关节转矩信息。其中，所述最终的关节转矩信息是机器人后续进行处理的基础，根据最终的关节转矩信息执行机器人的力控操作。即通过上述的方式得到去耦后的关节转矩信息后，重复所述计算步骤，根据去耦后的各关节的关节转矩信息和所述力学模型推算机器人各关节的关节受力信息，并进行去耦处理再次得到去耦后的关节转矩信息。也即，通过重复所述计算步骤，对所述关节转矩信息进行迭代处理，不断更新所述关节转矩信息，并输出最终的关节转矩信息，所述最终的关节转矩信息为最 后一次重复所述计算步骤所输出的关节转矩信息。
在本发明的一个实施例中，所述方法包括：当所述迭代次数大于或等于预设次数时，输出最终的关节转矩信息。参图2，设定各关节的关节转矩信息为T N，去耦后的关节转矩信息为T N’，N表示重复所述计算步骤的次数，第一次执行该方法时，N＝1，获取步骤获取各关节的关节转矩信息T 1，计算步骤基于获取步骤获取的关节转矩信息T 1计算关节受力信息，去耦步骤根据标定信息和所述关节受力信息对力矩传感器的关节转矩信息T N进行去耦处理，得到去耦合后的T N’，判断重复次数N是否大于或等于预设次数N preset，若判断大于，则输出最终的关节转矩信息T N’，否则，更新所述关节转矩信息T N+1＝T N’，即根据去耦合后的关节转矩信息T N’执行下一次计算步骤和去耦步骤，以再次对其进行去耦处理，再次提升其精度，同时增加重复次数N＝N+1。通过多次迭代处理，最终输出的关节转矩信息精度较好。
在本发明的另一实施例中，所述方法包括：当上一迭代周期输出的关节转矩信息和当前迭代周期输出的关节转矩信息差值小于预设差值时，输出最终的关节转矩信息。通过多次迭代处理，使得关节转矩信息的精度逐渐提升，当迭代次数大于或等于预设次数时、或者上一迭代周期输出的关节转矩信息和当前迭代周期输出的关节转矩信息差值小于预设差值时，说明关节转矩信息的精度已经相对较高，此时可以获得较高精度的关节转矩信息，进而使得机器人力控精度较好。参图3，设定各关节的转矩信息为T N，去耦后的关节转矩信息为T N’，N表示重复所述计算步骤的次数，第一次执行该方法时，N＝1，获取步骤获取各关节的的力矩传感器的关节转矩信息T 1，计算步骤基于获取步骤获取的关节转矩信息T 1计算关节受力信息，去耦步骤根据标定信息和所述关节受力信息对力矩传感器的关节转矩信息T N进行去耦处理，得到去耦合后的T N’，判断当前迭代周期输出的关节转矩信息T N’与上一迭代周期输出的关节转矩信息T N-1的差值是否小于或等于预设差值ΔT preset，若满足条件，则输出当前迭代周期的关节转矩信息T N’，否则，更新所述关节转矩信息T N’，即根据去耦合后的关节转矩信息T N’执行下一次计算步骤，以再次对其进行去耦处理，再次提升其精度，同时增加重复次数N＝N+1。通过多次迭代处理，最终输出的关节转矩信息精度较好。同时，设置T 0＝+∞，以避免产生错误的运算结果。
在本发明的一个实施例中，机器人关节的力矩传感器包括输入连接法兰 盘和输出连接法兰盘，所述关节力矩传感器用作所述机器人关节的输出连接法兰盘。即所述力矩传感器既用于检测关节的输出转矩，又用于关节的输出连接法兰盘，使得在不改变机器人关节结构设计的基础上，实现检测关节转矩信息的功能，同时本发明所提供的方法使得关节转矩信息的精度较高，机器人力控精度较好。
在本发明的一个实施例中，上述力矩传感器为转矩传感器，所述转矩传感器用于输出机器人的关节转矩信息，所述标定信息用于表示力和力矩对转矩的耦合影响。在其他的实施例中，所述力矩传感器还可能是三维力矩传感器等，通过所述力矩传感器获知关节转矩信息，并通过上述方法进行去耦处理，能够有效减小甚至消除外力对关节转矩信息检测的影响，进而提升机器人的力控精度。
以上优选实施例的有益效果在于：该方法通过预设关节力矩传感器的标定信息，获知外力对关节转矩信息的影响并进行去耦处理，使得关节转矩信息检测精度较高，机器人的力控精度较好。
本发明还用于提供一种多关节机器人，参图4-5，所述多关节机器人100包括若干关节和连杆，参图6，示出了本发明一个实施例的关节20的示意图，所述关节包括电机、减速器、力矩传感器，所述力矩传感器包括输入法兰连接盘和输出连接法兰连接盘，所述多关节机器人100包括：预设信息单元30，用于预设机器人各关节的标定信息和建立机器人的力学模型，所述标定信息表示关节受力对力矩传感器的关节转矩信息的耦合影响；获取单元40，用于获取各关节的关节转矩信息；计算单元50，电连接于所述预设信息单元30和获取单元40，用于根据所述力学模型和关节转矩信息获取各关节的关节受力信息；去耦单元60，电连接于所述力矩传感器22和计算单元50，用于根据所述标定信息和所述关节受力信息，对力矩传感器的关节转矩信息进行去耦处理，输出去耦后的关节转矩信息。进一步的，所述电机连接于所述减速器的输入端，所述输入法兰连接盘连接于所述减速器的输出端，所述输出连接法兰盘连接于所述关节20的外壳21。所述力矩传感器22连接于所述减速器的输出端和关节的外壳21，所述力矩传感器既用于检测关节转矩信息，又用作关节20的输出连接法兰盘，相比于单独安装关节的力矩传感器和输出连接法兰盘的方式，关节结构紧凑。机器人100的计算单元50确定关节20受力信息，去耦单元对力矩传感器22的关节转矩信息输出去耦合后的关节转矩 信息，从而提升了关节转矩信息的精度。
多关节机器人100包括获取单元40，以获取各关节20的关节转矩信息。在本发明的一个实施例中，所述获取单元为力矩传感器，通过各关节的力矩传感器获取各关节的关节转矩信息，和/或，所述获取单元包括电流检测单元，根据所述电流检测单元获取各关节的关节转矩信息，具体的，电流检测单元获取各关节的关节转矩信息，关节必要参数和本领域通用知识，进而获取各关节的关节转矩信息。采用电流检测单元的方式获取关节转矩信息时，计算单元基于该电流检测单元获取的关节转矩信息计算关节受力信息，去耦单元进而基于该关节受力信息对力矩传感器的关节转矩信息去耦得到高精度的关节转矩信息。
在本发明的一个实施例中，所述多关节机器人100的各关节20包括力矩传感器，所述力矩传感器22包括支撑梁和测量梁，所述测量梁粘贴应变片，所述支撑梁和测量梁连接于所述输入法兰连接盘和输出连接法兰盘之间，所述支撑梁用于增强所述力矩传感器的刚性。所述力矩传感器适于检测关节转矩信息以及保证刚性以使得其适于作为关节的输出连接法兰盘使用。本发明所提供的多关节机器人的结构紧凑、关节转矩信息精度好。多关节机器人的运动由各关节的转动而完成，检测关节转矩信息并对其去耦处理可以提升机器人的力控精度，优选的，所述力矩传感器为转矩传感器。
在一个具体的实施例中，所述多关节机器人100为六关节协作机器人。
以上所述实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进，这些都属于本发明的保护范围。因此，本发明专利的保护范围应以所附权利要求为准。

Claims (10)

  1. 一种提高机器人关节转矩检测精度的方法，所述机器人包括多个关节，各关节包括力矩传感器，其特征在于，所述方法包括：
    预设信息步骤：预设机器人各关节力矩传感器的标定信息以及建立机器人力学模型，所述标定信息表示关节受力对力矩传感器的关节转矩信息的耦合影响；
    获取步骤：获取各关节的关节转矩信息；
    计算步骤：根据所述机器人力学模型和关节转矩信息，推算机器人各关节的关节受力信息；
    去耦步骤：根据所述标定信息和所述关节受力信息，对力矩传感器的关节转矩信息进行去耦处理，输出去耦后的关节转矩信息。
  2. 根据权利要求1所述的方法，其特征在于，所述获取步骤包括：通过力矩传感器获取各关节的关节转矩信息；和/或，检测各关节的电流，据此获取各关节的关节转矩信息。
  3. 根据权利要求1所述的方法，其特征在于，所述方法包括：重复所述计算步骤和去耦步骤，迭代处理所述关节转矩信息，输出最终的关节转矩信息。
  4. 根据权利要求3所述的方法，其特征在于，所述方法包括：当所述迭代次数大于或等于预设次数时，输出最终的关节转矩信息。
  5. 根据权利要求3所述的方法，其特征在于，所述方法包括：当上一迭代周期输出的关节转矩信息和当前迭代周期输出的关节转矩信息差值小于或等于预设差值时，输出最终的关节转矩信息。
  6. 根据权利要求1所述的方法，其特征在于，所述力矩传感器包括输入连接法兰盘和输出连接法兰盘，所述力矩传感器用作所述机器人关节的输出连接法兰盘。
  7. 根据权利要求1所述的方法，其特征在于，所述力矩传感器为转矩传感器。
  8. 一种多关节机器人，包括若干关节和连杆，其特征在于，所述关节包括电机、减速器、力矩传感器，所述力矩传感器包括输入法兰连接盘和输出连接法兰盘，所述多关节机器人包括：
    预设信息单元，用于预设机器人各关节的标定信息和建立机器人的力学模型，所述标定信息表示关节受力对力矩传感器的关节转矩信息的耦合影响；
    获取单元，用于获取各关节的关节转矩信息；
    计算单元，电连接于所述预设信息单元和获取单元，用于根据所述力学模型 和关节转矩信息获取各关节的关节受力信息；
    去耦单元，电连接于力矩传感器和计算单元，用于根据所述标定信息和所述关节受力信息，对力矩传感器的关节转矩信息进行去耦处理，输出去耦后的关节转矩信息。
  9. 根据权利要求8所述的多关节机器人，其特征在于，所述获取单元为力矩传感器，和/或，所述获取单元包括电流检测单元，根据所述电流检测单元获取各关节的关节转矩信息。
  10. 根据权利要求8所述的多关节机器人，其特征在于，所述多关节机器人为六关节协作机器人。
PCT/CN2022/072969 2021-01-29 2022-01-20 一种提高机器人关节转矩检测精度的方法及多关节机器人 WO2022161245A1 (zh)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202110126502.9 2021-01-29
CN202110126502.9A CN114800465B (zh) 2021-01-29 2021-01-29 一种提高机器人关节转矩检测精度的方法及多关节机器人
CN202120255894.4 2021-01-29
CN202120255894 2021-01-29

Publications (1)

Publication Number Publication Date
WO2022161245A1 true WO2022161245A1 (zh) 2022-08-04

Family

ID=82654161

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/072969 WO2022161245A1 (zh) 2021-01-29 2022-01-20 一种提高机器人关节转矩检测精度的方法及多关节机器人

Country Status (1)

Country Link
WO (1) WO2022161245A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115674190A (zh) * 2022-09-30 2023-02-03 深圳市越疆科技有限公司 协作机械臂及其运动控制方法、碰撞检测方法、控制系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4547858A (en) * 1983-06-13 1985-10-15 Allied Corporation Dynamic control for manipulator
CN108312146A (zh) * 2018-01-31 2018-07-24 南京工程学院 模块化机器人分散控制方法及系统
CN108638072A (zh) * 2018-05-25 2018-10-12 珞石（山东）智能科技有限公司 用于工业机器人的高精度控制装置及方法
CN110142764A (zh) * 2019-05-15 2019-08-20 南京邮电大学 基于遗传算法的力矩姿态无线传感器的数据融合处理系统及处理方法
CN111469128A (zh) * 2020-04-21 2020-07-31 昆明理工大学 一种关节型机器人电流耦合信号分离提取方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4547858A (en) * 1983-06-13 1985-10-15 Allied Corporation Dynamic control for manipulator
CN108312146A (zh) * 2018-01-31 2018-07-24 南京工程学院 模块化机器人分散控制方法及系统
CN108638072A (zh) * 2018-05-25 2018-10-12 珞石（山东）智能科技有限公司 用于工业机器人的高精度控制装置及方法
CN110142764A (zh) * 2019-05-15 2019-08-20 南京邮电大学 基于遗传算法的力矩姿态无线传感器的数据融合处理系统及处理方法
CN111469128A (zh) * 2020-04-21 2020-07-31 昆明理工大学 一种关节型机器人电流耦合信号分离提取方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115674190A (zh) * 2022-09-30 2023-02-03 深圳市越疆科技有限公司 协作机械臂及其运动控制方法、碰撞检测方法、控制系统
CN115674190B (zh) * 2022-09-30 2024-05-07 深圳市越疆科技股份有限公司 协作机械臂及其运动控制方法、碰撞检测方法、控制系统

Similar Documents

Publication Publication Date Title
WO2021238049A1 (zh) 机械臂的多负载自适应重力补偿方法、装置及控制设备
WO2022161245A1 (zh) 一种提高机器人关节转矩检测精度的方法及多关节机器人
JP5550468B2 (ja) 力覚センサの校正方法
WO2021159662A1 (zh) 一种安装偏心状态下谐波减速器柔轮径向变形的检测方法
TW201914788A (zh) 用於機械臂之負載估測重力補償的方法及其負載估測重力補償系統
CN108326634B (zh) 一种面向超声振动辅助切削的原位切削力测量装置与方法
CN106774350A (zh) 双红外精确校准机器人与充电平台对接角度的机器人
CN109100073B (zh) 一种基于应变反演的六维力传感器及其测量方法
CN101100059A (zh) 柔性两轮自平衡机器人的姿态检测方法
CN102166755B (zh) 一种遥操作机器人的机械手末端三维接触力测量方法
WO2022121003A1 (zh) 机器人控制方法、装置、计算机可读存储介质及机器人
CN115946120B (zh) 机械臂控制方法、装置、设备和介质
CN115091455B (zh) 一种工业机器人定位误差补偿方法
CN110470223A (zh) 一种基于1d位移传感器的三维空间坐标测量方法
CN116901081B (zh) 机器人dh参数标定方法、装置、电子设备及存储介质
CN108108555B (zh) 一种装配尺寸链计算方法、装置及设备
CN114800465B (zh) 一种提高机器人关节转矩检测精度的方法及多关节机器人
CN113043325B (zh) 一种机器人关节运动状态的检测方法及装置
WO2024016892A1 (zh) 传感器外参标定方法、装置以及电子设备
CN114791333B (zh) 力传感器测量坐标系标定方法、装置、设备及存储介质
JP6431005B2 (ja) センサ装置
CN103324094A (zh) 臂架非线性变形在线求解装置及方法
CN114472588B (zh) 一种新型弯管加工过程碰撞干涉检测方法
CN114266776B (zh) 一种应用复合裂纹位移场函数的数字图像相关方法
CN106871857B (zh) 一种滚珠丝杠弯曲挠度检测装置及方法

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: 22745120

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22745120

Country of ref document: EP

Kind code of ref document: A1