A joint control system for a humanoid robot
Patent Information
- Application Number
- CN202522639482.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Utility models(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-12
- Publication Date
- 2026-11-03
- Estimated Expiration
- 2035-12-12
AI Technical Summary
In existing technologies, the joint control system of humanoid robots lacks deep integration of multi-dimensional sensor information, which makes it impossible for traditional rigid control schemes to dynamically adjust the output during close human-computer interaction, resulting in insufficient safety of human-computer interaction and weak scene adaptability.
The system employs integrated angle, torque, and force sensors. The signals from these sensors are collected and processed in real time by the control circuit to control the actuator to drive joint movement, thereby achieving smooth joint movement within a set threshold of torque and contact force.
It improves the motion precision and interaction safety of the joint control system, adapts to the compliant control requirements of service robots in scenarios such as hugging and helping, and enhances the comfort and safety of human-computer interaction.
Smart Images

Figure CN224927983U_ABST
Abstract
Description
Technical Field
[0001] This utility model relates to humanoid robots, and more specifically, it relates to a joint control system for humanoid robots. Background Technology
[0002] As the application of humanoid robots continues to expand in service industries (such as home companionship and medical assistance), education (such as programming demonstrations and interactive teaching), and industry (such as flexible assembly), the requirements for control performance of joints, as the core components for motion execution, are becoming increasingly stringent. In existing technologies, humanoid robot joint control systems mostly adopt rigid control schemes, relying on high-precision position / angle sensors and fixed-parameter control algorithms to achieve motion positioning in order to meet basic trajectory tracking requirements.
[0003] However, as humanoid robots move towards close-range human-machine interaction (such as hugging and helping each other), the limitations of traditional rigid control schemes are gradually becoming a bottleneck in application. This is mainly reflected in their inability to dynamically adjust output based on external contact force, resulting in insufficient safety and weak adaptability in human-machine interaction. Specifically, existing technologies lack deep integration of multi-dimensional sensor information, and the integration of the control system and actuator is low, with poor coordination between force control and position control. For example, most solutions rely solely on a single sensor (such as an angle sensor) combined with a fixed angle to execute the arm hugging program. The rigid, fixed force generated during the hug can easily cause discomfort to the user, limiting the overall performance improvement of joint control. Utility Model Content
[0004] The technical problem to be solved by this utility model is to provide a joint control system for humanoid robots that addresses the shortcomings of existing technologies. This system solves the problem that a single sensor cannot fully perceive the joint state in existing technologies, and also solves the problem of "large human-machine interaction impact" in traditional rigid control.
[0005] This utility model discloses a joint control system for a humanoid robot, comprising a first connecting arm, a second connecting arm, and a joint component installed between the two connecting arms. The joint component houses a driver, a control circuit, an angle sensor, and a torque sensor. Both the first and second connecting arms are equipped with force-tactile sensors near the joint component, with the force-tactile sensors mounted on the side of the connecting arm facing the humanoid robot arm to form a hugging motion. The control circuit collects electrical signals from the angle sensor, torque sensor, and force-tactile sensor, and controls the driver to rotate the connecting arms so that when the humanoid robot arm hugs an object, the values of the electrical signals from the torque sensor and the force-tactile sensor do not exceed corresponding set thresholds.
[0006] Preferably, the angle sensor is fixedly installed in the joint, and the rotating shaft of the angle sensor is fixed in the connecting arm.
[0007] Preferably, the angle sensor and torque sensor are integrated multi-functional sensors.
[0008] Preferably, the connecting arm one and connecting arm two have the same structure, both including an arm rod and connecting plates fixed at both ends of the arm rod, and two ear plates fixed on the end face of the connecting plate; the joint is rotatably mounted between the two ear plates, and the force tactile sensor is mounted on the side wall of the connecting plate.
[0009] Preferably, the diameter of the connecting disc is greater than the distance between the two ear plates.
[0010] Preferably, the control circuit includes a processor unit, a drive execution unit, and a power supply unit; the processor unit is connected to the angle sensor, torque sensor, and force tactile sensor via a CAN bus, and the processor unit is connected to the driver via the drive execution unit.
[0011] Beneficial effects The advantages of this utility model are: it integrates a circuit module based on a torque sensor, an angle sensor, a force and tactile sensor, a control unit, and a drive execution unit at the joint, which solves the defect of the existing technology that a single sensor cannot fully perceive the joint state. It also solves the defect of "large human-machine interaction impact" in traditional rigid control, and takes into account both motion accuracy and interaction safety, making it suitable for service robot "hugging, helping" and other scenarios. Attached Figure Description
[0012] Figure 1 This is a schematic diagram of the joint component in the joint control system of this utility model. Figure 2 This is an enlarged structural diagram of the joint portion in the joint control system of this utility model; Figure 3 This is a three-dimensional structural diagram of the joint component of this utility model; Figure 4 This is a schematic diagram of the control circuit part of the joint control system of this utility model.
[0013] Among them: 1-Connecting arm one, 2-Connecting arm two, 3-Joint component, 4-Arm, 5-Connecting plate, 6-Ear plate, 7-Force tactile sensor, 8-Multifunctional sensor, 9-X-rotation axis, 10-Y-rotation axis, 11-Processor unit, 12-Drive execution unit, 13-Power supply unit, 14-CAN bus. Detailed Implementation
[0014] The present invention will be further described below with reference to embodiments, but this does not constitute any limitation on the present invention. Any limited modifications made by any person within the scope of the claims of the present invention are still within the scope of the claims of the present invention. See Figures 1-4 This utility model discloses a joint control system for a humanoid robot, comprising a first connecting arm 1, a second connecting arm 2, and a joint component 3 installed between the two connecting arms. Specifically, the first connecting arm 1 and the second connecting arm 2 have identical structures, each including an arm 4, connecting discs 5 fixed to both ends of the arm 4, and two ear plates 6 fixed to the end faces of the connecting discs 5. The joint component 3 is rotatably mounted between the two ear plates 6, forming an X-axis rotation 9 and a Y-axis rotation 10 between the joint component 3 and the two connecting arms. A driver is installed along the path of these two rotation axes, the driver being fixed in the joint component 3, and its driving end being fixed to the ear plate 6, thereby realizing the rotation of the connecting arm.
[0015] To address the issue of traditional humanoid robot joint control where central processing unit (CPU) issues unified control and drive commands, this embodiment incorporates a control circuit in joint 3 that stores drive commands. By simply receiving control commands from the CPU, the circuit can invoke the corresponding drive commands to activate the actuator, thereby controlling the rotation of the connecting arm and optimizing CPU utilization.
[0016] The control circuit includes a processor unit 11, a drive execution unit 12, and a power supply unit 13. The power supply unit 13 supplies power to the processor unit 11, drive execution unit 12, driver, and various sensors in the joint. The processor unit 11 is connected to the driver via the drive execution unit 12, enabling the transmission and execution of drive commands. The processor unit 11 uses an STM32 processor with 128KB RAM + 1MB Flash memory. Its communication interface supports CAN bus / USB / serial port, and it is used for signal processing and command transmission and reception. The drive execution unit 12 uses a mature MOSFET bridge drive circuit with a maximum current of 20A. The driver employs a combination of a reducer and a servo motor. The servo motor has a rated torque of 50 N·m and a speed of 300 rpm; the reducer has a reduction ratio of 1:50. The drive execution unit 12 controls the driver to output the corresponding torque value according to the drive commands, driving the joint to the target angle. The power supply unit 13 uses a DC-DC converter with an input of 12~24V and an output of 5V / 12V / 24V, providing stable power supply for each module and adapting to the power supply system of the humanoid robot.
[0017] An angle sensor and a torque sensor are installed in joint 3. Specifically, the angle sensor and torque sensor are integrated multi-functional sensors 8. The sensor body is fixedly installed in joint 3, and the sensor shaft is fixed in the connecting arm, used to collect torque and angle during joint movement. The torque sensor has a range of 0.5~50 N·m and an accuracy of ±0.5%FS; the angle sensor has a resolution of 0.1° and a range of -180°~+180°. In addition, force tactile sensors 7 are installed on both connecting arms 1 and 2 near joint 3, and the force tactile sensors 7 are installed on the side of the connecting arm facing the humanoid robot arm to form a hug. Specifically, the force tactile sensors 7 are installed on the side wall of the connecting plate 5, and the diameter of the connecting plate 5 is larger than the distance between the two ear plates 6, so that when the joint comes into contact with an object, the connecting plate 5 with the force tactile sensors 7 can make contact with the object first, improving the detection accuracy. The force sensor 7 has a sensitivity of 0.1N and a response time of ≤5ms, and is used to collect the contact force when the humanoid robot's joints come into contact with external objects. The angle sensor, torque sensor, and force sensor 7 together provide multi-dimensional data support for control.
[0018] The processor unit 11 is wired to the angle sensor, torque sensor, and force sensor 7 in the joint via a CAN bus 14. Using the CAN bus 14 for synchronous data transmission reduces signal delay and ensures real-time control. Specifically, the control circuit acquires the electrical signals from the angle sensor, torque sensor, and force sensor 7, and controls the driver to rotate the connecting arm so that the values of the electrical signals from the torque sensor and force sensor 7 do not exceed corresponding set thresholds when the humanoid robot arm hugs an object. This means that the torque sensor directly measures the joint output torque, the angle sensor acquires the motion posture in real time, and the force sensor 7 captures external contact force. This solves the problem of a single sensor not being able to fully perceive the joint state, and also addresses the drawback of traditional rigid control's "large impact during human-machine interaction," balancing motion accuracy and interaction safety, and is suitable for service robot scenarios such as "hugging" and "helping."
[0019] Taking the scenario of a humanoid robot "embracing the user" as an example, the joint control system of this utility model is as follows: Step 1: After the system is powered on, the control unit loads the default parameters, and the initial angle of the arm is 0°.
[0020] Step 2: When the humanoid robot sends a control command to the control unit, the control unit drives the actuator according to the loaded drive command to bring the humanoid robot's arm closer to the user. For example, the arm is controlled to move at 0.5 rad / s, increasing the angle to 30°. During the movement of the humanoid robot's arm, the control unit collects the signal values of each sensor in real time. Assuming that when the humanoid robot's arm moves, the control unit collects a joint torque of 2 N·m through the torque sensor, and when the current angle collected by the angle sensor is 30°, the contact force collected by the force sensor 7 is 0.8 N, indicating that the humanoid robot's arm is in contact with the user.
[0021] The third step is to automatically switch to the compliant mode to perform the hugging action: The control unit controls the driver to continue working until the torque sensor feedback is above 3 N·m and the contact force feedback from the force sensor 7 is stable at 5.0 N, indicating that the "hugging user" action is completed.
[0022] Step 4, Task Complete: After the set hug time, the system automatically switches back to the initial mode, releases the hug action, and waits for the next instruction.
[0023] This invention utilizes multi-sensor fusion to determine whether the arm is in contact with the user during movement based on data collected by the sensors. Upon contact, torque and contact force are increased until a set threshold is reached, generating a hugging experience and thus achieving a smooth hugging motion. This invention specifically addresses the rigid control defects of existing technologies that use a single sensor combined with a fixed angle to execute arm hugging. It can be widely applied to joint control scenarios in fields such as service robots and educational robots, demonstrating significant technological innovation and practicality.
[0024] The above description is only a preferred embodiment of the present utility model. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the structure of the present utility model. These modifications and improvements will not affect the effectiveness of the present utility model or the practicality of the patent.
Claims
1. A joint control system for a humanoid robot, comprising a first connecting arm (1), a second connecting arm (2), and a joint component (3) mounted between the two connecting arms, wherein a driver is installed in the joint component (3), characterized in that, The joint (3) is equipped with a control circuit, an angle sensor and a torque sensor. Both the connecting arm 1 (1) and the connecting arm 2 (2) are equipped with force touch sensors (7) near the joint (3), and the force touch sensors (7) are installed on the side of the connecting arm facing the humanoid robot arm to form a hug. The control circuit collects the electrical signals of the angle sensor, the torque sensor and the force touch sensor (7) and controls the driver to drive the connecting arm to rotate so that the values of the electrical signals of the torque sensor and the force touch sensor (7) are not greater than the corresponding set threshold when the humanoid robot arm hugs the object.
2. A joint control system for a humanoid robot according to claim 1, characterized in that, The angle sensor is fixedly installed in the joint (3), and the rotating shaft of the angle sensor is fixed in the connecting arm.
3. A joint control system for a humanoid robot according to claim 1 or 2, characterized in that, The angle sensor and torque sensor are integrated multi-functional sensors (8).
4. A joint control system for a humanoid robot according to claim 1, characterized in that, The structure of the connecting arm one (1) and the connecting arm two (2) is the same, both including the arm (4) and the connecting plate (5) fixed at both ends of the arm (4), and two ear plates (6) fixed on the end face of the connecting plate (5); the joint (3) is rotatably installed between the two ear plates (6), and the force tactile sensor (7) is installed on the side wall of the connecting plate (5).
5. A joint control system for a humanoid robot according to claim 4, characterized in that, The diameter of the connecting plate (5) is greater than the distance between the two ear plates (6).
6. A joint control system for a humanoid robot according to claim 1, characterized in that, The control circuit includes a processor unit (11), a drive execution unit (12), and a power supply unit (13); the processor unit (11) is connected to the angle sensor, torque sensor, and force tactile sensor (7) via a CAN bus (14), and the processor unit (11) is connected to the driver via the drive execution unit (12).
