Control method, electronic device, medium, and program product

Abstract

The application relates to the technical field of robots, and discloses a control method, an electronic device, a medium and a program product. In the method, it is detected that at least one first joint of a robot system is running in a preset torque interval corresponding to the at least one first joint; thermal state information of each joint in the robot system is detected; and when the thermal state information corresponding to at least one joint of the at least one first joint satisfies a preset condition, the robot system is converted from a first configuration to a second configuration, so that at least part of a load is transferred from the joint whose thermal state information satisfies the preset condition to at least one second joint of the robot system. By utilizing the thermal load asymmetry related to the posture of the robot system, at least part of the load is transferred from the joint with poor thermal state to other joints, the thermal capacity on the motion chain is redistributed, the joint with exhausted heat is naturally recovered in a non-load-bearing state, and meanwhile, the joint with sufficient heat bears the load, thereby forming sustainable heavy-load operation capacity.

Description

Technical Field

[0001] This application relates to the field of robotics, and in particular to a control method, electronic device, medium, and program product. Background Technology

[0002] With the widespread application of multi-joint robot systems (such as humanoid robots and quadruped robots) in industrial warehousing, logistics handling, disaster relief and other scenarios, multi-joint robot systems need to have the ability to handle heavy loads in order to meet the high-load operation requirements in actual working conditions.

[0003] Currently, each joint actuator in a multi-joint robot system typically has a peak torque (τ_peak) capability that is higher than the continuous rated torque (τ_cont) (typically 3 to 5 times the continuous rated torque).

[0004] However, in current control strategies, in order to ensure that the multi-joint robot system can work continuously in all possible postures, the effective load capacity of the multi-joint robot system is determined by the continuous torque of the weakest joint in its worst stress posture.

[0005] Therefore, even though most joints still have sufficient thermal capacity margin in most postures, the rated payload of the multi-joint robot system is reduced by the limitations of a few weak links in extreme postures, resulting in the inherent peak torque capability of the hardware being basically unutilized. For example, the actual payload is only 20-30% of the peak capability of the actuator. Summary of the Invention

[0006] The purpose of this application is to provide a control method, electronic device, medium, and program product.

[0007] A first aspect of this application provides a control method applied to an electronic device. The method includes: detecting that at least one first joint of a robot system is operating within a preset torque range corresponding to the at least one first joint, wherein the minimum value of the preset torque range corresponding to each first joint is greater than the continuous rated torque value of the first joint; detecting thermal state information of each joint in the robot system, wherein the thermal state information is used to indicate the remaining heat capacity of the corresponding joint during task execution; and converting the robot system from a first configuration to a second configuration when the thermal state information of at least one of the at least one first joint meets the preset conditions, so as to transfer at least a portion of the load from the joint where the thermal state information meets the preset conditions to at least one second joint of the robot system.

[0008] In this embodiment, by utilizing the asymmetry of thermal load related to posture in the robot system, at least part of the load is transferred from the joints with poor thermal condition to other joints by converting the robot system from the first configuration to the second configuration. This achieves the redistribution of thermal capacity in the kinetic chain, allowing the joints that have exhausted their heat to recover naturally in a non-load-bearing state, while the joints with sufficient heat bear the load, thus forming a sustainable heavy-load operation capability.

[0009] Furthermore, when the joint thermal status information is detected to meet the preset conditions (i.e., the remaining heat capacity is lower than the safety threshold), the attitude conversion is actively triggered before the thermal limit is reached, rather than waiting for the overheating event to occur before performing derating or shutdown. This can avoid task interruption and ensure operational continuity.

[0010] Moreover, it requires no hardware modification to the robot system and can be directly deployed on existing robot systems with peak torque capability and joint mobility (including humanoid robots, quadruped robots, etc.), offering good compatibility and low implementation costs.

[0011] In one possible implementation of the first aspect described above, the thermal state information is determined based on at least one of the following parameters: the real-time temperature of the joint; the historical torque or historical drive current of the joint; the current torque or current drive current of the joint; the current posture of the limb segment in which the joint is located; and the thermal threshold temperature of the joint.

[0012] In this embodiment, by using multiple parameters such as real-time joint temperature, historical / current torque or drive current, limb segment posture, and thermal threshold temperature as the basis for determining thermal state information, the accuracy and robustness of thermal state assessment can be improved, so as to more accurately determine the remaining heat capacity of each joint, thereby providing a reliable data basis for configuration conversion decisions.

[0013] In one possible implementation of the first aspect above, the thermal state information of at least one of the first joints satisfies a preset condition, including: the remaining thermal capacity of at least one joint is less than or equal to a preset thermal threshold, the preset thermal threshold being a preset proportion of the maximum thermal capacity of at least one joint at the ambient temperature of the robot system.

[0014] In this embodiment of the application, by setting a preset thermal threshold as a triggering condition, a clear quantitative judgment standard can be provided for the thermal depletion state of the joint, so as to ensure that the load transfer is triggered in time when the joint thermal capacity is about to be depleted but has not yet reached the dangerous critical point, thereby realizing active thermal management rather than passive overheating response.

[0015] In one possible implementation of the first aspect above, detecting that at least one first joint of the robot system is operating within a preset torque range corresponding to at least one first joint includes: detecting that the torque value of at least one first joint is greater than the continuous rated torque value of at least one first joint; or detecting that the drive current value of at least one first joint is greater than the current threshold corresponding to the continuous rated torque of at least one first joint.

[0016] In this embodiment, the joint is determined to be operating within a preset torque range by detecting whether the joint torque value or the drive current value exceeds the threshold corresponding to the continuous rated value. This can quickly identify whether the joint is in a peak torque utilization state without the need for complex thermal model calculations, thereby reducing computational overhead.

[0017] In one possible implementation of the first aspect above, converting the robot system from the first configuration to the second configuration includes at least one of the following: shortening the effective length of the limb segment bearing the first joint to reduce the gravitational lever arm on the first joint; positioning the payload near the center of mass of the robot system; converting the robot system to a posture that transmits the load using a near-singular joint configuration; and configuring the joints of the robot system to bear at least part of the load using skeletal or mechanical restraints.

[0018] In this embodiment, configuration transformation is achieved by shortening the effective length of the limbs, bringing the payload closer to the center of mass, utilizing near-singular joint configurations, and using skeletal or mechanical constraints to bear the load. That is, the joint load is reduced from a geometric and mechanical structural perspective. These various methods can be used individually or in combination to adapt to different robot configurations and operational scenarios, effectively reducing the load-bearing burden on joints with insufficient heat capacity.

[0019] In one possible implementation of the first aspect described above, the method further includes: engaging a mechanical brake on a joint where the thermal state information meets preset conditions; and partially or completely disconnecting the actuator power supply to the joint engaging the mechanical brake.

[0020] In this embodiment, by engaging a mechanical brake on a joint that meets preset thermal conditions and partially or completely disconnecting the actuator power supply, the overheated joint can enter a zero-power or low-power thermal recovery mode, eliminating the source of resistive heat generation, achieving passive heat dissipation at the maximum rate, and reducing overall energy consumption.

[0021] In one possible implementation of the first aspect above, the engagement and disengagement of the mechanical brake are coordinated with the posture transition of the robot system in time; and the actuators of the corresponding joints are re-energized before the mechanical brake disengages.

[0022] In this embodiment, the engagement / disengagement of the mechanical brake and the attitude transition action are coordinated in time, and the actuator is re-energized before the brake disengages. This ensures that the joint does not undergo thermal recovery when locked and is energized in advance when movement is required. This achieves seamless integration of thermal management and motion control, thereby avoiding load drop or attitude loss caused by delayed brake disengagement or delayed actuator energization.

[0023] In one possible implementation of the first aspect above, the method further includes: corresponding to the robot system being in a second configuration, detecting the thermal state information of each joint of the robot system; corresponding to the thermal state information of the current load-bearing joint of the robot system meeting a preset condition, converting the robot system from the second configuration to a third configuration or a first configuration.

[0024] In this embodiment, after the robot is in the second configuration, the thermal state of each joint is continuously detected, and when the thermal state of the current load-bearing joint meets the preset conditions, the configuration conversion is triggered again, that is, a cyclic heat redistribution mechanism is formed, so that the load is transferred sequentially between different joint groups, realizing the alternating consumption and recovery of heat capacity, thereby maintaining long-term operation under load conditions that exceed the continuous torque capacity.

[0025] In one possible implementation of the first aspect above, the robot system includes one or more robots; and the robots include at least one of the following types: humanoid robots, quadruped robots, and mobile manipulation robots.

[0026] In this application embodiment, the method is applied to a robot system comprising one or more robots, including humanoid robots, quadruped robots, mobile manipulation robots, etc. This improves the versatility of the control method in this application; for example, it can be adapted to various multi-joint robot configurations, covering multiple application scenarios such as industrial handling, logistics warehousing, and disaster relief.

[0027] In one possible implementation of the first aspect above, the corresponding robot system includes a robot, the robot includes a plurality of joints, and at least one first joint and at least one second joint are joints in the robot; or, the corresponding robot system includes at least two robots, at least one first joint is a joint located in the same robot or a joint located in different robots, at least one second joint is a joint located in the same robot or a joint located in different robots, and at least one first joint and at least one second joint are located in the same robot or different robots.

[0028] In this embodiment, for single-robot scenarios, load transfer is performed between different joints of the same robot; for multi-robot collaborative scenarios, cross-robot load transfer is performed between joints of different robots. This allows thermal budget management to be extended from within the single-robot kinematic chain to the global multi-robot system, further enhancing the system's sustainable operational capability in heavy-duty collaborative tasks through cross-robot thermal redistribution.

[0029] A second aspect of this application provides an electronic device including a processor and a memory, the memory being used to store program instructions, and the processor being used to invoke the program instructions to execute any of the methods described in the first aspect above.

[0030] A third aspect of this application provides a computer-readable storage medium storing instructions that, when executed, implement any of the methods described in the first aspect above.

[0031] The fourth aspect of this application provides a computer program product including instructions that, when executed, implement any of the methods described in the first aspect above.

[0032] In the embodiments of this application, when the second to fourth aspects implement any one of the methods in the first aspect, they can achieve the same or similar technical effects as any one of the methods in the first aspect. Attached Figure Description

[0033] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0034] Figure 1 A flowchart of a control method is shown according to an embodiment of this application;

[0035] Figure 2A A schematic diagram of an initial configuration for the burst lift phase is shown according to an embodiment of this application;

[0036] Figure 2B A schematic diagram of a close-fitting configuration after a posture change is shown according to an embodiment of this application;

[0037] Figure 2C A schematic diagram of a continuous holding configuration of brake engagement is shown according to an embodiment of this application;

[0038] Figure 2DA schematic diagram of alternating limb load transfer during a cyclic heat redistribution stage is shown according to an embodiment of this application;

[0039] Figure 3 A flowchart of a model training method is shown according to an embodiment of this application;

[0040] Figure 4 A schematic diagram of the structure of a robot system 200 is shown according to an embodiment of this application;

[0041] Figure 5 A schematic diagram of the hardware structure of an electronic device 100 is shown according to an embodiment of this application. Detailed Implementation

[0042] The illustrative embodiments of this application include, but are not limited to, a control method, an electronic device, a medium, and a program product.

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described clearly and in detail below with reference to the accompanying drawings.

[0044] As mentioned earlier, each joint actuator in a multi-joint robot system typically has a peak torque (τ_peak) capability that is higher than the continuous rated torque (τ_cont) (typically 3 to 5 times the continuous rated torque).

[0045] However, in current control strategies, in order to ensure that the multi-joint robot system can work continuously in all possible postures, the effective load capacity of the multi-joint robot system is determined by the continuous torque of the weakest joint in its worst stress posture.

[0046] Therefore, even though most joints still have sufficient thermal capacity margin in most postures, the rated payload of the multi-joint robot system is reduced by the limitations of a few weak links in extreme postures, resulting in the inherent peak torque capability of the hardware being basically unutilized. For example, the actual payload is only 20-30% of the peak capability of the actuator.

[0047] For example, consider humanoid robot A moving heavy objects in a warehouse. Assume the continuous rated torque of robot A's shoulder joint is 30 N·m. When robot A carries the goods close to its chest, the shoulder joint lever arm is short, requiring only 20 N·m of torque (below the continuous rated torque). However, when the robot needs to lift the goods to a higher position and straightens its arm, the shoulder joint lever arm becomes longer, and the same weight of goods will subject it to 35 N·m of torque, exceeding the continuous rated torque.

[0048] In the current control strategy, to ensure that humanoid robot A can operate continuously without overheating and damage in all possible postures (including the adverse posture of extending its arm), the payload capacity of humanoid robot A is calibrated according to the shoulder joint not exceeding its continuous rated torque (i.e., 30 N·m) in the worst posture. This results in the shoulder joint only outputting about 67% of its continuous rated torque, even in the more effortless close-body handling posture, while the peak torque that the shoulder joint could have output instantaneously (e.g., 90-100 N·m) is completely unutilized. Ultimately, the payload capacity of humanoid robot A is only 20-30% of the peak capacity of each joint actuator.

[0049] In view of this, embodiments of this application provide a control method, the method comprising: detecting that at least one first joint of a robot system is operating within a preset torque range corresponding to at least one first joint, wherein the minimum value of the preset torque range corresponding to each first joint is greater than the continuous rated torque value of the first joint; detecting thermal state information of each joint in the robot system, wherein the thermal state information is used to indicate the remaining heat capacity of the corresponding joint during task execution; and, if the thermal state information of at least one of the at least one first joints satisfies a preset condition, converting the robot system from a first configuration to a second configuration to transfer at least a portion of the load from the joint whose thermal state information satisfies the preset condition to at least one second joint of the robot system.

[0050] It is understandable that by utilizing the asymmetry of thermal load related to posture in a robot system, and by converting the robot system from the first configuration to the second configuration, at least part of the load can be transferred from joints with poor thermal condition to other joints, thereby redistributing the heat capacity in the kinematic chain. This allows joints that have exhausted their heat to recover naturally in a non-load-bearing state, while joints with sufficient heat can bear the load, thus forming a sustainable heavy-load operation capability.

[0051] Furthermore, when the joint thermal status information is detected to meet the preset conditions (i.e., the remaining heat capacity is lower than the safety threshold), the attitude conversion is actively triggered before the thermal limit is reached, rather than waiting for the overheating event to occur before performing derating or shutdown. This can avoid task interruption and ensure operational continuity.

[0052] Moreover, it requires no hardware modification to the robot system and can be directly deployed on existing robot systems with peak torque capability and joint mobility (including humanoid robots, quadruped robots, etc.), offering good compatibility and low implementation costs.

[0053] To more clearly illustrate the technical solution of this application, the data processing method provided in the embodiments of this application will be described below with reference to the accompanying drawings.

[0054] Figure 1A flowchart illustrating a control method is shown according to an embodiment of this application. It can be understood that... Figure 1 The execution subject of the flowchart shown is electronic device 100. For ease of description, the following will refer to... Figure 1 When describing the flowchart shown, the execution subject of the flowchart will not be repeated.

[0055] like Figure 1 As shown, this process includes, but is not limited to:

[0056] S101: It is detected that at least one first joint of the robot system is operating in a preset torque range corresponding to at least one first joint, wherein the minimum value of the preset torque range corresponding to each first joint is greater than the continuous rated torque value of the first joint.

[0057] In some embodiments, the electronic device 100 detects that at least one first joint of the robot system is operating in a preset torque range corresponding to at least one first joint by detecting that the torque value of at least one first joint is greater than the continuous rated torque value of at least one first joint; or, detecting that the drive current value of at least one first joint is greater than the current threshold corresponding to the continuous rated torque of at least one first joint.

[0058] The torque value can be directly measured by a torque sensor installed at the joint, and the drive current value can be obtained by sampling the drive circuit of the joint actuator.

[0059] It is understandable that determining whether a joint is operating within a preset torque range by detecting whether the joint torque or drive current value exceeds the threshold corresponding to consecutive rated values ​​can quickly identify whether the joint is in peak torque utilization without complex thermal model calculations, thus reducing computational overhead. Furthermore, since there is a definite linear relationship between the output torque and drive current of the joint actuator, indirectly determining torque output by detecting the drive current is low-cost and highly real-time.

[0060] In some embodiments, the detection of whether at least one first joint of the robot system is operating within a preset torque range corresponding to at least one first joint is triggered by an operation command. For example, when the electronic device 100 receives a heavy-duty handling task, it actively commands one or more joints (i.e., at least one first joint) to operate within a preset torque range, and simultaneously begins to detect whether at least one first joint of the robot system is operating within the preset torque range corresponding to at least one first joint.

[0061] In other embodiments, the detection method is to detect whether at least one first joint of the robot system operates continuously within a preset torque range corresponding to at least one first joint. This application does not impose any limitations on this method.

[0062] S102: Detect the thermal state information of each joint in the robot system, wherein the thermal state information is used to indicate the remaining heat capacity of the corresponding joint during the execution of the task.

[0063] In some embodiments, thermal state information is determined based on at least one of the following parameters: real-time temperature of the joint; historical torque or historical drive current of the joint; current torque or current drive current of the joint; current posture of the limb segment in which the joint is located; and thermal threshold temperature of the joint.

[0064] The real-time temperature of the joint can be obtained directly by a thermal sensor or estimated using a thermal observer model. The joint's historical torque or historical drive current is used to calculate the accumulated heat over a past period and can be retrieved from the historical data storage of the electronic device 100. The joint's current torque or current drive current is used to calculate the heat generation rate at the current moment and can be obtained through a torque sensor or current sampling. The current posture of the limb segment containing the joint affects heat dissipation efficiency; different limb postures expose the joint to different airflow conditions or radiation environments. The current posture of the limb segment containing the joint can be obtained through a joint position sensor. The joint's thermal threshold is preset according to the joint actuator's specifications and indicates the highest safe temperature the joint can reach.

[0065] In some embodiments, the remaining heat capacity of a joint is calculated using a thermal budget model. The thermal budget model expresses the rate of change of thermal state as the difference between a heat generation term and a heat recovery term, where the heat generation term is proportional to the square of the current torque of the joint, and the heat recovery term is proportional to the difference between the current heat capacity and the maximum heat capacity, and the heat recovery term changes with the real-time temperature of the joint and the current posture of the limb segment in which the joint is located. By integrating this differential equation, an estimate of the remaining heat capacity at any given time can be obtained.

[0066] For example, for each joint Ji, the thermal budget Bi(t) is defined as the remaining heat capacity of the joint actuator before it reaches its thermal limit. The thermal budget model can be referenced from the following formula (1) or formula (2).

[0067] Formula (1)

[0068] Formula (2)

[0069] in, The rate of change of thermal state, For heat generation, For heat recovery, This is a temperature-dependent heat generation coefficient (proportional to I²R resistance loss, as resistance increases with temperature). The heat dissipation coefficient is temperature-dependent (depending on the characteristics of the cooling system, the current posture of the limb segment where the joint is located, and environmental conditions). This is the maximum heat capacity (maximum heat capacity at ambient temperature). This represents the instantaneous torque output of joint i.

[0070] In some embodiments, the thermal state can be divided into multiple levels based on the proportion of remaining heat capacity. A remaining heat capacity proportion greater than 50% indicates sufficient joint heat capacity for normal operation; a remaining heat capacity proportion greater than 30% and less than or equal to 50% indicates rapid heat capacity depletion requiring enhanced monitoring; a remaining heat capacity proportion greater than 20% and less than or equal to 30% indicates the joint heat capacity is approaching the depletion threshold, necessitating preparation for configuration conversion; and a remaining heat capacity proportion greater than 0 and less than or equal to 20% indicates the joint heat capacity has reached the heat depletion condition, triggering configuration conversion.

[0071] It is understandable that by using multiple parameters such as real-time joint temperature, historical / current torque or drive current, limb segment posture, and thermal threshold temperature as the basis for determining thermal state information, the accuracy and robustness of thermal state assessment can be improved, so as to more accurately determine the remaining heat capacity of each joint, thereby providing a reliable data foundation for configuration conversion decisions.

[0072] S103: When the thermal state information of at least one of the first joints meets a preset condition, the robot system is switched from the first configuration to the second configuration to transfer at least part of the load from the joint whose thermal state information meets the preset condition to at least one second joint of the robot system.

[0073] In some embodiments, the thermal state information of at least one of the first joints satisfies a preset condition, including: the remaining thermal capacity of at least one joint is less than or equal to a preset thermal threshold, wherein the preset thermal threshold is a preset proportion of the maximum thermal capacity of at least one joint at the ambient temperature of the robot system.

[0074] In other words, when the remaining heat capacity Bi(t) of joint Ji is greater than the preset heat threshold B_threshold, the heat budget of joint Ji is available. When the remaining heat capacity Bi(t) of joint Ji is less than or equal to the preset heat threshold B_threshold, the heat budget of joint Ji is exhausted. The goal of the control strategy is to maintain the remaining heat capacity of all joints above the preset heat threshold through coordinated management.

[0075] Among them, the preset ratio is less than 50%, for example, the preset ratio ranges from 20% to 30%. This corresponds to the joint's remaining heat capacity being less than or equal to the preset heat threshold, indicating that the joint has entered or is about to enter a state of heat depletion and is no longer suitable to continue bearing the current load.

[0076] It is understandable that by setting a preset thermal threshold as a trigger condition, a clear quantitative judgment standard can be provided for the thermal depletion state of the joint, so as to ensure that the load transfer is triggered in time when the joint's thermal capacity is about to be depleted but has not yet reached the dangerous critical point, thereby achieving active thermal management rather than passive overheating response.

[0077] In some embodiments, the electronic device 100 converts the robot system from a first configuration to a second configuration by at least one of the following: shortening the effective length of the limb segment carrying the first joint to reduce the gravitational lever arm on the first joint; positioning the payload near the center of mass of the robot system; converting the robot system to a posture that transmits load using a near-singular joint configuration; and configuring the joints of the robot system to bear at least a portion of the load using skeletal or mechanical restraints.

[0078] For example, shortening the effective length of a load-bearing limb refers to contracting the limb segment that bears the first joint, such as bending an arm or leg to reduce the lever arm length. According to the principle that torque equals force multiplied by the lever arm, a reduced lever arm results in a corresponding reduction in the joint's load torque. Moving the effective load closer to the center of mass means moving the weight towards the robot's body center to reduce the load's eccentricity relative to the joint, thereby reducing the required balancing torque of the joint. Utilizing near-singular joint configurations refers to transitioning to a posture close to a kinematic singularity. Near a singular configuration, the torque requirement of the joint can approach zero, significantly reducing the actuator's output requirements. Utilizing skeletal or mechanical restraints for load bearing means transferring part of the load through the robot's rigid skeleton and mechanical restraint devices, rather than having the actuators bear it entirely, thus utilizing the robot's own structural rigidity to replace the actuator's output.

[0079] It is understandable that different conversion methods are suitable for different scenarios. For example, for overheating of the shoulder joint of a humanoid robot, a combination of shortening the effective limb length (bending the elbow) and bringing the payload closer to the center of gravity (holding the cargo to the chest) can be chosen. For overheating of the leg joint of a quadruped robot, a combination of using a near-singular joint configuration (adjusting the leg posture to a vertical support position) and using mechanical limiting for load bearing can be chosen. This application does not impose any limitations on this.

[0080] It is understandable that configuration transformation can be achieved by shortening the effective length of limbs, bringing the payload closer to the center of mass, utilizing near-singular joint configurations, and using skeletons or mechanical constraints to bear the load. In other words, it reduces the joint load from a geometric and mechanical structural perspective. Multiple methods can be used individually or in combination to adapt to different robot configurations and operating scenarios, effectively reducing the load-bearing burden on joints with insufficient heat capacity.

[0081] It is understood that in this application, "configuration" refers to the body posture or layout pattern formed by the position, angle, and relative spatial relationship between the joints of the robot system.

[0082] Specifically, configuration is a complete description of the state of freedom of all joints of a robot, determining the robot's overall shape, the degree of limb extension, the spatial position of the end effector, and the force relationships between the joints. For example, a humanoid robot "holding a heavy object with its arm outstretched" is one configuration, while "holding the object close to its chest with its arm bent" is another. A quadruped robot "standing posture" is one configuration, while "crawling posture" is another.

[0083] In the technical solution of this application, the purpose of switching from the "first configuration" to the "second configuration" is to change the load distribution of each joint: by adjusting the body posture of the robot system (the position, angle and relative spatial relationship between at least some joints and limbs), the gravitational load torque is transferred from joints with insufficient heat capacity to joints with sufficient heat capacity, or by shortening the lever arm and moving the effective load closer to the center of mass, etc., the torque requirement of a specific joint is reduced.

[0084] Understandable. Figure 1 The implementation process shown is only an illustrative example; in other embodiments, Figure 1 The process shown may include more or fewer steps, and this application does not limit this.

[0085] For example, in some embodiments, the electronic device 100 may also engage a mechanical brake on a joint where the thermal state information meets preset conditions, and partially or completely disconnect the actuator power supply of the joint engaging the mechanical brake.

[0086] In other words, after the configuration conversion is completed and the load has been successfully transferred, the electronic device 100 identifies the joint that can be braked, issues a brake engagement command to lock the joint position with the mechanical brake, and after confirming that the brake has been engaged, cuts off the power supply to the joint actuator or reduces it below the holding current, and the joint enters a passive heat dissipation mode, where heat is dissipated through natural convection and radiation.

[0087] It is understandable that engaging a mechanical brake and partially or completely disconnecting the actuator power supply on a joint that meets preset thermal conditions can allow the overheated joint to enter a zero-power or low-power thermal recovery mode, eliminating the source of resistive heat generation, achieving passive heat dissipation at the maximum rate, and reducing overall energy consumption.

[0088] In some embodiments, the engagement and disengagement of the mechanical brakes are coordinated with the robot system's posture transitions in time. Furthermore, the actuators of the corresponding joints are re-energized before the mechanical brakes disengage.

[0089] In other words, brake engagement occurs after the configuration change is complete and the load has been successfully transferred, while brake disengagement occurs before the next configuration change begins. Within a certain time window before brake disengagement, the actuator is first powered back on to enable the joint to output torque, and then the brake is disengaged. To prevent load drop, a certain overlap period can be set between brake disengagement and the actuator fully establishing torque. The length of this overlap period can be dynamically adjusted according to the actuator's response characteristics.

[0090] It is understandable that the engagement / disengagement of the mechanical brake and the attitude transition action are coordinated in time, and the actuator is re-energized before the brake disengages. This ensures that the joint does not undergo thermal recovery when locked and is energized in advance when movement is required. This achieves seamless integration of thermal management and motion control, thereby avoiding load drop or attitude loss caused by delayed brake disengagement or delayed actuator energization.

[0091] For example, in some embodiments, when the robot system is in the second configuration, the thermal state information of each joint of the robot system is detected; when the thermal state information of the current load-bearing joint of the robot system meets the preset conditions, the robot system is switched from the second configuration to the third configuration or the first configuration.

[0092] In other words, in the second configuration, the previously overheated joint is passively cooling itself through brake locking, while the second joint, bearing the load, gradually deteriorates in its thermal state due to continuous load. When the thermal state of the second joint also reaches a preset condition, the electronic device 100 determines whether to proceed with the next cycle. If the thermal capacity of the first joint has recovered to a sufficient level, the load is transferred back to the first joint, i.e., the configuration reverts to the first configuration; if the first joint has not fully recovered, the load is transferred to the third joint, i.e., the configuration reverts to the third configuration. This process is repeated, forming a cyclical pattern of alternating load-bearing and recovery among multiple joints.

[0093] It is understandable that after the robot is in the second configuration, it continues to monitor the thermal state of each joint, and triggers a new configuration conversion when the thermal state of the current load-bearing joint meets the preset conditions. That is, a cyclic heat redistribution mechanism is formed, so that the load is transferred sequentially between different joint groups, realizing the alternating consumption and recovery of heat capacity, thereby maintaining long-term operation under load conditions that exceed the continuous torque capacity.

[0094] It can be understood that, through the above method, the operation process of the robot system can be divided into four stages: explosive lifting stage, posture transformation stage, continuous holding stage with brake engagement, and cyclic heat redistribution stage.

[0095] During the burst lifting phase, selected joints in the primary load-bearing kinematic chain (as an example of at least one first joint) reach peak torque to initiate the lifting of the heavy-duty payload. The thermal budget of each participating joint begins to deplete at a rate proportional to the square of the joint torque output. Electronics 100 monitors the thermal budget of all participating joints in real time and calculates the remaining time window before the joint with the highest thermal stress reaches its depletion threshold. This time window is related to the maximum duration of the burst phase.

[0096] During the posture transition phase, the body posture transition is initiated before the thermal budget (remaining thermal capacity) of any joint is less than or equal to a preset thermal threshold. For example, the primary load-bearing responsibility is transferred from a heat-depleted joint (first joint) to a heat-sufficient joint (second joint) that is not bearing weight or is lightly loaded in the first configuration. Alternatively, the effective lever arm on thermally stressed joints is reduced by reconfiguring the kinematic chain geometry (e.g., shrinking limb segments to reduce lever arm length, repositioning the payload closer to the robot's center of mass). Alternatively, a configuration is transitioned to one where the skeletal structure and mechanical advantages bear a larger proportion of the load, reducing actuator torque requirements. Alternatively, the heat-depleted joint (first joint) is positioned in a configuration with enhanced cooling exposure.

[0097] During the sustained holding phase with brakes engaged, the robot control system maintains the payload in the post-transition posture. Joints that are currently unloaded or lightly loaded during the burst phase can engage their corresponding mechanical brakes.

[0098] During the cyclic heat redistribution phase, the thermal budgets of all joints are continuously monitored, and further attitude transitions are initiated as needed. As the thermal budgets of initially depleted joints recover (e.g., joints with brake engagement) and the current load-bearing joints approach depletion, new attitude transitions are triggered.

[0099] For example, Figures 2A to 2D According to an embodiment of this application, a schematic diagram of the sequence of body posture changes of a humanoid robot during the operation of carrying a 40 kg box is shown.

[0100] Figure 2A The initial configuration of the explosive lifting phase is shown. The humanoid robot lifts the box from the ground with its outstretched arms extended. Figure 2A In the configuration shown, the humanoid robot's arms are fully extended, and the lever arms of the shoulder and elbow joints are relatively long, requiring high torque output (marked as peak torque in the figure, reaching 4 times the continuous rated torque). During the explosive lifting phase, the peak torque capability of the joints is used to quickly lift the box off the ground.

[0101] Figure 2B This illustrates the close-fitting configuration after a posture change. The humanoid robot shifts the box from an extended-arm position to a chest-hugging position, with its arms bent and the box pressed tightly against its body. Figure 2B In the configuration shown, the lever arms of the shoulder and elbow joints are shortened (labeled as short lever arms in the figure), and the required maintaining torque is reduced (labeled as reduced to 1.5 times the continuous rated torque in the figure). Furthermore, this close-fitting configuration helps to expose heat-depleting joints to a better heat dissipation environment.

[0102] Figure 2C The diagram illustrates a continuous holding configuration with the brakes engaged. In this close-fitting configuration, the humanoid robot's shoulder and elbow joints, having exhausted their heat, engage the mechanical brakes. The areas labeled "Brake Engagement" and "Heat Dissipation" indicate that these joints are passively cooling themselves without current. The robot maintains its posture with zero power consumption.

[0103] Figure 2D The diagram illustrates the alternating limb load transfer during the cyclic thermal redistribution phase. Once the joint thermal budget of the humanoid robot's left arm is restored, the load is transferred from the right to the left, achieving alternating load-bearing between the left and right arms. The diagram labels the switching processes of "alternating left and right arms" and "brake engagement / release."

[0104] It is understandable that the above control method can be implemented through a neural network model trained by reinforcement learning.

[0105] For example, Figure 3 A flowchart illustrating a model training method is shown according to an embodiment of this application. It can be understood that... Figure 3 The flowchart shown is executed by either electronic device 100 or a simulation server (the following explanation will still use electronic device 100 as an example). For ease of description, the following will refer to... Figure 3 When describing the flowchart shown, the execution subject of the flowchart will not be repeated.

[0106] like Figure 3 As shown, this process includes, but is not limited to:

[0107] S301: Construct a physical simulation environment.

[0108] In some embodiments, the electronic device 100 constructs a physical simulation environment for the rigid body dynamics and joint-by-joint thermodynamics of the robot system. The rigid body dynamics section simulates the robot's kinematic chains, joint actuation, contact dynamics, ground friction, and gravitational effects. The thermodynamics section configures a thermal budget state variable for each simulated joint actuator, which serves as a function of joint torque output, temperature-dependent dissipation, and body posture configuration.

[0109] In some embodiments, the thermal budget dynamics of each joint follow the above formula (1), which will not be elaborated here.

[0110] S302: Configure domain randomization parameters.

[0111] In some embodiments, at the start of each training round, the electronic device 100 randomly selects the following parameters within a preset range: the heat generation coefficient αᵢ varies randomly within the range of ±20% to ±30%; the heat dissipation coefficient βᵢ varies randomly within a preset range; the ambient temperature varies randomly within the range of 10 degrees Celsius (°C) to 40°C; the actuator efficiency degradation curve varies randomly with temperature; the initial thermal budget state of each joint is randomly set between full recovery and partial depletion; and the mass and center of mass of the payload vary randomly within a preset range.

[0112] It is understandable that multi-level domain randomization can enhance the adaptability of the trained model to various uncertainties in the real physical environment, thereby improving the transfer effect from simulation to reality.

[0113] S303: Initialize training rounds.

[0114] In some embodiments, the electronic device 100 randomly initializes the robot system's posture, payload parameters, and initial thermal budget states of each joint, and resets the simulation environment.

[0115] In some embodiments, the initialization range covers a variety of possible working postures and load conditions of the robot system to enhance the generalization capability of the strategy.

[0116] S304: Observation status.

[0117] In some embodiments, at each time step, a state observation vector is obtained from the simulation environment. The state observation vector includes: the position angle, velocity, current output torque, normalized thermal budget level (range [0,1]), mechanical brake engagement state, payload mass and center of mass position, robot overall attitude parameters (including center of mass position, orientation and angular velocity), ground contact force and inertial measurement unit data.

[0118] S305: Perform the action.

[0119] In some embodiments, a motion vector is output based on the current state observation. The motion vector includes: target position or target torque for each joint; brake engagement or disengagement commands for each joint equipped with a mechanical brake; and target body configuration parameters for attitude transition.

[0120] S306: Update simulation status.

[0121] In some embodiments, the kinematic state, dynamic state, and thermal budget state of each joint of the robot system are updated based on motion commands.

[0122] S307: Calculate the reward value.

[0123] In some embodiments, the reward function R(s, a) takes a multi-objective composite form.

[0124] For example, the reward function includes five components: lift reward (R_lift), which provides positive incentives based on the payload's height achievement and positional stability; balance reward (R_balance), which penalizes deviations of the centroid from the support polygon and excessive angular velocity amplitude; thermal budget reward (R_thermal), which employs a two-layer shape-based penalty structure; energy efficiency reward (R_efficiency), which provides negative incentives based on the total electrical energy consumption of all joints; and posture reward (R_pose), which imposes a small penalty on behaviors that deviate from the preferred posture.

[0125] The two-layer structure of the thermal budget reward is as follows: The first layer is a soft quadratic penalty, defined as the penalty value being proportional to the square of the preset thermal threshold when the thermal budget of each joint is less than the preset thermal threshold. The second layer is a hard penalty, defined as a larger fixed penalty value being set when the thermal budget of any joint is less than the critical thermal limit (usually set to 5% of the maximum thermal capacity).

[0126] It is understandable that soft-penalty guided models take mitigation measures such as attitude transition or brake engagement before the thermal limit is reached, while hard-penalty guided models force the avoidance of any behavior that may lead to overheating risks.

[0127] For example, the reward function R can be referred to the following formula (3).

[0128] Formula (3)

[0129] in, To increase the weight of the reward, To balance the weighting of the rewards, The weight corresponding to the heat budget reward. The weighting corresponding to the energy efficiency reward. R_lift represents the weights corresponding to the attitude reward, R_balance represents the lift reward, R_thermal represents the thermal budget reward, R_efficiency represents the energy efficiency reward, and R_pose represents the attitude reward.

[0130] For example, the penalty structure for the thermal budget reward (R_thermal) can be referenced in the following formula (4).

[0131] Formula (4)

[0132] Among them, λ is the weight coefficient (a positive number) of the soft penalty, which is used to control the intensity of the soft quadratic penalty. The larger λ is, the greater the penalty the agent feels when the thermal budget is below the threshold, and thus it will more actively avoid too low thermal budget. B_threshold is the depletion threshold (preset thermal threshold), usually set to 20% to 30% of the maximum heat capacity. When the joint thermal budget is less than this threshold, it indicates that the joint has entered or is about to enter the thermal depletion state. Bi(t) is the remaining heat capacity of joint i at time t, usually normalized to the interval [0,1] (0 means completely depleted, 1 means completely recovered). μ is the weight coefficient (a positive number) of the hard penalty, which is used to control the intensity of the hard penalty. μ is usually set to be relatively large to make the model avoid triggering the hard penalty condition. δ(Bi(t)<B_critical) is an indicator function. When Bi(t) < B_critical, δ = 1; otherwise δ = 0. This function is used to judge whether the thermal budget of any joint is below the critical limit. B_critical is the critical thermal limit, usually set to about 5% of the maximum heat capacity. When the joint thermal budget is less than this value, it means that an unacceptable thermal risk state has been reached.

[0133] S308: Store the experience and update the model parameters.

[0134] In some embodiments, the experience data composed of the current state, action, reward, and next state is stored in the experience pool. Every preset number of time steps, a batch of experience data is sampled from the experience pool, and the policy gradient is calculated using the reinforcement learning algorithm to update the model parameters.

[0135] For example, proximal policy optimization is used as the training algorithm. Alternatively, algorithms suitable for continuous control tasks such as flexible actor-critic are used. This application does not limit the specific training method.

[0136] S309: Judge whether the current episode meets the termination condition.

[0137] Among them, the termination conditions include: the payload is lifted to the target height and maintained for a preset duration (task success); the payload drops or the robot falls (task failure); the thermal budget of any joint drops below the critical limit (thermal budget violation); or, the maximum number of time steps in a single episode is reached.

[0138] In some embodiments, if the judgment result is no, then step S304 is executed; if the judgment result is yes, then S310 is executed.

[0139] S310: Judge whether the training has reached the convergence condition.

[0140] The convergence conditions include: the average reward value is stable or grows slowly; the frequency of hot budget violations drops to near zero; the task success rate reaches the preset target (e.g., above 90%); and the hot budget curves of each joint exhibit a regular periodic oscillation pattern.

[0141] In some embodiments, if the determination result is negative, step S303 is executed; if the determination result is positive, step S311 is executed.

[0142] S311: Output the trained model.

[0143] It is understandable that through the above training process, the model can gradually transition from random exploratory behavior to strategic operational behavior, spontaneously emerging a coordinated pattern similar to that of human weightlifters: explosive lifting, posture transition, brake engagement, and cyclic heat redistribution.

[0144] It is understood that the types of electronic devices 100 include, but are not limited to: embedded control systems integrated into the robot body, independent industrial control computers, robot external servers or cloud computing platforms, edge computing devices, and general-purpose computers or workstations used in the simulation training phase.

[0145] Electronic device 100 may also be a robot body. The types of robots include, but are not limited to: humanoid robots, quadruped robots, mobile manipulation robots, hexapod robots and other multi-legged robots, as well as any multi-joint robot system with redundant degrees of freedom that can support the same payload in multiple kinematic configurations. This application does not limit the specific type of robot.

[0146] It is understood that during the simulation training phase, the electronic device 100 can be a general-purpose computer, workstation, or high-performance computing cluster used to run the physical simulation environment and reinforcement learning training algorithms. After training is completed, the trained model parameters can be exported and deployed to the embedded controller of the robot body or an industrial control computer to perform inference. This application does not limit the specific implementation of the electronic device 100; as long as it can perform the functions of thermal state detection, thermal budget calculation, configuration conversion decision, and brake control described in this application, it falls within the scope of the electronic device 100 described in this application.

[0147] For example, Figure 4 A schematic diagram of the structure of a robot system 200 is shown according to an embodiment of this application.

[0148] like Figure 4 As shown, the robot system 200 includes a robot platform 210 with multiple articulated joints (J1, J2, ... Jn) and a central control unit 220.

[0149] Each joint in the robot platform 210 is driven by an electric actuator (A1, A2, ... An), enabling each joint to generate torque in at least two operating modes: a continuous rated torque (τ_cont) that can be sustained indefinitely, and a peak torque (τ_peak), where τ_peak is greater than or equal to 2 × τ_cont, and the range of τ_peak is typically from 3 × τ_cont to 5 × τ_cont, with the duration determined by the thermal state of the actuator.

[0150] Each actuator (Ai) is equipped with a thermal sensor (Ti) that provides real-time temperature measurement or temperature estimation via a thermal observer model, and an optional mechanical brake or holding mechanism (Bi) that can maintain the joint position without current input.

[0151] The central control unit 220 receives thermal sensor data, kinematic state data (joint position, velocity), and payload estimation data from force / torque sensors or based on current estimation, and executes a thermal budget sensing control strategy.

[0152] It is understood that the form of robot platform 210 includes, but is not limited to: humanoid robots with two legs and two arms, quadruped robots equipped with one or more manipulators, hexapod robots with manipulator attachments, mobile manipulator robots with articulated bases, or any multi-joint robot system with redundant degrees of freedom that can support the same payload in multiple kinematic configurations.

[0153] Understandable. Figure 4 The structure of the robot system 200 shown is only an example. In other embodiments, the robot system 200 may include more or fewer modules, which is not limited here.

[0154] It is understood that if the robot platform 210 consists of multiple robots, the central control unit 220 may be a processing unit configured outside the robot platform 210. If the robot platform 210 is a multi-joint robot, the central control unit 220 may be a processing unit configured inside the robot platform 210, such as a processor integrated into the robot.

[0155] For example, Figure 5 A schematic diagram of the hardware structure of an electronic device 100 is shown according to an embodiment of this application.

[0156] like Figure 5 As shown, the electronic device 100 includes one or more (only one is shown in the figure) processors 110, memory 120, communication interface 130, and bus 140. The processors 110, memory 120, and communication interface 130 are interconnected via the bus 140.

[0157] In some embodiments, the processor 110 may include one or more processing units, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU), or an application-specific integrated circuit, for executing related programs to achieve the functions required by the modules in the compilation apparatus of this application embodiment, or to execute the control method of this application method embodiment.

[0158] The memory 120 may include one or more memories for storing data or one or more applications. The memory may be read-only memory (ROM), static storage device, dynamic storage device, random access memory (RAM), high-speed random access memory, double data rate synchronous dynamic RAM (DDR), high bandwidth memory (HBM), or non-volatile memory, such as at least one disk storage device, flash memory device, universal flash storage (UFS), etc.

[0159] The processor 110 can also be an integrated circuit chip with signal processing capabilities. In implementation, each step of the data processing method of this application can be completed by software instructions in the processor 110. The aforementioned processor 110 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit, a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in the memory 120. The processor 110 reads the information in the memory 120 and, in conjunction with its hardware, performs the functions required by the units included in the compilation apparatus of this application embodiment, or executes the control method of this application method embodiment.

[0160] Communication interface 130 is used to enable communication between electronic device 100 and other devices or communication networks. Communication interface 130 may include wired or wireless communication interfaces, so that electronic device 100 can access the Internet via wired or wireless means, and obtain data from or send data to other devices based on the Internet.

[0161] Bus 140 is used to connect processor 110, memory 120, communication interface 130 and other possible modules or circuits.

[0162] In some embodiments, the electronic device 100 can be used to execute the control method corresponding to the method embodiment described above. To avoid repetition, it will not be described again here.

[0163] Understandable. Figure 5 The structure of the electronic device 100 shown is only an example. In other embodiments, the electronic device 100 may include more or fewer modules, which is not limited here.

[0164] This application also provides a computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set, which is loaded and executed by a processor to implement the control methods provided in the above-described method embodiments.

[0165] This application also provides a program product that includes instructions that, when executed by a device, enable the device to implement the control method provided in this application.

[0166] In some embodiments, a chip is also provided, the chip including a processor coupled to a memory for executing a computer program or instructions stored in the memory, such that the chip implements the control methods provided in the above-described method embodiments.

[0167] In the accompanying drawings, some structural or methodological features may be shown in a specific arrangement and / or order. However, it should be understood that such a specific arrangement and / or order may not be necessary. Rather, in some embodiments, these features may be arranged in a manner and / or order different from that shown in the illustrative drawings. Furthermore, the inclusion of structural or methodological features in a particular figure does not imply that such features are required in all embodiments, and in some embodiments, these features may be omitted or may be combined with other features.

[0168] It should be noted that all units / modules mentioned in the device embodiments of this application are logical units / modules. Physically, a logical unit / module can be a physical unit / module, a part of a physical unit / module, or a combination of multiple physical units / modules. The physical implementation of these logical units / modules themselves is not the most important factor; the combination of functions implemented by these logical units / modules is the key to solving the technical problems proposed in this application. Furthermore, to highlight the innovative aspects of this application, the above-described device embodiments of this application have not introduced units / modules that are not closely related to solving the technical problems proposed in this application. This does not mean that the above-described device embodiments do not contain other units / modules.

[0169] It should be noted that in the examples and description of this patent, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0170] Although this application has been illustrated and described with reference to certain preferred embodiments thereof, those skilled in the art should understand that various changes in form and detail may be made thereto without departing from the spirit and scope of this application.

Claims

1. A control method, characterized in that, Applied to electronic devices, the method includes: The robot system is detected to be operating in a preset torque range corresponding to the at least one first joint, wherein the minimum value of the preset torque range corresponding to each first joint is greater than the continuous rated torque value of the first joint. The thermal state information of each joint in the robot system is detected, wherein the thermal state information is used to indicate the remaining heat capacity of the corresponding joint during the execution of the task; If the thermal state information of at least one of the at least one first joints meets a preset condition, the robot system is switched from a first configuration to a second configuration to transfer at least a portion of the load from the joint whose thermal state information meets the preset condition to at least one second joint of the robot system; wherein, The thermal state information of at least one of the at least one first joints satisfies preset conditions, including: The remaining heat capacity of at least one joint is less than or equal to a preset heat threshold, which is a preset proportion of the maximum heat capacity of the at least one joint corresponding to the ambient temperature of the robot system.

2. The method according to claim 1, characterized in that, The thermal state information is determined based on at least one of the following parameters: Real-time temperature of the joint; Historical torque or historical drive current of the joint; The current torque or current drive current of the joint; The current posture of the limb segment where the joint is located; The thermal threshold of a joint.

3. The method according to claim 1, characterized in that, The detection of at least one first joint of the robot system operating within a preset torque range corresponding to the at least one first joint includes: The torque value of the at least one first joint is detected to be greater than the continuous rated torque value of the at least one first joint; or, The drive current value of the at least one first joint is detected to be greater than the current threshold corresponding to the continuous rated torque of the at least one first joint.

4. The method according to claim 1, characterized in that, The conversion of the robot system from the first configuration to the second configuration includes at least one of the following: Shorten the effective length of the limb segment supporting the first joint to reduce the gravitational lever arm on the first joint; Position the payload close to the center of mass of the robot system; The robot system is then switched to a posture that utilizes a near-singular joint configuration to transmit load-bearing forces. The joints of the robotic system are configured to bear at least part of the load using bones or mechanical restraints.

5. The method according to claim 1, characterized in that, The method further includes: A mechanical brake is engaged on the joint where the thermal state information satisfies the preset conditions. Partially or completely disconnect the actuator power supply to the joint of the mechanical brake.

6. The method according to claim 5, characterized in that, The engagement and disengagement of the mechanical brake are coordinated with the posture transition actions of the robot system in time; and... Before the mechanical brake disengages, the actuator of the corresponding joint is re-energized.

7. The method according to claim 1, characterized in that, The method further includes: Corresponding to the robot system being in the second configuration, the thermal state information of each joint of the robot system is detected; If the thermal state information of the current load-bearing joint of the robot system meets the preset conditions, the robot system will be converted from the second configuration to the third configuration or the first configuration.

8. The method according to claim 1, characterized in that, The robotic system includes one or more robots; and... The robot includes at least one of the following types: humanoid robot, quadruped robot, and mobile operation robot.

9. The method according to claim 1, characterized in that, The robot system includes a robot comprising multiple joints, and the at least one first joint and the at least one second joint are joints within the robot; or, The robot system includes at least two robots, the at least one first joint is a joint located in the same robot or in different robots, the at least one second joint is a joint located in the same robot or in different robots, and the at least one first joint and the at least one second joint are located in the same robot or in different robots.

10. An electronic device, characterized in that, It includes a processor and a memory, the memory being used to store program instructions, and the processor being used to invoke the program instructions to perform the method of any one of claims 1 to 9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed, implement the method of any one of claims 1 to 9.

12. A computer program product, characterized in that, The computer program product includes instructions that, when executed, implement the method of any one of claims 1 to 9.

Images