Edge trigger motion control method and system for joints of upper limbs of humanoid robot

Abstract

The application discloses a kind of edge trigger action control method and system of humanoid robot upper limb joint, the edge trigger action control method of humanoid robot upper limb joint includes the following steps: step S1, current action instruction issued by host computer is received by communication interface;Step S2, the current action instruction is compared with the last executed action instruction cached by edge trigger, and comparison result is obtained;Step S3, when the comparison result represents that the current action instruction is different from the last executed action instruction, then it is determined as rising edge trigger, and the current action instruction is issued.The application provides a kind of edge trigger action control method and system of humanoid robot upper limb joint, introduces hardware level edge trigger comparison mechanism in instruction issuing path, only triggers issuing at the moment of instruction change, filters the same instruction of repeated issuing, while maintaining the real-time performance of instruction response, and decouples with host computer software.

Description

Technical Field

[0001] This invention relates to a method and system for edge-triggered motion control of upper limb joints in a humanoid robot, specifically a method and system for edge-triggered motion control of upper limb joints in a humanoid robot. Background Technology

[0002] Currently, motion control of the upper limb joints of humanoid robots is typically achieved by issuing motion commands from a host computer. In scenarios with high-frequency command issuance, the same motion command may be issued repeatedly, causing the upper limb joint motion commands to be repeatedly triggered, resulting in motion jitter or even abnormal motion. Existing technologies mainly use software debouncing, timer debouncing, and state machines to process motion commands, but all have certain limitations. Software debouncing logic is highly coupled with business logic, and modifying the debouncing strategy requires changes to the core control logic; software processing has unpredictable latency, and may miss command changes in high-frequency issuance scenarios. Timer debouncing solutions, on the other hand, increase motion response latency and reduce the real-time performance of human-computer interaction.

[0003] In addition, a robot anti-shake method based on the VLA (Vision-Language-Action) model has emerged. Its core idea is to acquire the robot's current observation image before the VLA model initiates the current inference round, obtain a subset of action sequences from the current moment to the first moment from the action sequence predicted in the previous inference round, predict the observation image for the first moment based on the current observation image and the action sequence subset, and finally call the VLA model to perform the current inference process based on the observation image for the first moment, obtaining the predicted action sequence for this round and controlling the robot to switch actions. This method has the following shortcomings: 1. It relies on the inference prediction accuracy of the VLA model, and the model inference itself has uncertainties and delays, making it impossible to guarantee hard real-time response; 2. Predicting the observation image requires additional computing resources and GPU acceleration, increasing system cost and power consumption; 3. Software-level anti-shake processing cannot cover the hardware-level problem of repeated command issuance, and shaking may still occur in high-frequency command issuance scenarios.

[0004] Therefore, how to avoid motion jitter caused by repeated command issuance and maintain the real-time performance of command response in high-frequency command issuance scenarios is an urgent problem to be solved in this field. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a method and system for edge-triggered motion control of the upper limb joints of a humanoid robot. The method introduces a hardware-level edge-triggered comparison mechanism in the command issuance path, triggers the issuance only at the moment of command change, filters out repeated issuance of the same command, maintains the real-time performance of command response, and decouples it from the host computer software.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: This invention provides a method for edge-triggered motion control of upper limb joints in a humanoid robot, comprising the following steps: Step S1: Receive the current action command issued by the host computer through the communication interface; Step S2: Compare the current action instruction with the cached previously executed action instruction using edge triggering to obtain the comparison result; Step S3: When the comparison result indicates that the current action instruction is different from the previously executed action instruction, it is determined to be a rising edge trigger, the current action instruction is sent, and the current action instruction is updated to the cache as the new previously executed action instruction; when the comparison result indicates that the current action instruction is the same as the previously executed action instruction, a default empty instruction is sent.

[0007] Furthermore, after the current action command is issued, an action timer is used to keep track of the time; when the action timer expires, a default reset command is issued to reset the current action of the upper limb joint.

[0008] Furthermore, the edge-triggered comparison in step S2 specifically includes the following steps: The current action instruction is compared bit by bit with the previously executed action instruction using an XOR gate comparison circuit. The XOR results of each bit are then combined and output. A trigger signal is output when the combined result is non-zero, and a hold signal is output when the combined result is zero.

[0009] Another aspect of the present invention provides a system that applies the edge-triggered motion control method for the upper limb joint of the humanoid robot, including an instruction generation module, an action instruction cache register, an edge-triggered comparator, an action timer, a communication module, a robot motion control board, and an upper limb joint; The instruction generation module is used to issue the current action instruction; The action instruction cache register is used to cache the previously executed action instruction; The edge-triggered comparator is used to perform an edge-triggered comparison between the current action instruction and the previously executed action instruction cached in the action instruction cache register, and outputs a trigger signal when the current action instruction is different from the previously executed action instruction, and outputs a hold signal when the current action instruction is the same as the previously executed action instruction. The communication module is used to acquire the current action command at a preset frequency, and send the current action command or default empty command to the robot motion control board according to the output of the edge-triggered comparator, so as to drive the upper limb joint to perform the action; The action timer is used to start timing after the action command is issued, and to trigger the issuance of a default reset command after the timing expires.

[0010] Furthermore, the communication module is a DDS communication module.

[0011] Furthermore, the preset frequency is 500Hz.

[0012] Furthermore, the edge-triggered comparator is an XOR gate comparator circuit.

[0013] Furthermore, the action timer is a reloadable counter.

[0014] By adopting the above technical solution, the present invention has the following beneficial effects: This invention compares the current action command with the previously executed action command using a hardware-level edge-triggered comparator, triggering the command only on the rising edge of the command change. Identical commands are filtered out during the command transmission period and are not triggered repeatedly, thus avoiding motion jitter and abnormalities caused by repeated command transmission.

[0015] 2. This invention places the edge-triggered comparison mechanism at the hardware level, completely decoupling it from the host computer software. It can achieve instruction debouncing without modifying the business code, which helps to reduce the coupling between the debouncing logic and the business code, and also avoids the problem of unpredictable delay in software debouncing processing.

[0016] 3. The hardware-level XOR comparator used in this invention can complete the comparison operation within a single clock cycle, achieving nanosecond-level instruction change detection delay, avoiding unpredictable delays caused by software debouncing, and not affecting the real-time performance of high-frequency instruction issuance.

[0017] 4. The motion timer used in this invention provides configurable motion duration, which can support differentiated duration settings for different motion types. This is beneficial for achieving differentiated control of different motions and avoids increased motion response delay caused by timer debouncing.

[0018] 5. Compared with software debouncing schemes based on model inference prediction, this invention adopts a hardware edge detection mechanism, does not rely on model inference accuracy, does not have the problem of prediction error accumulation, does not require additional model inference computing resources, has a faster response speed, and the system is more stable and reliable. Attached Figure Description

[0019] Figure 1 This is a flowchart of the edge-triggered motion control method for the upper limb joints of the humanoid robot of the present invention; Figure 2 This is a schematic diagram of the overall architecture of the system that applies the edge-triggered motion control method of the humanoid robot upper limb joint of the present invention; Figure 3 This is a circuit schematic diagram of the edge-triggered comparator of the present invention; Figure 4This is a comparison waveform diagram of the edge-triggered and level-triggered signals of the present invention. Detailed Implementation

[0020] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0021] Example 1 like Figure 1 As shown, this embodiment provides a method for edge-triggered motion control of the upper limb joints of a humanoid robot, including the following steps: Step S1: Receive the current action command issued by the host computer through the communication interface.

[0022] Step S2: Compare the current action instruction with the cached previously executed action instruction using edge-triggered comparison to obtain the comparison result. The edge-triggered comparison specifically includes the following steps: The current action instruction is compared bit by bit with the previously executed action instruction using an XOR gate comparison circuit. The XOR results of each bit are then combined and output. A trigger signal is output when the combined result is non-zero, and a hold signal is output when the combined result is zero.

[0023] Edge-triggered comparison is performed at the hardware level, generating a single trigger only when the instruction changes. Consecutive identical instructions are filtered out, preventing duplicate triggers.

[0024] Step S3: When the comparison result indicates that the current action instruction is different from the previously executed action instruction, it is determined to be triggered by a rising edge, the current action instruction is sent, and the current action instruction is updated to the cache as the new previously executed action instruction; when the comparison result indicates that the current action instruction is the same as the previously executed action instruction, a default empty instruction is sent to filter out duplicate sent identical action instructions.

[0025] In this embodiment, after the current action command is issued, an action timer is used to keep track of the time. When the action timer expires, a default reset command is issued to reset the current action of the upper limb joint.

[0026] Example 2 like Figure 2 As shown, this embodiment provides a system that applies the edge-triggered motion control method for the upper limb joint of a humanoid robot as described in Embodiment 1, including an instruction generation module, an action instruction cache register, an edge-triggered comparator, an action timer, a communication module, a robot motion control board, and an upper limb joint.

[0027] The instruction generation module is used to issue the current action instruction.

[0028] The action instruction cache register uses 4 bytes of non-volatile Flash to cache the last executed action instruction.

[0029] The edge-triggered comparator is used to perform an edge-triggered comparison between the current action instruction and the previously executed action instruction cached in the action instruction cache register. It outputs a trigger signal when the current action instruction is different from the previously executed action instruction, and outputs a hold signal when the current action instruction is the same as the previously executed action instruction.

[0030] The communication module is a DDS communication module, which is used to acquire the current action command at a preset frequency of 500Hz, and send the current action command or default empty command to the robot motion control board according to the output of the edge-triggered comparator to drive the upper limb joints to perform actions.

[0031] The action timer is used to start timing after an action command is issued, and to trigger the issuance of a default reset command after the timing expires.

[0032] In this embodiment, the edge-triggered comparator is an XOR gate comparison circuit, such as... Figure 3 As shown, a CMOS XOR gate chip of model 74HC86 is used. The input receives two 8-bit parallel data streams, namely the currently pending instruction A[7:0] and the already executed cached instruction B[7:0]. An edge-triggered comparator performs a bit-by-bit XOR comparison on the corresponding bits of the two parallel data streams, obtaining 8 comparison results, which are then merged to obtain the final edge detection result. When the two data streams are different, the merged result is non-zero, and a high-level trigger signal is output to enable the communication module to issue the current action instruction; when the two data streams are the same, the merged result is zero, and a low-level hold signal is output to issue a default empty instruction. Figure 3 As shown in the timing diagram, when the instruction changes, the comparison result is non-zero, and the trigger signal generates a high-level pulse, corresponding to a rising edge trigger; consecutive identical instructions are filtered out and no trigger is generated.

[0033] A comparative test of edge-triggered and level-triggered methods is as follows: Figure 4 As shown in the figure, the upper half represents traditional level triggering, and the lower half represents edge triggering in this embodiment. The test conditions are a 500Hz command issuance frequency, with the same command lasting for 1 second. The example command is a hand-waving gesture. Under traditional level triggering, the same command is repeatedly issued during the valid high-level duration, causing the upper limb joint to be interrupted approximately 30 times per second. This repeated interruption and restarting results in incomplete and intermittent movements, with continuous upper limb shaking. With the edge triggering method of this invention, triggering occurs only once on the rising edge of the command change. The same command is automatically filtered out and not repeatedly triggered during its duration, resulting in only one interruption per second for the upper limb joint, allowing for a smooth and complete completion of the movement. Regarding the delay from the command layer to the execution layer, the average delay for level triggering is 5.2ms, while the average delay for this embodiment is 0.8ms.

[0034] The action timer in this embodiment is a reloadable counter. The timing duration of the action timer can be preset to different values ​​according to the action type. Different timing durations can be set according to the action type. For example, the waving action WAVE corresponds to 2 seconds, the grabbing action GRAB corresponds to 3 seconds, and the pointing action POINT corresponds to 5 seconds.

[0035] The specific embodiments described above further illustrate the technical problems, technical solutions, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for edge-triggered motion control of upper limb joints in a humanoid robot, characterized in that, Includes the following steps: Step S1: Receive the current action command issued by the host computer through the communication interface; Step S2: Compare the current action instruction with the cached previously executed action instruction using edge triggering to obtain the comparison result; Step S3: When the comparison result indicates that the current action instruction is different from the previously executed action instruction, it is determined to be a rising edge trigger, the current action instruction is sent, and the current action instruction is updated to the cache as the new previously executed action instruction; when the comparison result indicates that the current action instruction is the same as the previously executed action instruction, a default empty instruction is sent.

2. The edge-triggered motion control method for the upper limb joints of a humanoid robot according to claim 1, characterized in that: After the current action command is issued, an action timer is used to keep track of the time. When the action timer expires, a default reset command is issued to reset the current action of the upper limb joint.

3. The edge-triggered motion control method for the upper limb joints of a humanoid robot according to claim 1, characterized in that: The edge-triggered comparison in step S2 specifically includes the following steps: The current action instruction is compared bit by bit with the previously executed action instruction using an XOR gate comparison circuit. The XOR results of each bit are then combined and output. A trigger signal is output when the combined result is non-zero, and a hold signal is output when the combined result is zero.

4. A system employing the edge-triggered motion control method for the upper limb joints of a humanoid robot as described in any one of claims 1 to 3, characterized in that: It includes an instruction generation module, an action instruction buffer register, an edge-triggered comparator, an action timer, a communication module, a robot motion control board, and upper limb joints; The instruction generation module is used to issue the current action instruction; The action instruction cache register is used to cache the previously executed action instruction; The edge-triggered comparator is used to perform an edge-triggered comparison between the current action instruction and the previously executed action instruction cached in the action instruction cache register, and outputs a trigger signal when the current action instruction is different from the previously executed action instruction, and outputs a hold signal when the current action instruction is the same as the previously executed action instruction. The communication module is used to acquire the current action command at a preset frequency, and send the current action command or default empty command to the robot motion control board according to the output of the edge-triggered comparator, so as to drive the upper limb joint to perform the action; The action timer is used to start timing after the action command is issued, and to trigger the issuance of a default reset command after the timing expires.

5. The system according to claim 4, characterized in that: The communication module is a DDS communication module.

6. The system according to claim 4, characterized in that: The preset frequency is 500Hz.

7. The system according to claim 4, characterized in that: The edge-triggered comparator is an XOR gate comparator circuit.

8. The system according to claim 4, characterized in that: The action timer is a reloadable counter.

Images