Robot motion control method and system

Abstract

The invention relates to the technical field of robot motion control, in particular to a robot motion control method and system.The robot motion control method comprises the following steps that in the process that a port filter is screwed to a brewing head, counter-acting force information between the port filter and the brewing head is obtained through a force sensor on a mechanical arm tail end executor; meanwhile, angle information of the mechanical arm in the screwing-in process of the port filter is obtained; according to the counter-acting force information and the angle information, the actual state of the port filter and the interface of the brewing head is evaluated, and locking mechanical parameters and screwing-in motion parameters are adjusted; and in the screwing process, when the force sensor feeds back that abnormal resistance exists, posture adjustment of the mechanical arm is triggered, and the screwing-in action is tried again. The screw-in accuracy and stability can be improved.

Description

Technical Field

[0001] This invention relates to the technical field of robot motion control, and specifically to a robot motion control method and system. Background Technology

[0002] In the modern business environment, coffee robots, as an important tool for automated coffee making, face several uncertainties that make it difficult for them to consistently and accurately produce high-quality espresso. First, the natural variability of coffee beans, such as differences in density, moisture content, particle hardness, and roasting level, leads to fluctuations in the bulk density and volume of the coffee grounds. Second, long-term operation of the robot causes wear and tear on mechanical parts, and changes in equipment condition (such as wear or accuracy drift in the grinder sprocket, port filter, and tamper) exacerbate the unpredictability of coffee grounds bulking characteristics. Furthermore, environmental factors such as changes in temperature and humidity also affect the physical properties of the coffee grounds, further increasing the complexity of motion control methods.

[0003] Traditional motion control methods for coffee robots are typically based on fixed trajectories and preset parameters, which cannot adapt to changes in the physical state of coffee grounds in real time. For example, fixed tamping force or depth cannot cope with real-time changes in the volume and density of coffee grounds, leading to inconsistent tamping and consequently affecting extraction quality. Inconsistent tamping can result in over- or under-extraction, affecting the flavor and mouthfeel of the coffee. Summary of the Invention

[0004] The purpose of this invention is to address the aforementioned shortcomings by proposing a robot motion control method and system.

[0005] The present invention adopts the following technical solution: A robot motion control method, the method comprising the following steps: During the process of tightening the port filter into the brewing head, the force sensor on the end effector of the robotic arm obtains the reaction force information between the port filter and the brewing head, and at the same time obtains the angle information of the robotic arm during the tightening of the port filter; Based on the reaction force and angle information, assess the actual state of the interface between the port filter and the brewing head, and adjust the locking mechanical parameters and screw-in motion parameters. During the tightening process, if the force sensor reports abnormal resistance, it triggers the robotic arm to adjust its posture and attempt to tighten again.

[0006] This technical solution enables the perception and evaluation of real-time mechanical and kinematic information during the screwing-in process of the port filter, thereby dynamically adjusting the locking and screwing-in parameters and promptly adjusting the posture when encountering abnormal resistance. This effectively addresses uncertainties such as changes in coffee powder condition and wear of mechanical parts, improving the accuracy and stability of screwing in.

[0007] Furthermore, when the force sensor reports abnormal resistance, the steps to trigger the robotic arm's attitude adjustment and re-attempt the screwing-in motion include: Simultaneously acquire the non-screwed axial force signal fed back by the force sensor and the sound signal acquired by the acoustic sensor; Spectral analysis and abnormal noise pattern recognition are performed on the sound signal to obtain the analysis results and extract the stuck acoustic features; When the force sensor detects that the force value of the non-screwed axial force signal reaches the warning threshold, the analysis results are queried. When the acoustic sensor detects a stuck noise corresponding to the stuck acoustic characteristics, the stuck event is confirmed by combining the analysis results. The jam release vector is calculated based on the jamming direction fed back by the force sensor and the abnormal noise direction determined by the acoustic sensor. Based on the jam release vector, send a robotic arm posture adjustment command; Perform the adjustment action; Try the inward rotation motion again.

[0008] Furthermore, when the force sensor detects that the force value of a non-screwed axial force signal reaches the warning threshold, the steps for querying and analyzing the results include: Continuously monitor the non-intrusive axial force signal fed back by the force sensor; Based on the non-intrusive axial force signal, calculate the mean and standard deviation of the non-intrusive axial force signal within a preset time period; The warning threshold is dynamically adjusted based on the mean and standard deviation. When the force value of the non-rotating axial force signal exceeds the dynamically adjusted warning threshold, query the analysis results.

[0009] Furthermore, based on the direction of the jamming as fed back by the force sensor and the direction of the abnormal noise as determined by the acoustic sensor, the steps for calculating the jamming release vector include: The direction of the jamming is fed back by the force sensor and the direction of the abnormal noise is determined by the acoustic sensor; When the angle between the direction of jamming and the direction of abnormal noise exceeds a preset threshold, analyze the current rotation angle and the rate of change of torsional torque of the robotic arm. Based on the current screw-in angle of the robotic arm and the rate of change of torsional torque, determine the weights of the jamming direction fed back by the force sensor and the abnormal noise direction judged by the acoustic sensor. Based on the weights, the jamming direction and the abnormal noise direction are fused together to calculate the jamming release vector.

[0010] Furthermore, when the force value of the non-rotating axial force signal exceeds the dynamically adjusted warning threshold, the steps for querying the analysis results include: Continuously monitor vibration sensor data at each joint of the robotic arm; Based on vibration sensor data, identify specific vibration patterns associated with wear or loosening of robotic arm components; Extract time-domain and frequency-domain features from sound signals; Based on the time domain and frequency domain characteristics, it is matched with a preset library of abnormal noise patterns of the robotic arm itself; The duration, gradient, and peak characteristics of the non-intrusive axial force signal from the force sensor were analyzed. Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, a mechanical characteristic benchmark is compared with that generated based on the mechanical characteristics of preset typical jamming events and external instantaneous interference events. When the force value of the non-rotational axial force signal exceeds the dynamically adjusted warning threshold, the specific vibration mode identified, the abnormal mode matched by the acoustic sensor, and the duration, gradient and peak characteristics of the non-rotational axial force signal are comprehensively judged. When a specific vibration pattern is identified, and the acoustic sensor matches the abnormal pattern, and the non-rotating axial force signal has a long duration and a gentle change gradient, it is determined that the robot arm itself is abnormal. When the duration of the non-rotating axial force signal is short and the change gradient is drastic, and the sound signal is a momentary impact sound but without a specific vibration mode, it is judged as an external momentary interference. When the non-rotating axial force signal exhibits the mechanical characteristics of a preset typical jamming event, and the sound signal exhibits the acoustic characteristics of jamming, and there is no evidence of abnormality in the robotic arm's own components or external instantaneous interference, the jamming event is confirmed. Query analysis results.

[0011] Furthermore, the step of comparing the duration, gradient, and peak characteristics of the non-entrapment axial force signal with a mechanical characteristic benchmark generated based on the mechanical characteristics of preset typical jamming events and external instantaneous disturbance events includes: Continuously monitor the non-screwed axial force signal and record its duration, gradient, and peak characteristics. Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, combined with the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity, the mechanical characteristic benchmarks of typical jamming events and external instantaneous interference events under the current interface state are dynamically generated. The duration, gradient, and peak characteristics of the non-rotating axial force signal are compared with mechanical characteristic benchmarks. Based on the comparison results, determine the event type of the non-rotating axial force signal.

[0012] Furthermore, the steps for dynamically generating mechanical characteristic benchmarks for typical jamming events and external transient disturbance events under the current interface state include: Continuously monitor the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity; When the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity change, the change magnitude is compared with the preset change threshold. When the change exceeds the change threshold, the interface state update mechanism is triggered. The duration, gradient, and peak characteristics of the non-rotating axial force signal within the preset time window are analyzed. Combined with the cleanliness of the current brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity, the mechanical characteristic benchmark is calibrated in real time. When the cleanliness of the brewing head interface decreases, the number of times the brewing head interface has been used increases, or the ambient temperature and humidity fluctuate, adjust the mechanical response range of typical jamming events in the mechanical characteristic reference. Tighten the mechanical response range of external transient disturbance events in the mechanical characteristic reference.

[0013] Furthermore, when the cleanliness of the brewing head interface decreases, the number of times the brewing head interface has been used increases, or the ambient temperature and humidity fluctuate, the steps for adjusting the mechanical response range of typical jamming events in the mechanical characteristic reference include: The interface area of ​​the brewing head is divided into multiple sub-areas; For each sub-region, analyze the historical mechanical response data of jamming events in the sub-region, and combine it with the cleanliness of the current brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity to calculate the adjustment coefficient of the mechanical response range of the sub-region. Based on the mechanical response range adjustment coefficient, the upper and lower limits of the mechanical response range of the jamming event in each sub-region are adjusted, and the adjusted mechanical response range is updated to the mechanical characteristic benchmark.

[0014] Furthermore, the steps for tightening the mechanical response range of external transient disturbance events in the mechanical characteristic reference include: Continuously monitor the duration, gradient, and peak characteristics of the non-rotating axial force signal; Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, the type of external transient disturbance event can be identified. When the type of external transient disturbance event is a minor collision, the mechanical response range of the external transient disturbance event in the mechanical characteristic reference is tightened according to the collision intensity of the non-screwed axial force signal and the duration of the collision event. When the type of external transient interference event is a sudden change in environmental noise, the mechanical response range of the external transient interference event in the mechanical characteristic reference is tightened according to the noise spectrum characteristics and duration of the noise characteristics of the sound signal. When the type of external transient disturbance event is an occasional micro-motion inside the equipment, the mechanical response range of the external transient disturbance event in the mechanical characteristic benchmark is tightened according to the vibration mode and vibration duration collected by the vibration sensor.

[0015] This application also discloses a robot motion control system applied to a robot motion control method, the system comprising: The data acquisition module acquires the reaction force information between the port filter and the brewing head through the force sensor on the end effector of the robotic arm during the process of tightening the port filter into the brewing head, and at the same time acquires the angle information of the robotic arm during the process of tightening the port filter. The processing module evaluates the actual state of the interface between the port filter and the brewing head based on the reaction force information and angle information, and adjusts the locking mechanical parameters and screw-in motion parameters. During the tightening process, when the force sensor reports abnormal resistance, the trigger module triggers the robotic arm to adjust its posture and attempt to tighten again.

[0016] Through modular design, the system can efficiently acquire, process, and trigger actions, providing a reliable platform for the implementation of robot motion control methods.

[0017] This application significantly improves the extraction quality of coffee, overcomes the limitations of traditional fixed tamping force or depth methods that are difficult to adapt to real-time changes, and achieves a smarter and more stable coffee making process.

[0018] To further understand the features and technical content of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are for reference and illustration only and are not intended to limit the present invention. Attached Figure Description

[0019] Figure 1 This is a flowchart of a robot motion control method according to the present invention; Figure 2 This is a schematic diagram of the structure of a robot motion control system according to the present invention. Detailed Implementation

[0020] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can understand the advantages and effects of the present invention from the content disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the spirit of the present invention. Furthermore, the accompanying drawings of the present invention are for simple illustrative purposes only and are not depictions of actual dimensions; this is stated in advance. The following embodiments will further describe the relevant technical content of the present invention in detail, but the disclosed content is not intended to limit the scope of protection of the present invention.

[0021] This embodiment provides a robot motion control method and system, combined with Figure 1 and Figure 2 As shown.

[0022] refer to Figure 1 A robot motion control method, the method comprising the following steps: During the process of tightening the port filter into the brewing head, the force sensor on the end effector of the robotic arm obtains the reaction force information between the port filter and the brewing head, and at the same time obtains the angle information of the robotic arm during the tightening of the port filter; Based on the reaction force and angle information, assess the actual state of the interface between the port filter and the brewing head, and adjust the locking mechanical parameters and screw-in motion parameters. During the tightening process, if the force sensor reports abnormal resistance, it triggers the robotic arm to adjust its posture and attempt to tighten again.

[0023] The "port filter" mentioned in this application typically refers to the component in a coffee machine used to hold coffee grounds, which is connected to the "brewing head" by a screw-on mechanism to form a sealed extraction chamber. The "robotic arm end effector" is the foremost component of the robotic arm, responsible for directly interacting with the port filter and performing operations such as gripping and screwing. A "force sensor" is a device capable of measuring mechanical quantities (such as pressure, tension, torque, etc.), used here to sense the interaction force between the port filter and the brewing head in real time. "Reaction force information" refers to the interaction force data generated between the port filter and the brewing head during the screwing process; this data reflects the resistance, friction, etc., during the screwing process. "Angle information" refers to the real-time rotation angle data of the joints or end effector of the robotic arm when performing the screwing action, used to accurately describe the screwing process and posture. "Locking mechanical parameters" include, but are not limited to, screwing torque and axial pressure; these parameters directly affect the tightness of the connection between the port filter and the brewing head. "Screwing motion parameters" refer to the robot arm's speed, acceleration, path planning, etc. during the screwing process. These parameters determine the smoothness and efficiency of the screwing action.

[0024] During the tightening process of the port filter into the brewing head, a force sensor on the end effector of the robotic arm acquires information about the reaction force between the port filter and the brewing head, as well as the angle information of the robotic arm during the tightening process. Specifically, the force sensor can be configured as a triaxial force sensor, capable of simultaneously measuring force and torque in the X, Y, and Z directions. At the start of the tightening action, the force sensor continuously collects mechanical data of the contact surface between the port filter and the brewing head. For example, when the port filter begins to make threaded contact with the brewing head, the force sensor detects changes in axial pressure and torsional torque. Simultaneously, the encoder or angle sensor of the robotic arm records the rotation angle of each joint in real time, thereby calculating the attitude and tightening angle of the end effector. This information can be transmitted to the robot controller via wired or wireless means for further processing.

[0025] Based on the reaction force and angle information, the actual condition of the interface between the port filter and the brewing head is assessed, and the locking mechanical parameters and screwing motion parameters are adjusted. For example, a mechanical model can be preset, which describes the relationship between the reaction force and angle when the port filter is screwed into the brewing head under ideal conditions. By comparing the real-time acquired reaction force and angle information with this model, it is possible to assess whether there are any abnormalities in the interface, such as thread misalignment, foreign object jamming, or uneven coffee powder accumulation. If the assessment results show that the interface condition deviates from the ideal situation, such as an abnormal increase in reaction force within a certain angle range, the locking mechanical parameters can be adjusted accordingly, such as reducing the upper limit of the tightening torque to avoid damage, or adjusting the screwing motion parameters, such as slowing down the screwing speed or changing the screwing path, to adapt to the actual condition of the interface.

[0026] During the tightening process, if the force sensor detects abnormal resistance, it triggers a robotic arm attitude adjustment and attempts to tighten the thread again. Specifically, an abnormal resistance threshold can be set. When the reaction force detected by the force sensor (such as non-tightening axial force or torsional torque) exceeds this threshold, the system determines that abnormal resistance exists. At this point, the robot controller immediately stops the current tightening action and triggers the robotic arm attitude adjustment program. Attitude adjustment may include small axial or radial displacements, or small angular oscillations, designed to resolve any potential jamming or realign the threads. After adjustment, the robotic arm will attempt to tighten the thread again with the modified motion parameters.

[0027] This application further proposes a procedure for triggering the robotic arm's attitude adjustment and re-attempting the rotation action when the force sensor reports abnormal resistance, including: Simultaneously acquire the non-screwed axial force signal fed back by the force sensor and the sound signal acquired by the acoustic sensor; Spectral analysis and abnormal noise pattern recognition are performed on the sound signal to obtain the analysis results and extract the stuck acoustic features; When the force sensor detects that the force value of the non-screwed axial force signal reaches the warning threshold, the analysis results are queried. When the acoustic sensor detects a stuck noise corresponding to the stuck acoustic characteristics, the stuck event is confirmed by combining the analysis results. The jam release vector is calculated based on the jamming direction fed back by the force sensor and the abnormal noise direction determined by the acoustic sensor. Based on the jam release vector, send a robotic arm posture adjustment command; Perform the adjustment action; Try the inward rotation motion again.

[0028] Specifically, during the tightening process of the port filter, a force sensor on the end effector of the robotic arm not only acquires information on the reaction force in the screwing direction but also simultaneously collects non-screwing axial force signals. Non-screwing axial force signals refer to force components that are not parallel to the main axis of the screwing action; abnormal changes in these components often indicate potential jamming or interference. Simultaneously, acoustic sensors collect sound signals in real time during the screwing process. These acoustic sensors can be positioned near the end effector of the robotic arm or in the brewing head area to capture sounds related to the screwing operation.

[0029] Among these, spectral analysis of sound signals refers to converting the acquired sound signal into frequency domain information using methods such as Fourier transform to reveal the intensity distribution of different frequency components in the sound signal. Abnormal sound pattern recognition refers to using machine learning algorithms or a pre-set acoustic feature library to perform pattern matching on the spectral analysis results and identify abnormal sound patterns that differ from normal incoming sounds. Sticking acoustic features specifically refer to sound patterns with specific frequency and time characteristics, such as friction sounds, impact sounds, or scraping sounds, generated when there is sticking between the port filter and the brewing head.

[0030] In practical applications, when the force sensor detects a non-rotating axial force signal whose value reaches the warning threshold, it indicates the presence of potential abnormal resistance. At this point, the system will query the analysis results obtained from previous spectral analysis and abnormal noise pattern recognition of the sound signal to obtain more comprehensive abnormal information.

[0031] Furthermore, when the acoustic sensor detects a jamming noise corresponding to a preset jamming acoustic characteristic, the system can more reliably confirm the occurrence of the jamming event by combining the analysis results of the non-screwing axial force signal fed back by the force sensor.

[0032] Therefore, to effectively release the jam, a jam release vector needs to be calculated. This vector is calculated based on the jam direction fed back by the force sensor and the abnormal noise direction determined by the acoustic sensor. The force sensor can provide information on the direction of the abnormal force, such as the direction in which the non-rotating axial force increases. The acoustic sensor can determine the specific direction of the abnormal noise by analyzing its propagation path or locating the sound source.

[0033] Based on the calculated jam release vector, the system generates and sends a robotic arm attitude adjustment command. This command instructs the robotic arm to perform minor, targeted attitude adjustments, such as slight translation or rotation along the direction of the jam release vector, to eliminate the jam. After performing the adjustment, the robotic arm will attempt the screwing action again to verify that the jam has been successfully released and to continue the tightening process.

[0034] In some preferred embodiments, it is assumed that a robotic arm is screwing a port filter into the brewing head. During the screwing process, a force sensor suddenly detects a significant increase in the non-screwing axial force signal in a certain direction, exceeding a preset warning threshold. Simultaneously, the acoustic sensor, after spectral analysis, identifies a high-frequency scraping sound, a sound pattern that highly matches a preset jamming acoustic characteristic. The system, after comprehensive judgment, confirms that a jamming event has occurred. Furthermore, the non-screwing axial force signal from the force sensor indicates that the jamming occurs at the right edge of the port filter, and the acoustic sensor, through sound source localization technology, determines that the abnormal noise also mainly originates from the right side. Based on this information, the system calculates a slightly upward-leftward translational release vector. Subsequently, the robotic arm receives a command and makes a slight attitude adjustment along this vector direction, for example, translating upward-leftward by 0.5 mm and fine-tuning the angle. After adjustment, the robotic arm attempts the screwing action again; this time, the non-screwing axial force signal returns to normal, the scraping sound disappears, and the port filter is successfully screwed into the brewing head, completing the tightening operation.

[0035] This application further proposes that when the force sensor detects a non-rotating axial force signal whose force value reaches a warning threshold, the steps for querying and analyzing the results include: Continuously monitor the non-intrusive axial force signal fed back by the force sensor; Based on the non-intrusive axial force signal, calculate the mean and standard deviation of the non-intrusive axial force signal within a preset time period; The warning threshold is dynamically adjusted based on the mean and standard deviation. When the force value of the non-rotating axial force signal exceeds the dynamically adjusted warning threshold, query the analysis results.

[0036] Specifically, continuous monitoring of the non-rotational axial force signal from the force sensor aims to acquire real-time data on changes in axial force during the rotation process, providing a continuous data stream for subsequent analysis. The non-rotational axial force signal refers to the force component that is not parallel to the direction of rotation; an abnormal increase in this component typically indicates jamming or obstruction. Further, based on the monitored non-rotational axial force signal, the mean and standard deviation of the signal are calculated over a preset time period. The mean reflects the average axial force level under current operating conditions, while the standard deviation characterizes the degree of fluctuation in the force signal. These two statistics quantify the mechanical characteristics of the current rotation process. Based on this, the warning threshold is dynamically adjusted according to the calculated mean and standard deviation. This means the warning threshold is no longer fixed but can adaptively adjust according to the current mechanical environment. For example, when the mean and standard deviation indicate a large range of normal mechanical fluctuations under the current operating environment, the warning threshold can be widened accordingly; conversely, when the fluctuation range is small, the warning threshold can be tightened to improve detection sensitivity. Finally, when the force value of the non-rotating axial force signal exceeds the dynamically adjusted warning threshold, the system will trigger the query and analysis of the results to further confirm whether there is a jamming event.

[0037] Specifically, continuous monitoring ensures real-time perception of changes in force signals. By calculating the mean and standard deviation, the system can establish a "baseline" and "normal fluctuation range" for the non-rotating axial force signal under current operating conditions, thereby gaining a more accurate understanding of the current mechanical state. Because the warning threshold can be dynamically adjusted based on these real-time statistical data, it can adapt to changes in the condition of the port filter and brewing head interface, such as wear, differences in cleanliness, or fluctuations in ambient temperature, thus avoiding false alarms or missed alarms caused by fixed thresholds.

[0038] In some preferred embodiments, it is assumed that a force sensor continuously monitors the non-screwing axial force signal during the robot's tightening operation of the port filter. In the initial tightening stage, the mean and standard deviation of the non-screwing axial force signal may be low because the interface is not yet fully in contact. At this time, the system dynamically sets a relatively low warning threshold based on these statistics to ensure sensitive detection of early slight resistance. As the tightening process progresses, if there is slight friction or unevenness between the port filter and the brewing head interface, the non-screwing axial force signal may fluctuate slightly. The system updates the mean and standard deviation in real time and adjusts the warning threshold accordingly. For example, if the interface is slightly worn due to long-term use, causing a slight increase in the axial force during normal tightening and a wider fluctuation range, the dynamically adjusted warning threshold will also be raised accordingly, thereby avoiding misjudging normal wear-induced mechanical changes as jamming events. Conversely, if the interface is very clean and tightly fitted, the force signal fluctuation is small, and the warning threshold will be tightened to more sensitively capture any abnormal force value changes. When the force value of the non-rotating axial force signal suddenly exceeds this dynamically adjusted threshold by a large margin, the system will immediately query and analyze the results to confirm whether jamming has occurred. This dynamic adjustment mechanism enables the system to distinguish between normal operational fluctuations and genuine abnormal resistance, thereby improving the accuracy of jamming detection and the system's intelligence level.

[0039] This application further proposes a step for calculating the jam release vector based on the jamming direction fed back by the force sensor and the abnormal noise direction determined by the acoustic sensor, including: The direction of the jamming is fed back by the force sensor and the direction of the abnormal noise is determined by the acoustic sensor; When the angle between the direction of jamming and the direction of abnormal noise exceeds a preset threshold, analyze the current rotation angle and the rate of change of torsional torque of the robotic arm. Based on the current screw-in angle of the robotic arm and the rate of change of torsional torque, determine the weights of the jamming direction fed back by the force sensor and the abnormal noise direction judged by the acoustic sensor. Based on the weights, the jamming direction and the abnormal noise direction are fused together to calculate the jamming release vector.

[0040] Specifically, comparing the direction of jamming reported by the force sensor and the direction of abnormal noise determined by the acoustic sensor aims to preliminarily assess the consistency of the jamming source or direction perceived by the two different types of sensors. Force sensors typically directly sense the contact force between the robotic arm's end effector and the port filter and brewing head, thereby inferring the physical direction of the jamming. Acoustic sensors, on the other hand, analyze the spectrum and abnormal noise patterns of the sound signal to identify the specific acoustic characteristics produced when jamming occurs, and based on this, determine the direction of the abnormal noise source.

[0041] Specifically, when the angle between the direction of jamming and the direction of abnormal noise exceeds a preset threshold, it indicates a significant difference or uncertainty in the information fed back by the two sensors. In this case, to more accurately calculate the jamming release vector, further analysis of the robot arm's current screw-in angle and the rate of change of torsional torque is needed. The robot arm's current screw-in angle reflects its specific posture and position during the screw-in process, while the rate of change of torsional torque indicates the dynamic characteristics of resistance changes during screw-in. These parameters are crucial for understanding the underlying causes of jamming and selecting an appropriate release strategy.

[0042] In practical applications, the weights of the force sensor's feedback on the direction of jamming and the acoustic sensor's judgment of abnormal noise can be determined based on the robotic arm's current screw-in angle and the rate of change of torsional torque. For example, when the robotic arm is at a specific screw-in angle and the rate of change of torsional torque shows a certain pattern, it may mean that the force sensor's data is more reliable, or the acoustic sensor's data is more indicative. By dynamically adjusting the weights, the information from both sensors can be utilized more flexibly.

[0043] Therefore, based on the weights, the directions of the jamming and abnormal noise are fused to calculate the jamming release vector. The fusion process can employ weighted averaging, fuzzy logic, or other multi-sensor data fusion algorithms. In this way, the feedback from both sensors can be comprehensively considered, combined with the robotic arm's own motion state, to generate a more accurate and robust jamming release vector, guiding the robotic arm to adjust its posture.

[0044] When there is a significant difference in the direction of the jamming reported by two sensors, analyzing the real-time kinematic and dynamic parameters of the robotic arm can provide a deeper understanding of the nature and cause of the jamming. For example, if the torsional torque of the robotic arm changes drastically at a specific angle, it may indicate that the jamming is due to geometric interference rather than simple friction. In this case, the direction of the force sensor may be more valuable. This approach avoids misjudgments caused by single or incomplete information, making the calculation of the jamming release vector more intelligent and adaptive. This weighted fusion mechanism ensures that the robotic arm receives more accurate attitude adjustment commands in complex or uncertain jamming scenarios, thereby improving the success rate and efficiency of jamming release.

[0045] As a specific implementation, suppose that during the process of screwing the port filter into the brewing head, the force sensor reports the jamming direction along the positive X-axis, while the acoustic sensor detects the abnormal noise direction along the positive Y-axis, with a 90-degree angle between them, which obviously exceeds the preset 15-degree threshold. At this point, the system immediately analyzes the current screwing angle of the robotic arm and the rate of change of the torsional torque. For example, if the current screwing angle of the robotic arm is 30 degrees, and the rate of change of the torsional torque shows a continuous and gradually increasing resistance in the X-axis direction, while the resistance change in the Y-axis direction is not significant, this may indicate that the force sensor feedback in the X-axis direction is more reliable. Based on this analysis, the system can assign a higher weight (e.g., 0.7) to the jamming direction reported by the force sensor and a lower weight (e.g., 0.3) to the abnormal noise direction detected by the acoustic sensor. Subsequently, the two directions are fused using a weighted average to calculate the final jamming release vector. For example, the jamming release vector might be calculated as a direction along the positive X-axis deviating slightly from the positive Y-axis, thereby guiding the robotic arm to make precise attitude adjustments to more effectively release the jam. This dynamic adjustment and fusion mechanism enables the robotic arm to flexibly weigh the reliability of different sensor information according to the actual situation, thereby achieving more intelligent and precise jamming removal.

[0046] This application further proposes that when the force value of the non-rotating axial force signal exceeds the dynamically adjusted warning threshold, the steps for querying the analysis results include: Continuously monitor vibration sensor data at each joint of the robotic arm; Based on vibration sensor data, identify specific vibration patterns associated with wear or loosening of robotic arm components; Extract time-domain and frequency-domain features from sound signals; Based on the time domain and frequency domain characteristics, it is matched with a preset library of abnormal noise patterns of the robotic arm itself; The duration, gradient, and peak characteristics of the non-intrusive axial force signal from the force sensor were analyzed. Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, a mechanical characteristic benchmark is compared with that generated based on the mechanical characteristics of preset typical jamming events and external instantaneous interference events. When the force value of the non-rotational axial force signal exceeds the dynamically adjusted warning threshold, the specific vibration mode identified, the abnormal mode matched by the acoustic sensor, and the duration, gradient and peak characteristics of the non-rotational axial force signal are comprehensively judged. When a specific vibration pattern is identified, and the acoustic sensor matches the abnormal pattern, and the non-rotating axial force signal has a long duration and a gentle change gradient, it is determined that the robot arm itself is abnormal. When the duration of the non-rotating axial force signal is short and the change gradient is drastic, and the sound signal is a momentary impact sound but without a specific vibration mode, it is judged as an external momentary interference. When the non-rotating axial force signal exhibits the mechanical characteristics of a preset typical jamming event, and the sound signal exhibits the acoustic characteristics of jamming, and there is no evidence of abnormality in the robotic arm's own components or external instantaneous interference, the jamming event is confirmed. Query analysis results.

[0047] Specifically, continuous monitoring of vibration sensor data at each joint of the robotic arm refers to the real-time acquisition of vibration signals by installing vibration sensors at various key joints of the robotic arm. These vibration signals can reflect the operating status of internal components of the robotic arm, such as bearing wear and gear loosening. Based on the vibration sensor data, specific vibration modes related to wear or loosening of robotic arm components can be identified. For example, through methods such as spectrum analysis and time-domain statistical feature analysis, vibration characteristics corresponding to known fault modes can be extracted and compared with a preset fault mode library.

[0048] The extraction of time-domain and frequency-domain features from sound signals involves processing the sound signals collected by acoustic sensors, such as performing Fourier transforms to obtain frequency-domain information, or calculating time-domain features such as root mean square values ​​and peak values. Based on these time-domain and frequency-domain features, the signals can be matched with a pre-defined library of abnormal noise patterns of the robotic arm. This library contains typical sound features under different conditions, such as normal operation of the robotic arm, component wear, and loosening, thereby helping to determine the source of abnormal sounds.

[0049] In practical applications, the analysis of the duration, gradient, and peak characteristics of force sensors in non-entrapment axial force signals aims to deeply explore the dynamic changes in force signals. Duration reflects the duration of the abnormal force, gradient reveals the rate of force increase or decrease, and peak characteristics indicate the maximum intensity of the abnormal force. These characteristics are used to compare with mechanical characteristic benchmarks generated based on preset typical jamming events and external transient disturbance events to distinguish between abnormal events of different natures. The mechanical characteristic benchmarks are pre-established and include the typical performance of various known abnormal events (such as typical jamming, minor collisions, environmental noise, etc.) on force signals.

[0050] When the force value of the non-entering axial force signal exceeds the dynamically adjusted warning threshold, the system will make a comprehensive judgment. This judgment process combines the specific vibration mode identified by the vibration sensor, the abnormal mode matched by the acoustic sensor, and the duration, gradient, and peak characteristics of the non-entering axial force signal. Through multi-source information fusion, the nature of the abnormal event can be diagnosed more accurately.

[0051] Specifically, when a specific vibration pattern is identified and the acoustic sensor matches the abnormal pattern, and the non-rotating axial force signal has a long duration and a gradual change gradient, this usually indicates that there is a persistent, slowly developing abnormality inside the robotic arm, such as increased friction due to component wear or loosening, and is therefore judged as an abnormality of the robotic arm's own components.

[0052] When the duration of the non-rotating axial force signal is short and the change gradient is drastic, and the sound signal is a momentary impact sound without a specific vibration mode, this is consistent with the characteristics of external momentary interference. For example, when a robotic arm accidentally and slightly collides with an external object during its movement, it is therefore judged as external momentary interference.

[0053] When the non-rotating axial force signal exhibits the mechanical characteristics of a typical jamming event, and the sound signal exhibits the acoustic characteristics of jamming, without evidence of abnormality in the robotic arm's own components or external transient interference, the occurrence of a jamming event can be confirmed. This means that the possibility of robotic arm malfunction and external transient interference has been ruled out, thus attributing the anomaly to actual jamming between the port filter and the brewing head.

[0054] This application's solution addresses the potential for misjudgment based solely on non-rotational axial force signal thresholds by introducing multimodal sensor data (vibration sensor data, acoustic sensor data) and performing more refined feature analysis on force sensor signals. Specifically, when the force value of the non-rotational axial force signal exceeds a dynamically adjusted warning threshold, the system no longer simply queries the analysis results. Instead, it first continuously monitors the vibration sensor data of each joint of the robotic arm to identify the presence of specific vibration patterns related to wear or loosening of robotic arm components, thereby determining whether the anomaly originates from the robotic arm itself. Simultaneously, time-domain and frequency-domain features of the sound signal are extracted and matched against a pre-defined library of abnormal noise patterns inherent to the robotic arm, further assisting in determining the source of the abnormal sound. Furthermore, the duration, gradient, and peak characteristics of the non-rotational axial force signal are analyzed in depth and compared with a mechanical feature benchmark generated based on the mechanical characteristics of pre-defined typical jamming events and external instantaneous interference events to distinguish different types of mechanical responses. By comprehensively evaluating this multi-source information, the system can effectively distinguish abnormal events into malfunctions of the robotic arm's own components, transient external interference, or genuine jamming events, thereby avoiding misoperations or omissions caused by relying on a single mechanical threshold. For example, when an malfunction of the robotic arm's own components is detected, the system can trigger a maintenance alarm instead of adjusting the posture; when a transient external interference is detected, the event can be ignored or a minor adjustment can be made; only when a jamming event is clearly confirmed will the system perform robotic arm posture adjustments and re-attempt the screwing-in action, ensuring the accuracy and effectiveness of the response.

[0055] Specifically, by introducing vibration sensor data and refined analysis of sound signals, combined with the dynamic characteristics of non-screwing axial force signals, the system can achieve refined classification and diagnosis of abnormal events. This allows the system to accurately distinguish between mechanical responses caused by abnormalities in the robotic arm's own components, external transient interference, and actual jamming between the port filter and the brewing head, thus avoiding misjudgments that may occur in traditional methods. For example, relying solely on force threshold judgment, slight wear on robotic arm components or transient external collisions may trigger unnecessary attitude adjustments, leading to reduced efficiency. The solution in this application, through multimodal information fusion and intelligent judgment, ensures that attitude adjustments are only performed when a jamming event actually occurs, thereby reducing invalid operations, improving the efficiency and reliability of the robot's tightening operation, and extending the service life of the robotic arm.

[0056] In some preferred embodiments, it is assumed that during the tightening of the port filter, the force sensor detects that the force value of the non-screwed axial force signal exceeds the dynamically adjusted warning threshold. At this time, the system will initiate a multimodal diagnostic process.

[0057] Specifically, if the vibration sensor data displays a specific high-frequency vibration pattern consistent with wear on the robotic arm's joint bearings, and the sound signal collected by the acoustic sensor, after spectral analysis, matches an abnormal noise pattern generated by friction within the robotic arm's internal components, and the non-rotational axial force signal has a long duration and a gradual change gradient (e.g., the force value slowly rises and remains at a high level), then the system will comprehensively determine that this abnormality is due to a malfunction in the robotic arm's own components. In this case, the system will not trigger rotational attitude adjustment, but may instead issue a maintenance alarm to the operator, recommending a check of the robotic arm's health condition.

[0058] For example, if the force sensor detects a non-screwing axial force signal that is a transient impact with a very short duration, a drastic gradient, and a high peak value, while the acoustic sensor picks up a crisp "bang," but the vibration sensor data does not show any specific pattern related to the abnormality of the robotic arm components, the system will determine that this abnormality is an external transient disturbance, such as the robotic arm accidentally and slightly bumping into the casing of a nearby coffee machine during movement. In this case, the system may briefly pause the screwing action and then immediately resume it without performing complex attitude adjustments.

[0059] This application further proposes the following steps for comparing the above-mentioned mechanical feature benchmarks generated based on the duration, gradient, and peak characteristics of the non-rotating axial force signal with the mechanical features of preset typical jamming events and external instantaneous disturbance events: Continuously monitor the non-screwed axial force signal and record its duration, gradient, and peak characteristics. Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, combined with the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity, the mechanical characteristic benchmarks of typical jamming events and external instantaneous interference events under the current interface state are dynamically generated. The duration, gradient, and peak characteristics of the non-rotating axial force signal are compared with mechanical characteristic benchmarks. Based on the comparison results, determine the event type of the non-rotating axial force signal.

[0060] Specifically, the duration of the non-entrapped axial force signal refers to the time from when the force sensor detects abnormal resistance until the resistance returns to normal or reaches a steady state, reflecting the instantaneous or persistent nature of the event. The gradient of change refers to the rate of change of the non-entrapped axial force signal over a short period, characterizing the suddenness or gradualness of the event. The peak characteristic refers to the maximum force value reached by the non-entrapped axial force signal during the event, indicating the intensity of the event. These characteristics collectively constitute a quantitative description of the abnormal mechanical event.

[0061] The cleanliness of the brew head's interface refers to the absence of coffee residue, scale, or other foreign matter on its surface, directly affecting the coefficient of friction and screw-in resistance. The number of times the brew head's interface has been used refers to the total number of screw-in and screw-out operations it has undergone, reflecting the degree of wear. Ambient temperature and humidity refer to the temperature and humidity of the operating environment, which can affect the physical properties and frictional performance of the materials. These parameters are considered key factors influencing the actual condition of the interface.

[0062] In practical applications, dynamically generating mechanical characteristic benchmarks for typical jamming events and external transient interference events under the current interface state refers to the system's real-time adjustment and optimization of the expected range or threshold of the mechanical characteristics (such as duration, gradient of change, and peak value) of typical jamming events and external transient interference events based on real-time collected data on the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and ambient temperature and humidity data, combined with historical data and preset models. For example, when the interface cleanliness decreases, the system may appropriately widen the response range for the mechanical characteristics of jamming events, because a slight increase in resistance may be caused by dirt rather than actual jamming. Conversely, when the interface is in optimal condition, the system may tighten the response range to improve the sensitivity of jamming detection.

[0063] Therefore, by comparing the duration, gradient, and peak characteristics of the currently monitored non-rotating axial force signal with dynamically generated mechanical characteristic benchmarks, the type of the current abnormal mechanical event can be determined more accurately. For example, by calculating the similarity or distance between the feature vectors of the current feature and the feature vectors of different event types (such as typical jamming or external transient interference) in the benchmark, the best-matching event type can be obtained.

[0064] Because the mechanical characteristic benchmark can adapt to interface wear, contamination, and environmental changes in real time, the comparison and judgment of the duration, gradient, and peak characteristics of non-screwing axial force signals are more accurate. For example, when the brewing head interface wears slightly due to long-term use, its normal screwing resistance curve may change slightly. If the initial fixed benchmark is still used for judgment, the normal wear effect may be misjudged as a jamming event. By dynamically adjusting the benchmark, the system can identify this normal change caused by wear and exclude it from jamming events, thus avoiding false alarms. At the same time, when the interface cleanliness decreases or the ambient humidity increases, leading to increased friction, the dynamic benchmark can adjust the mechanical response range to "typical jamming" accordingly, ensuring that it can still accurately identify real jamming events under these conditions, rather than attributing them to environmental interference.

[0065] This application further proposes steps for dynamically generating mechanical characteristic benchmarks for typical jamming events and external transient disturbance events under the current interface state, including: Continuously monitor the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity; When the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity change, the change magnitude is compared with the preset change threshold. When the change exceeds the change threshold, the interface state update mechanism is triggered. The duration, gradient, and peak characteristics of the non-rotating axial force signal within the preset time window are analyzed. Combined with the cleanliness of the current brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity, the mechanical characteristic benchmark is calibrated in real time. When the cleanliness of the brewing head interface decreases, the number of times the brewing head interface has been used increases, or the ambient temperature and humidity fluctuate, adjust the mechanical response range of typical jamming events in the mechanical characteristic reference. Tighten the mechanical response range of external transient disturbance events in the mechanical characteristic reference.

[0066] Specifically, continuously monitoring the cleanliness of the brewing head's interface, the number of times the interface has been used, and the ambient temperature and humidity can be understood as acquiring the physical state and environmental parameters of the brewing head's interface area in real time through various sensors integrated into the robot system, such as vision sensors, counters, and temperature and humidity sensors. Cleanliness can be quantitatively assessed using image recognition technology to determine the amount of dirt and residue on the interface surface; the number of uses is obtained by recording the cumulative number of screw-in and screw-out operations; and the ambient temperature and humidity are directly measured by environmental sensors.

[0067] Specifically, when the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity change, the magnitude of the change is compared with a preset change threshold. The purpose is to identify state changes that may have a significant impact on the mechanical properties of the interface. For example, a decrease in cleanliness exceeding a certain percentage, an increase in the number of uses reaching a certain milestone, or temperature and humidity fluctuations exceeding the normal operating range can all be considered significant changes that require subsequent adjustments.

[0068] In practical applications, when the change exceeds a threshold, the interface status update mechanism is triggered. This means the system automatically initiates a process to reassess and adjust parameters related to the current interface status. This mechanism ensures the dynamic adaptability of the mechanical characteristic reference, enabling it to reflect actual wear, contamination, or environmental impacts on the interface.

[0069] Furthermore, by analyzing the duration, gradient, and peak characteristics of the non-rotating axial force signal within a preset time window, and combining this with the cleanliness of the brewing head interface, the number of times the interface has been used, and the ambient temperature and humidity, the mechanical characteristic benchmark is calibrated in real time. This means that after a significant change in the interface state, the system uses recently collected mechanical data, combined with the latest interface state information, to finely adjust the mechanical characteristic benchmark. This ensures the accuracy of the benchmark, enabling it to more accurately reflect the mechanical performance of typical jamming and transient interference under current operating conditions.

[0070] As a preferred implementation, when the cleanliness of the brewing head interface decreases, the number of times the brewing head interface has been used increases, or there are fluctuations in ambient temperature and humidity, the mechanical response range of typical jamming events in the mechanical characteristic benchmark is adjusted. This is to account for the fact that interface wear, contamination, or environmental changes may cause the mechanical performance of jamming events to become more complex or uncertain. For example, unclean or worn interfaces may cause even slight jamming to produce a large mechanical response, or may expand the mechanical characteristic range of jamming. Through adjustment, potential jamming events can be identified more broadly, avoiding missed detections.

[0071] Meanwhile, tightening the mechanical response range for external transient interference events in the mechanical characteristic benchmark aims to reduce the possibility of false alarms for external transient interference events when the interface condition is poor or the environment fluctuates. For example, when the ambient temperature and humidity fluctuate significantly, some mechanical signals similar to transient interference may be generated. By tightening the response range, the accuracy of identifying genuine external transient interference can be improved, avoiding misjudging interface problems or environmental noise as external transient interference.

[0072] Specifically, when the interface status changes due to decreased cleanliness, increased usage, or fluctuations in ambient temperature and humidity, the system triggers an interface status update mechanism and calibrates the mechanical characteristic benchmark by combining real-time non-screw-in axial force signal characteristics. Thus, for typical jamming events, the mechanical response range is appropriately adjusted, for example, expanded, to accommodate changes in mechanical characteristics that may be caused by interface wear or contamination, thereby improving the detection sensitivity of jamming events. Simultaneously, for external transient interference events, the mechanical response range is tightened to reduce the risk of misjudgment in complex or unstable environments, thereby improving the accuracy of interference event identification. This adaptive benchmark adjustment mechanism enables the system to more accurately distinguish different types of abnormal events, effectively solving the problem of misjudgment or missed detection that may occur with traditional fixed benchmarks under varying operating conditions.

[0073] In some preferred embodiments, it is assumed that a coffee robot is performing a port filter tightening operation. Initially, the brew head interface is clean and used infrequently, and the ambient temperature and humidity are stable; at this time, the mechanical characteristic benchmark is set to a relatively strict range. As the robot operates for an extended period, the cleanliness of the brew head interface gradually decreases; for example, coffee residue begins to accumulate at the interface threads, and the number of uses also increases significantly. Furthermore, the ambient temperature and humidity may fluctuate due to seasonal changes or air conditioning operation. In this situation, the solution of this application continuously monitors the decrease in the cleanliness of the brew head interface, the increase in the number of uses, and the fluctuations in ambient temperature and humidity. When these changes exceed a preset threshold, the interface status update mechanism is triggered. The system analyzes the duration, gradient, and peak characteristics of recently acquired non-screwed axial force signals and, combined with the current actual state of the interface, performs real-time calibration of the mechanical characteristic benchmark. Specifically, since decreased interface cleanliness and increased usage may cause the mechanical response of jamming events to become more complex or less obvious, the system adjusts the mechanical response range of typical jamming events in the mechanical characteristic benchmark, for example, by appropriately widening its upper and lower limits, to ensure that even jamming caused by minor residue can be effectively identified. Meanwhile, considering that fluctuations in ambient temperature and humidity may introduce additional transient mechanical disturbances, the system will tighten the mechanical response range of external transient disturbance events in the mechanical characteristic benchmark, thereby improving the identification accuracy of true external transient disturbances and avoiding misjudging environmental noise or slight vibrations as external disturbances that require attitude adjustment.

[0074] When the cleanliness of the brewing head interface decreases, the number of times the brewing head interface has been used increases, or the ambient temperature and humidity fluctuate, the steps to adjust the mechanical response range of typical jamming events in the mechanical characteristic reference include: The interface area of ​​the brewing head is divided into multiple sub-areas; For each sub-region, analyze the historical mechanical response data of jamming events in the sub-region, and combine it with the cleanliness of the current brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity to calculate the adjustment coefficient of the mechanical response range of the sub-region. Based on the mechanical response range adjustment coefficient, the upper and lower limits of the mechanical response range of the jamming event in each sub-region are adjusted, and the adjusted mechanical response range is updated to the mechanical characteristic benchmark.

[0075] Specifically, dividing the interface area of ​​the brewing head into multiple sub-regions refers to logically or physically subdividing the entire annular or planar area of ​​the interface into several smaller, independent sub-regions based on the interface's geometry, stress distribution characteristics, or the location of historical jamming events. For example, the interface area can be divided into multiple sector-shaped or annular sub-regions along the circumferential or radial direction, or divided according to specific sections of the thread. The purpose is to achieve refined management and localized perception of the interface status.

[0076] For each sub-region, historical jamming event mechanical response data is analyzed. This data, combined with the cleanliness of the current brewing head interface, the number of times the interface has been used, and ambient temperature and humidity, is used to calculate an adjustment coefficient for the sub-region's mechanical response range. Historical jamming event mechanical response data can be understood as the duration, gradient, and peak characteristics of the non-screwing axial force signal fed back by the force sensor when a jamming event occurs in that specific sub-region during past operations. This data reflects the typical jamming mechanical performance of that sub-region under different conditions. Combining this data with the cleanliness of the current brewing head interface, the number of times the interface has been used, and ambient temperature and humidity allows for a more comprehensive assessment of the current state of that sub-region. For example, if the cleanliness of a sub-region is significantly reduced, and historical data shows that that region is more prone to a specific type of jamming under low cleanliness conditions, an adjustment coefficient reflecting this trend can be calculated. This adjustment coefficient aims to quantify the degree of influence of the current sub-region's state on the mechanical response range of jamming events.

[0077] In practical applications, the upper and lower limits of the mechanical response range for jamming events in each sub-region are adjusted based on the mechanical response range adjustment coefficient, and the adjusted mechanical response range is then updated to the mechanical characteristic benchmark. This means that for each sub-region, the mechanical response range for jamming events is no longer fixed, but dynamically adjusted based on its own historical data and current environmental conditions. For example, if the adjustment coefficient of a certain sub-region indicates an increased risk of jamming, the upper and lower limits of the mechanical response range of that sub-region can be appropriately tightened, enabling it to sensitively identify even small abnormal mechanical signals as jamming events; conversely, if the adjustment coefficient indicates a decreased risk of jamming, the range can be appropriately widened. This measure aims to ensure that the mechanical characteristic benchmark can accurately adapt to the actual changes in the local area of ​​the interface.

[0078] This application's solution divides the interface area of ​​the brewing head into multiple sub-regions and adjusts the mechanical response range of each sub-region independently, thus solving the precision deficiency problem that may exist in traditional overall adjustment schemes. It is precisely because of the in-depth analysis of the historical jamming event mechanical response data of each sub-region, combined with real-time information such as the current cleanliness of the sub-region, the number of uses, and the ambient temperature and humidity, that the mechanical response range adjustment coefficient can more accurately reflect the actual state of the local area. Through this refined adjustment, the mechanical characteristic benchmark can adaptively respond to the differential changes in different parts of the interface, avoiding local area identification errors caused by overall adjustment.

[0079] In some preferred embodiments, it is assumed that the interface area of ​​a brewing head is divided into four equally divided sector-shaped sub-regions: sub-region A, sub-region B, sub-region C, and sub-region D. During a tightening operation, the force sensor detects abnormal resistance at sub-region B. At this time, the system first queries the historical mechanical response data of sub-region B for jamming events. For example, historical data shows that when the interface cleanliness of sub-region B is low, the peak force of a typical jamming event is usually between 50N and 70N. At the same time, the system obtains that the cleanliness of the current sub-region B has decreased below a preset threshold, and the number of times the sub-region has been used has increased significantly. Based on this information, the system calculates the mechanical response range adjustment coefficient of sub-region B, which indicates that the mechanical response range of the jamming event in sub-region B needs to be tightened by 10%. Therefore, the upper and lower limits of the mechanical response range of the jamming event in sub-region B are adjusted to 45N to 63N. When the force sensor detects a non-screwed axial force signal of 55N in sub-region B, since it falls within the adjusted range of 45N to 63N, the system can immediately confirm that the abnormal resistance is a jamming event and trigger the corresponding robotic arm posture adjustment. Meanwhile, even if the cleanliness, number of uses, and ambient temperature and humidity of sub-regions A, C, and D do not change significantly, their mechanical response range remains unchanged, avoiding unnecessary adjustments and thus ensuring the accuracy and adaptability of jamming identification throughout the entire interface area.

[0080] This application further proposes steps for tightening the mechanical response range of external transient disturbance events in the mechanical characteristic reference, including: Continuously monitor the duration, gradient, and peak characteristics of the non-rotating axial force signal; Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, the type of external transient disturbance event can be identified. When the type of external transient disturbance event is a minor collision, the mechanical response range of the external transient disturbance event in the mechanical characteristic reference is tightened according to the collision intensity of the non-screwed axial force signal and the duration of the collision event. When the type of external transient interference event is a sudden change in environmental noise, the mechanical response range of the external transient interference event in the mechanical characteristic reference is tightened according to the noise spectrum characteristics and duration of the noise characteristics of the sound signal. When the type of external transient disturbance event is an occasional micro-motion inside the equipment, the mechanical response range of the external transient disturbance event in the mechanical characteristic benchmark is tightened according to the vibration mode and vibration duration collected by the vibration sensor.

[0081] Specifically, continuously monitoring the duration, gradient, and peak characteristics of the non-entrapped axial force signal refers to the system acquiring and recording the dynamic changes of the non-entrapped axial force signal fed back by the force sensor in real time. The duration refers to the length of time the force signal exceeds a certain reference value; the gradient refers to the rate of change of the force signal per unit time, reflecting the severity of the force change; and the peak characteristic refers to the maximum value reached by the force signal within a specific time period. These characteristics are used to quantify the mechanical performance of external transient disturbance events.

[0082] Furthermore, based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, the type of external transient interference event can be identified. For example, a signal with a short duration, a drastic gradient, and a high peak value may indicate a transient impact or collision; a signal with a longer duration, a gentle gradient, but a lower peak value may be associated with some persistent but low-intensity interference. Through a pre-defined classification model or rule, these mechanical characteristics can be mapped to specific interference types, such as minor collisions, sudden changes in environmental noise, or occasional micro-motions within the equipment.

[0083] When an external transient disturbance event is identified as a minor collision, the system tightens the mechanical response range of the external transient disturbance event in the mechanical characteristic benchmark based on the collision intensity of the non-screwed axial force signal and the duration of the collision event. The collision intensity can be characterized by the peak or average force value of the non-screwed axial force signal, while the duration of the collision event directly reflects the transient nature of the collision. By analyzing these parameters, the benchmark can be adjusted more accurately to avoid misclassifying minor, harmless collisions as jamming events.

[0084] When the type of external transient interference event is a sudden change in environmental noise, the system tightens the mechanical response range of the external transient interference event in the mechanical characteristic benchmark based on the noise spectrum characteristics and duration of the noise characteristic of the sound signal. The noise spectrum characteristics can be obtained through spectral analysis methods such as Fourier transform on the sound signal collected by the acoustic sensors to identify the frequency distribution and energy concentration areas of the noise. The duration of the noise characteristic indicates the duration of the noise event. By combining these acoustic characteristics, it is possible to effectively distinguish between mechanical fluctuations caused by environmental noise and actual jamming phenomena.

[0085] When the type of external transient disturbance event is an occasional micro-motion within the equipment, the system tightens the mechanical response range of the external transient disturbance event in the mechanical characteristic benchmark based on the vibration mode and duration collected by the vibration sensor. The vibration mode refers to the specific vibration frequency, amplitude, and waveform generated by each joint or component of the robotic arm during micro-motion, which can be identified through vibration sensor data. The vibration duration records the duration of the micro-motion event. By analyzing these vibration characteristics, mechanical signal fluctuations caused by occasional micro-motions of the robotic arm's own structure or internal components can be eliminated, thereby improving the accuracy of jamming event detection.

[0086] Specifically, when a minor collision is detected, the system adjusts the threshold based on the intensity and duration of the collision to ensure that only collisions reaching a certain level are considered; when a sudden change in environmental noise is detected, the system uses the spectral characteristics of the sound signal to filter out the mechanical response related to the noise; when an occasional micro-motion inside the device is detected, the system eliminates internal interference based on the vibration mode and duration.

[0087] In some preferred embodiments, this application is implemented as follows: Suppose that during the tightening of the port filter, the non-screwed axial force signal fed back by the force sensor suddenly fluctuates.

[0088] First, the system continuously monitors the duration, gradient, and peak characteristics of the non-rotating axial force signal.

[0089] If the monitoring results show that the duration of the non-intrusive axial force signal is extremely short (e.g., less than 50 milliseconds), the change gradient is drastic, and the peak value is relatively low, while the acoustic sensor does not detect obvious jamming noise, but there is a momentary impact sound in the sound signal, then the system will identify the event as a "minor collision". In this case, the system will tighten the mechanical response range of external instantaneous interference events in the mechanical characteristic benchmark based on the collision intensity (e.g., peak force) and the duration of the collision event. For example, it may appropriately lower the upper limit of the mechanical response of this type of event to avoid misjudging such harmless minor collisions as jamming.

[0090] If the monitoring results show that the duration of the non-rotating axial force signal is short, the gradient of change is moderate, and the peak value is not high, and at the same time, after spectral analysis of the sound signal collected by the acoustic sensor, a sudden increase in energy is found within a specific frequency range, and the duration of this noise feature is short, then the system will identify the event as an "abrupt change in environmental noise." In this case, the system will tighten the mechanical response range of the external instantaneous interference event in the mechanical characteristic benchmark based on the noise spectrum characteristics of the sound signal (e.g., identifying high-frequency noise) and the duration of the noise feature. For example, within the mechanical response range corresponding to the noise frequency, the threshold will be increased to filter out the mechanical fluctuations caused by environmental noise.

[0091] If the monitoring results show that the duration of the non-rotating axial force signal is relatively long (e.g., exceeding 100 milliseconds), the change gradient is gentle, and the peak value is low, and at the same time, the vibration sensor identifies a specific vibration mode related to occasional micro-motions of internal components of the robotic arm (such as joint motors or reducers) in the joint data of each joint, and the duration of the vibration matches the duration of the force signal, then the system will identify the event as "incidental micro-motions inside the equipment." In this case, the system will tighten the mechanical response range of external instantaneous interference events in the mechanical characteristic benchmark based on the vibration mode and duration collected by the vibration sensor. For example, within the mechanical response range related to the vibration mode, the threshold will be tightened to eliminate interference caused by micro-motions within the robotic arm itself.

[0092] refer to Figure 2 This application proposes a robot motion control system for implementing a robot motion control method. The system includes: The data acquisition module acquires the reaction force information between the port filter and the brewing head through the force sensor on the end effector of the robotic arm during the process of tightening the port filter into the brewing head, and at the same time acquires the angle information of the robotic arm during the process of tightening the port filter. The processing module evaluates the actual state of the interface between the port filter and the brewing head based on the reaction force information and angle information, and adjusts the locking mechanical parameters and screw-in motion parameters. During the tightening process, when the force sensor reports abnormal resistance, the trigger module triggers the robotic arm to adjust its posture and attempt to tighten again.

[0093] The data acquisition module is configured to acquire key mechanical and kinematic data in real time during the tightening of the port filter into the brewing head. Specifically, this module uses a force sensor integrated into the robotic arm's end effector to accurately measure the reaction force generated between the port filter and the brewing head. Simultaneously, the module can also acquire angular information of the robotic arm during the port filter tightening process, for example, through the robotic arm's own encoder or an external vision system. This information forms the basis for subsequent evaluation and control.

[0094] Furthermore, the processing module is configured to receive and analyze the reaction force and angle information acquired by the acquisition module. Based on this real-time data, the module evaluates the actual state of the interface between the port filter and the brewing head, such as determining whether there is misalignment, jamming, or thread misalignment. Based on the evaluation results, the processing module can dynamically adjust the locking mechanical parameters, such as screw-in torque and axial force limitation, as well as the screw-in motion parameters, such as screw-in speed and acceleration, to adapt to the actual situation of the interface and ensure a smooth and accurate tightening process.

[0095] Furthermore, the trigger module is configured to continuously monitor the mechanical data fed back by the force sensor during the tightening process. When the force sensor reports abnormal resistance, such as a sudden increase in force or the appearance of a force in an unexpected direction, the trigger module will immediately identify this anomaly. Once abnormal resistance is detected, the trigger module will immediately trigger the robotic arm to adjust its posture and instruct the robotic arm to attempt to tighten again in order to relieve the abnormal resistance and avoid damage to the port filter or brewing head.

[0096] The content disclosed above is only a preferred and feasible embodiment of the present invention, and is not intended to limit the scope of protection of the present invention. Therefore, all equivalent technical changes made based on the content of the present invention specification and drawings are included within the scope of protection of the present invention. Furthermore, the elements therein can be updated as technology develops.

Claims

1. A robot motion control method, characterized in that, The method includes the following steps: During the process of tightening the port filter into the brewing head, the force sensor on the end effector of the robotic arm obtains the reaction force information between the port filter and the brewing head, and at the same time obtains the angle information of the robotic arm during the tightening of the port filter; Based on the reaction force and angle information, assess the actual state of the interface between the port filter and the brewing head, and adjust the locking mechanical parameters and screw-in motion parameters. During the tightening process, if the force sensor reports abnormal resistance, it triggers the robotic arm to adjust its posture and attempt to tighten again.

2. The robot motion control method as described in claim 1, characterized in that, When the force sensor detects abnormal resistance, the steps that trigger the robotic arm to adjust its posture and attempt the rotation again include: Simultaneously acquire the non-screwed axial force signal fed back by the force sensor and the sound signal acquired by the acoustic sensor; Spectral analysis and abnormal noise pattern recognition are performed on the sound signal to obtain the analysis results and extract the stuck acoustic features; When the force sensor detects that the force value of the non-screwed axial force signal reaches the warning threshold, the analysis results are queried. When the acoustic sensor detects a stuck noise corresponding to the stuck acoustic characteristics, the stuck event is confirmed by combining the analysis results. The jam release vector is calculated based on the jamming direction fed back by the force sensor and the abnormal noise direction determined by the acoustic sensor. Based on the jam release vector, send a robotic arm posture adjustment command; Perform the adjustment action; Try the inward rotation motion again.

3. The robot motion control method as described in claim 2, characterized in that, When the force sensor detects that the force value of the non-rotating axial force signal reaches the warning threshold, the steps for querying and analyzing the results include: Continuously monitor the non-intrusive axial force signal fed back by the force sensor; Based on the non-intrusive axial force signal, calculate the mean and standard deviation of the non-intrusive axial force signal within a preset time period; The warning threshold is dynamically adjusted based on the mean and standard deviation. When the force value of the non-rotating axial force signal exceeds the dynamically adjusted warning threshold, query the analysis results.

4. The robot motion control method as described in claim 2, characterized in that, The steps for calculating the jam release vector, based on the jamming direction fed back by the force sensor and the abnormal noise direction determined by the acoustic sensor, include: The direction of the jamming is fed back by the force sensor and the direction of the abnormal noise is determined by the acoustic sensor; When the angle between the direction of jamming and the direction of abnormal noise exceeds a preset threshold, analyze the current rotation angle and the rate of change of torsional torque of the robotic arm. Based on the current screw-in angle of the robotic arm and the rate of change of torsional torque, determine the weights of the jamming direction fed back by the force sensor and the abnormal noise direction judged by the acoustic sensor. Based on the weights, the jamming direction and the abnormal noise direction are fused together to calculate the jamming release vector.

5. The robot motion control method as described in claim 3, characterized in that, When the force value of the non-rotating axial force signal exceeds the dynamically adjusted warning threshold, the steps for querying the analysis results include: Continuously monitor vibration sensor data at each joint of the robotic arm; Based on vibration sensor data, identify specific vibration patterns associated with wear or loosening of robotic arm components; Extract time-domain and frequency-domain features from sound signals; Based on the time domain and frequency domain characteristics, it is matched with a preset library of abnormal noise patterns of the robotic arm itself; The duration, gradient, and peak characteristics of the non-intrusive axial force signal from the force sensor were analyzed. Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, a mechanical characteristic benchmark is compared with that generated based on the mechanical characteristics of preset typical jamming events and external instantaneous interference events. When the force value of the non-rotational axial force signal exceeds the dynamically adjusted warning threshold, the specific vibration mode identified, the abnormal mode matched by the acoustic sensor, and the duration, gradient and peak characteristics of the non-rotational axial force signal are comprehensively judged. When a specific vibration pattern is identified, and the acoustic sensor matches the abnormal pattern, and the non-rotating axial force signal has a long duration and a gentle change gradient, it is determined that the robot arm itself is abnormal. When the duration of the non-rotating axial force signal is short and the change gradient is drastic, and the sound signal is a momentary impact sound but without a specific vibration mode, it is judged as an external momentary interference. When the non-rotating axial force signal exhibits the mechanical characteristics of a preset typical jamming event, and the sound signal exhibits the acoustic characteristics of jamming, and there is no evidence of abnormality in the robotic arm's own components or external instantaneous interference, the jamming event is confirmed. Query analysis results.

6. The robot motion control method as described in claim 5, characterized in that, The steps for comparing the duration, gradient, and peak characteristics of the non-entrapment axial force signal with a mechanical characteristic benchmark generated based on the mechanical characteristics of preset typical jamming events and external instantaneous disturbance events include: Continuously monitor the non-screwed axial force signal and record its duration, gradient, and peak characteristics. Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, combined with the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity, the mechanical characteristic benchmarks of typical jamming events and external instantaneous interference events under the current interface state are dynamically generated. The duration, gradient, and peak characteristics of the non-rotating axial force signal are compared with mechanical characteristic benchmarks. Based on the comparison results, determine the event type of the non-rotating axial force signal.

7. The robot motion control method as described in claim 6, characterized in that, The steps for dynamically generating mechanical characteristic benchmarks for typical jamming events and external transient disturbance events under the current interface state include: Continuously monitor the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity; When the cleanliness of the brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity change, the change magnitude is compared with the preset change threshold. When the change exceeds the change threshold, the interface state update mechanism is triggered. The duration, gradient, and peak characteristics of the non-rotating axial force signal within the preset time window are analyzed. Combined with the cleanliness of the current brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity, the mechanical characteristic benchmark is calibrated in real time. When the cleanliness of the brewing head interface decreases, the number of times the brewing head interface has been used increases, or the ambient temperature and humidity fluctuate, adjust the mechanical response range of typical jamming events in the mechanical characteristic reference. Tighten the mechanical response range of external transient disturbance events in the mechanical characteristic reference.

8. The robot motion control method as described in claim 7, characterized in that, When the cleanliness of the brewing head interface decreases, the number of times the brewing head interface has been used increases, or the ambient temperature and humidity fluctuate, the steps to adjust the mechanical response range of typical jamming events in the mechanical characteristic reference include: The interface area of ​​the brewing head is divided into multiple sub-areas; For each sub-region, analyze the historical mechanical response data of jamming events in the sub-region, and combine it with the cleanliness of the current brewing head interface, the number of times the brewing head interface has been used, and the ambient temperature and humidity to calculate the adjustment coefficient of the mechanical response range of the sub-region. Based on the mechanical response range adjustment coefficient, the upper and lower limits of the mechanical response range of the jamming event in each sub-region are adjusted, and the adjusted mechanical response range is updated to the mechanical characteristic benchmark.

9. A robot motion control method as described in claim 7, characterized in that, The steps for tightening the mechanical response range of external transient disturbance events in the mechanical characteristic reference include: Continuously monitor the duration, gradient, and peak characteristics of the non-rotating axial force signal; Based on the duration, gradient, and peak characteristics of the non-rotating axial force signal, the type of external transient disturbance event can be identified. When the type of external transient disturbance event is a minor collision, the mechanical response range of the external transient disturbance event in the mechanical characteristic reference is tightened according to the collision intensity of the non-screwed axial force signal and the duration of the collision event. When the type of external transient interference event is a sudden change in environmental noise, the mechanical response range of the external transient interference event in the mechanical characteristic reference is tightened according to the noise spectrum characteristics and duration of the noise characteristics of the sound signal. When the type of external transient disturbance event is an occasional micro-motion inside the equipment, the mechanical response range of the external transient disturbance event in the mechanical characteristic benchmark is tightened according to the vibration mode and vibration duration collected by the vibration sensor.

10. A robot motion control system, applied to the robot motion control method as described in claim 1, characterized in that, The system includes: The data acquisition module acquires the reaction force information between the port filter and the brewing head through the force sensor on the end effector of the robotic arm during the process of tightening the port filter into the brewing head, and at the same time acquires the angle information of the robotic arm during the process of tightening the port filter. The processing module evaluates the actual state of the interface between the port filter and the brewing head based on the reaction force information and angle information, and adjusts the locking mechanical parameters and screw-in motion parameters. During the tightening process, when the force sensor reports abnormal resistance, the trigger module triggers the robotic arm to adjust its posture and attempt to tighten again.

Images