Robot linear rack guide collision detection method, device and electronic equipment

Abstract

This invention provides a method, device, and electronic device for collision detection of a robot linear rack and pinion guide. The method achieves theoretical modeling of the linear rack and pinion guide based on motion state switching and online friction force identification. On this basis, a dynamic threshold function is designed, and collision detection of the guide is achieved by combining time-delay compensated commands and torque feedback data. This scheme has the advantages of strong robustness and low false trigger rate, while also taking into account the response speed and communication latency of the control system, thus improving the safety of the human-machine interaction process at a lower cost.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and more specifically, to a method, apparatus, and electronic device for detecting collisions on a robot linear rack and pinion guideway. Background Technology

[0002] The structure of a linear rack and pinion guide can be divided into a linear guide, a rack, a drive system, and a robot base. Its transmission principle can be simply described as follows: the guide drive motor drives the output shaft gear to mesh with the rack and move forward, thereby realizing the translation of the base and the robot.

[0003] The aforementioned external axes of the guide rails are widely used in robot systems. During the guide rail's movement, they are prone to interference with the environment, leading to collisions. Without collision detection and corresponding response strategies, significant safety hazards exist. However, collision detection functionality for linear rack and pinion guide rails has not yet been widely commercialized in the industry, and related technical solutions are scarce. According to publicly available information, only FANUC possesses a similar function, monitoring the disturbance torque range under a specific trajectory and setting a collision detection threshold accordingly. When the disturbance torque during guide rail operation exceeds the threshold, the system alarms and stops movement.

[0004] Based on publicly available information from the survey and common methods used in collision detection for robot bodies, the implementation route for collision detection technology of linear rack and pinion guides can be as follows: 1. FANUC collision detection based on disturbance torque threshold; 2. Referring to the current loop collision detection method commonly used in robot bodies, linear guides can also achieve collision detection based on theoretical dynamic models; 3. External sensor fusion detection, adding physical sensors to key parts of the guide rail, and detecting whether a collision has occurred based on feedback data; 4. Vision / laser-based pre-collision detection, perceiving and avoiding obstacles in advance.

[0005] Existing collision detection technologies for linear rack and pinion guides rely on disturbance torque threshold-based methods. These methods measure the range of disturbance torque under specific operating conditions (e.g., 100% rated speed, no-load reciprocating operation) and then set a fixed threshold for collision detection. However, this strategy lacks robustness. When the robot is in different poses, carrying different end-effector loads, or moving synchronously with the guide rail, the system's inertia and dynamic characteristics change, causing fluctuations in background disturbance torque and easily leading to false triggers. Setting the threshold too high to improve anti-interference capability significantly reduces collision sensitivity, resulting in detection lag or missed detections. Furthermore, its detection accuracy is highly dependent on the accuracy of the preset threshold, exhibiting poor adaptability to changes in operating conditions.

[0006] The detection method based on theoretical dynamics requires the establishment of an accurate dynamic model of the linear rack and pinion guide. However, due to the strong dynamic coupling between the guide and the robot body, the model is highly complex and parameter identification is difficult. At the same time, limited by the response speed and communication delay of the control system, this method is prone to misjudgment due to model prediction errors during dynamic processes such as motion start-stop, reversal, etc.

[0007] The method of detection based on external sensor fusion increases hardware cost and wiring complexity, and the long-term durability and reliability of sensors in harsh industrial environments face severe challenges, which restricts its engineering application.

[0008] Vision / laser-based pre-collision detection achieves non-contact early warning by establishing an environmental perception zone, but it also suffers from problems such as high cost, susceptibility to light / dust interference, and perception blind spots. In addition, the large amount of data processing poses a challenge to the real-time performance of the control system.

[0009] In summary, it is of great significance to develop a low-cost linear rack and pinion guide collision detection method that is robust, takes into account the response speed of the control system and communication delay, and has a low false trigger rate. Summary of the Invention

[0010] The purpose of this invention is to provide a method, device, and electronic device for detecting collisions in a robot linear rack and pinion guide, so as to balance the response speed of the control system with communication latency and improve the safety of the human-computer interaction process.

[0011] In a first aspect, the present invention provides a collision detection method for a robot linear rack guideway, the method comprising: Response delay data and communication delay data are set for the response speed and communication delay of the control system of the robot linear rack and pinion guide respectively; Based on the communication delay data, perform a time-domain alignment operation between the instruction data and the feedback data; Design a motion state machine by combining the response delay data and the time-domain aligned instruction data; The linear rack and pinion guide is controlled to perform reciprocating motion at different speeds, and torque feedback data is collected based on the switching of the running state corresponding to the motion state machine. Based on the torque feedback data, online friction force identification is performed to conduct theoretical modeling and obtain the theoretical torque of the guide rail; During operation, collision detection is performed based on torque feedback data within the sliding window, theoretical torque of the guide rail, and the set dynamic threshold function to obtain the detection results.

[0012] In an optional implementation, the step of designing a motion state machine by combining the response delay data and the time-domain aligned instruction data includes: The linear rack and pinion guide is divided into multiple motion states based on its actual physical characteristics; Multiple counters are initialized, and a motion state machine is set up. Each counter is used to quantize different motion states and their switching states, and the motion state machine is used to maintain the switching between the motion states based on the count values ​​of each counter.

[0013] In an optional implementation, the step of controlling the linear rack and pinion guide to perform reciprocating motion at different speeds and collecting torque feedback data based on the switching of the operating state corresponding to the motion state machine includes: Based on the torque variation law of the linear rack and pinion guide when switching between different operating states, the reciprocating motion trajectory at different speeds is designed; The linear rack and pinion guide rail is controlled to perform reciprocating motion according to the reciprocating motion trajectory at different speeds; During the reciprocating motion, torque feedback data is collected based on the motion state switching status maintained by the motion state machine, and the torque feedback data is subjected to mean filtering.

[0014] In an optional implementation, the step of performing online friction force identification based on the torque feedback data to perform theoretical modeling and obtain the theoretical torque of the guide rail includes: Based on the torque feedback data, the Coulomb friction coefficient and the viscous friction coefficient are determined using the least squares method. The theoretical Coulomb torque of the guide rail is determined based on the Coulomb friction coefficient, and the theoretical viscous friction torque of the guide rail is determined based on the viscous friction coefficient. The theoretical torque of the guide rail is obtained by combining the theoretical inertial torque, the theoretical Coulomb torque, and the theoretical viscous friction torque of the guide rail.

[0015] In an optional implementation, the command data includes speed command data; The steps of determining the theoretical Coulomb torque of the guide rail based on the Coulomb friction coefficient and determining the theoretical viscous friction torque of the guide rail based on the viscous friction coefficient include: The theoretical Coulomb torque of the guide rail is determined based on the time-domain aligned speed command data and the Coulomb friction coefficient. The theoretical viscous friction torque of the guide rail is determined based on the time-domain aligned speed command data and the viscous friction coefficient.

[0016] In an optional implementation, the operation process is divided into an online friction force identification stage and a normal operation stage, and the dynamic threshold function includes a first dynamic threshold function for the online friction force identification stage and a second dynamic threshold function for the normal operation stage. The first dynamic threshold function is constructed based on the collision detection level, the basic threshold, the temporally aligned speed command data, the maximum allowable speed, and the upper limit of the variable threshold. The second dynamic threshold function is constructed based on the collision detection level, the current weight of the moving part, the total weight of the moving part in the online friction identification stage, the temporally aligned speed command data, the basic threshold, and the upper limit of the variable threshold.

[0017] In an optional implementation, the step of performing collision detection based on torque feedback data within the sliding window, the theoretical torque of the guide rail, and a set dynamic threshold function includes: Determine the current stage. If the current stage is the online friction identification stage, obtain the maximum and minimum values ​​of the torque feedback data within the sliding window, calculate the first difference between the maximum and minimum values, and compare the first difference with the first dynamic threshold function to perform collision detection. If the current operation is in the normal operation phase, the torque feedback data and the theoretical torque of the guide rail within the sliding window are obtained, the second difference between the torque feedback data and the theoretical torque of the guide rail is calculated, and the second difference is compared with the second dynamic threshold function to perform collision detection.

[0018] In an optional implementation, the step of comparing the second difference with the second dynamic threshold function to perform collision detection includes: Detect whether the second difference is greater than a set multiple of the second dynamic threshold function. If it is greater than the set multiple of the second dynamic threshold function, then determine that a collision has been triggered. If the value is not greater than a set multiple of the second dynamic threshold function, then it is determined whether the value is greater than the second dynamic threshold function and the duration reaches the set time. If the value is greater than the second dynamic threshold function and the duration reaches the set time, then a collision is determined to be triggered.

[0019] Secondly, the present invention provides a collision detection device for a robot linear rack and pinion guide, the device comprising: The setting module is used to set response delay data and communication delay data for the control system of the robot linear rack and pinion guide. The alignment operation module is used to perform time-domain alignment operations between instruction data and feedback data based on the communication delay data; The design module is used to design a motion state machine by combining the response delay data and the time-domain aligned instruction data. The control and acquisition module is used to control the linear rack and pinion guide to perform reciprocating motion at different speeds, and to acquire torque feedback data based on the switching of the running state corresponding to the motion state machine. The modeling module is used to perform online friction force identification based on the torque feedback data in order to perform theoretical modeling and obtain the theoretical torque of the guide rail; The detection module is used to perform collision detection during operation based on torque feedback data within a sliding window, theoretical guide rail torque, and a set dynamic threshold function, and obtain the detection results.

[0020] Thirdly, the present invention provides an electronic device including a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method described in any of the foregoing embodiments.

[0021] Compared to existing technologies, this invention provides a collision detection method, device, and electronic device for a robot linear rack and pinion guide. It sets response delay data and communication delay data to address the response speed and communication latency of the robot's linear rack and pinion guide control system. Time-domain alignment of command data and feedback data is performed based on the communication delay data. A motion state machine is designed by combining the response delay data and the time-latency aligned command data. The linear rack and pinion guide is controlled to perform reciprocating motion at different speeds, and torque feedback data is collected based on the corresponding operating state switching of the motion state machine. Online friction force identification is performed based on the torque feedback data to perform theoretical modeling and obtain the theoretical torque of the guide rail. During operation, collision detection is performed based on the torque feedback data within a sliding window, the theoretical torque of the guide rail, and a set dynamic threshold function to obtain the detection result.

[0022] In this scheme, a theoretical model of the linear rack and pinion guide is achieved based on motion state switching and online friction force identification. On this basis, a dynamic threshold function is designed, and collision detection of the guide is achieved by combining the command and torque feedback data after time delay compensation. This scheme has the advantages of strong robustness and low false trigger rate, while taking into account the response speed and communication delay of the control system. It can improve the safety of the human-machine interaction process at a lower cost. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart of a robot linear rack and pinion guide collision detection method provided in an embodiment of the present invention; Figure 2 for Figure 1 A flowchart of the sub-steps included in S13; Figure 3 for Figure 1 A flowchart of the sub-steps included in S14; Figure 4 for Figure 1 A flowchart of the sub-steps included in S15; Figure 5 This is a schematic diagram illustrating the implementation state of recognizing speed commands under the collision detection method provided in this embodiment of the invention. Figure 6 This is a functional block diagram of a robot linear rack and pinion guide collision detection device provided in an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention.

[0026] Please see Figure 1 The flowchart below shows a robot linear rack and pinion guide collision detection method provided in an embodiment of the present invention. This robot linear rack and pinion guide collision detection method can be executed by a robot linear rack and pinion guide collision detection device. The robot linear rack and pinion guide collision detection device can be implemented by software and / or hardware and can be configured in an electronic device, such as a computer device, server, robot controller, etc. The detailed steps of the robot linear rack and pinion guide collision detection method are described below.

[0027] S11 sets response delay data and communication delay data for the response speed and communication delay of the control system of the robot linear rack and pinion guide respectively; S12, perform time-domain alignment operation between instruction data and feedback data based on the communication delay data; S13, Design a motion state machine by combining the response delay data and the instruction data after time domain alignment; S14, control the linear rack guide to perform reciprocating motion at different speeds, and collect torque feedback data based on the switching of the running state corresponding to the motion state machine; S15, Based on the torque feedback data, perform online friction force identification to perform theoretical modeling and obtain the theoretical torque of the guide rail; S16, during operation, collision detection is performed based on the torque feedback data within the sliding window, the theoretical torque of the guide rail, and the set dynamic threshold function to obtain the detection results.

[0028] In this embodiment, the specific implementation use is a robot system with a linear rack and pinion guide as the external axis. First, a theoretical modeling method for linear rack and pinion guides based on motion state switching and online friction force identification is provided. Furthermore, a method for dynamic collision detection and response considering load and operating speed is provided.

[0029] Specifically, a delay compensation mechanism is first introduced, setting different delay data, such as the number of delay cycles, for the response speed and communication delay of the robot's linear rack and pinion guide control system. For theoretical dynamic characteristics, the transition of motion states should be accompanied by an instantaneous change in torque. However, constrained by the response speed of the control system, the actual torque response inevitably undergoes a transition process from transient to steady state. The communication between the robot body and the linear rack and pinion guide has an inherent transmission delay, causing the control commands and state feedback to be out of sync in time. Specifically, feedback signals such as speed, acceleration, and torque maintain strict time consistency through a synchronous acquisition mechanism, while the speed / acceleration commands issued by the controller lead the feedback due to the unidirectional transmission delay, resulting in a misalignment of command data and feedback data in the time domain.

[0030] Combining the above two points, this embodiment introduces a delay compensation mechanism, including response delay data used to compensate for the response speed of the control system. and communication delay data used to compensate for communication delays. .

[0031] Based on this, a time-domain alignment operation is performed between command data and feedback data based on communication latency data. Specifically, the time domain alignment of command data (velocity command and acceleration command) and feedback data is achieved using the sliding window method. After alignment via the sliding window, at the current sampling time... At that time, the instruction data and feedback data can be represented as follows: (1) In the formula, and They are respectively The time guide rail command data (including speed command and acceleration command); , and They are respectively Real-time feedback data (including guide rail speed, acceleration, and torque feedback data).

[0032] Then, combine the response delay data with the time-domain aligned instruction data to design the motion state machine. Please refer to [link / reference]. Figure 2 In this embodiment, this step can be implemented in the following way: S131, the linear rack guide is divided into multiple motion states based on actual physical characteristics; S132, initialize multiple counters and set up a motion state machine, wherein each counter is used to quantize different motion states and their switching states, and the motion state machine is used to maintain the switching between the motion states based on the count values ​​of each counter.

[0033] In this embodiment, the motion states are divided based on actual physical characteristics. As mentioned above, for a linear rack and pinion guide that can be approximated as a single-degree-of-freedom system, the switching of motion states should be accompanied by an instantaneous change in torque. From the perspective of theoretical dynamic characteristics, this mainly involves two types of abrupt torques: inertial torque and Coulomb torque. Therefore, the motion states of the linear rack and pinion guide can be divided into four working conditions: A. Uniform Motion (including stationary motion): In this state, the system is in a steady state, and the torque feedback mainly consists of constant torque and disturbance torque. The constant torque during uniform motion includes Coulomb friction torque and viscous friction torque; when stationary, the theoretical value of the constant torque should be zero. B. Acceleration / Deceleration: When the system is in the steady-state acceleration / deceleration phase, the torque feedback mainly consists of inertial torque, Coulomb torque, and velocity-related viscous friction torque. C. Transition from acceleration / deceleration to uniform velocity: This stage is the transient process of changing the state of motion. Due to the sudden application or removal of the inertial torque, the system exhibits obvious transient response characteristics; D. Velocity reversal: This stage is also a transient process. Due to the reversal of the motion direction, the Coulomb friction torque undergoes a sudden change in sign, and the step change in torque puts the system in a transient phase.

[0034] Based on this, combined with response latency data The motion state machine is designed based on time-domain aligned instruction data. First, five counters are defined and initialized: stationary, constant speed, acceleration, deceleration, and speed reversal. Second, the increment and reset conditions for the five counters are explained as follows: 1) "Stationary" counter: When the linear rack and pinion guide is not moving, the value increases with the control cycle. To prevent the value from overflowing when the function is activated, the value is limited. When the guide moves, the value is reset. 2) "Acceleration" counter: When the acceleration command after delay compensation is greater than 0, it increments with the control cycle; when the acceleration command is less than 0, or when the acceleration command is 0 but the "Station" counter value has not reached the limit, the value is reset. 3) "Deceleration" counter: When the acceleration command after delay compensation is less than 0, it increments with the control cycle; when the acceleration command is greater than 0, or when the acceleration command is 0 but the "Station" counter value has not reached the limit, this value is reset. 4) "Uniform Speed" Counter: When the acceleration command after delay compensation is 0 and the value of the "Stationary" counter has not reached the limit, it increments with the control cycle; when the acceleration command is not 0, this value is reset. 5) "Velocity Reversal" Counter: When the motion state is "D-Velocity Reversal" and the counter value has not reached... When the motion state is "B-acceleration / deceleration" and the counter value is [value missing], the counter increments with the control cycle. The value will be reset.

[0035] Finally, based on the counter values, the transition conditions of the motion state machine are explained: 1) "A-Uniform Motion (including stationary)" motion state: The "uniform speed" counter is greater than or equal to When the time comes, transition to that state; 2) "B-Acceleration / Deceleration" motion state: The "Acceleration" and "Deceleration" counters are greater than or equal to And the "speed reversal" counter is 0 or When the time comes, transition to that state; 3) Motion state of "C-adjustment between acceleration / deceleration and uniform speed": The "acceleration" counter is greater than 0 and less than 1. Or the "deceleration" counter is greater than 0 and less than 0. Or, the "uniform speed" counter is greater than 0 and less than 0. When the time comes, transition to that state; 4) "D-velocity reversal" motion state: at the current sampling time The transition condition for this state can be represented by the following equation: (2) In the formula, The speed command for the previous moment after delay compensation; The speed command for the current moment after delay compensation.

[0036] Based on this, the linear rack and pinion guide is controlled to perform reciprocating motion at different speeds. Torque feedback data is collected based on the switching of the operating states corresponding to the motion state machine. Please refer to [link / reference]. Figure 3 This step can be achieved in the following ways: S141, Design the reciprocating motion trajectory at different speeds based on the torque change law of the linear rack guide when switching between different operating states; S142, control the linear rack guide to perform reciprocating motion according to the reciprocating motion trajectory at different speeds; S143, during the reciprocating motion, torque feedback data is collected based on the motion state switching status maintained by the motion state machine, and the torque feedback data is subjected to mean filtering.

[0037] First, based on the above analysis of the state switching process of the linear rack guide during motion, the T-type speed programming can be divided into two cases: those with and without a uniform speed segment. Specifically, there are: 1) No uniform speed segment: "A-uniform speed (including stationary)" → "C-transition between acceleration / deceleration and uniform speed" → "B-acceleration / deceleration" → "C-transition between acceleration / deceleration and uniform speed" → "B-acceleration / deceleration" → "C-transition between acceleration / deceleration and uniform speed" → "A-uniform speed (including stationary)"; 2) There is a uniform speed segment: "A-uniform speed (including stationary)" → "C-transition between acceleration / deceleration and uniform speed" → "B-acceleration / deceleration" → "C-transition between acceleration / deceleration and uniform speed" → "A-uniform speed (including stationary)" → "C-transition between acceleration / deceleration and uniform speed" → "B-acceleration / deceleration" → "C-transition between acceleration / deceleration and uniform speed" → "A-uniform speed (including stationary)".

[0038] As mentioned above, the components of torque typically differ under different motion states. Therefore, during state transitions, torque may exhibit continuous changes from its original value or undergo abrupt changes. Taking a uniform velocity segment as an example, from a decomposition perspective, we have: 1) "A-Uniform speed (including stationary)" → "C-Transition between acceleration / deceleration and uniform speed": Initial value of torque feedback; 2) "C - The transition from acceleration / deceleration to uniform speed" → "B - Acceleration / deceleration": Superimposed inertial torque + Coulomb torque (acceleration); 3) "B-Acceleration / Deceleration" → "C-Transition between Acceleration / Deceleration and Uniform Motion": Superimposed increasing viscous friction torque (acceleration); 4) "C - The transition between acceleration / deceleration and uniform speed" → "A - Uniform speed (including stationary)": Cancel the inertial torque; 5) "A - Uniform speed (including stationary)" → "C - Transition between acceleration / deceleration and uniform speed": remain unchanged; 6) "C - The transition from acceleration / deceleration to uniform speed" → "B - Acceleration / deceleration": Superimposed inertial torque (deceleration); 7) "B-Acceleration / Deceleration" → "C-Transition between Acceleration / Deceleration and Uniform Motion": Superimposed decreasing viscous friction torque (deceleration); 8) "C - The transition between acceleration / deceleration and uniform speed" → "A - Uniform speed (including stationary speed)": Cancel the inertial torque + Coulomb torque.

[0039] Based on the physical laws governing torque changes during motion state transitions, a reciprocating motion trajectory for the long stroke (ensuring a relatively long uniform speed segment) of the guide rail can be designed at different operating speeds. This trajectory is then used to control the linear rack guide rail to perform reciprocating motion at different speeds. Torque feedback data is collected during the guide rail's motion, coinciding with motion state transitions. Considering the measurement noise from the sensors, the torque feedback data is filtered by mean, as shown below: (3) In the formula, The current sampling time Torque feedback data after mean filtering; To adjust the sliding window size; for At time 1, the original torque of the guide rail is fed back.

[0040] Based on the above, online friction force identification is performed using torque feedback data to perform theoretical modeling and obtain the theoretical torque of the guide rail. For details, please refer to [link to relevant documentation]. Figure 4 This step can be achieved in the following ways: S151, Based on the torque feedback data, the Coulomb friction coefficient and the viscous friction coefficient are determined using the least squares method; S152, the theoretical Coulomb torque of the guide rail is determined based on the Coulomb friction coefficient, and the theoretical viscous friction torque of the guide rail is determined based on the viscous friction coefficient; S153, combining the theoretical inertial torque, theoretical Coulomb torque, and theoretical viscous friction torque of the guide rail, the theoretical torque of the guide rail is obtained.

[0041] Torque feedback data is recorded based on the motion state switching during the reciprocating motion of the guide rail. ,have: 1) When the motion state changes from "B - acceleration / deceleration" to "C - transition between acceleration / deceleration and uniform motion", or from "A - uniform motion (including stationary)" to "C - transition between acceleration / deceleration and uniform motion", the torque feedback data is the original data. ; 2) When the motion state switching situation is different from point 1, the torque feedback data is the filtered data. .

[0042] Once the data recording is complete, and considering the decomposed perspective of motion state switching, the change in viscous friction torque during forward and reverse rotation can be obtained from equations (4) and (5): (4) (5) In the formula, and These represent the changes in viscous friction torque during forward and reverse rotation. From the perspective of decomposing the uniform velocity segment, there are 8 stages, while the forward and reverse reciprocating motion has 16 stages. The corresponding changes in viscous friction torque are the differences between different speeds in the same direction of motion. The sample size of the running speed in the reciprocating motion trajectory (e.g., 10%, 30% of the global speed).

[0043] From equations (4) and (5), combined with the least squares identification method, the viscous friction coefficients during forward and reverse rotation can be obtained: (6) (7) In the formula, and These are the identification values ​​of the viscous friction coefficient during forward and reverse rotation, respectively. This refers to the interval between different running speeds in the reciprocating motion trajectory. This value remains consistent in this invention, meaning the speed is... Linearly increasing.

[0044] From equations (6) and (7), the Coulomb friction coefficients for both forward and reverse rotation can be further obtained: (8) (9) In the formula, and These are the Coulomb friction coefficients for forward and reverse rotation, respectively. The initial velocity of the guide rail's reciprocating motion.

[0045] An approximate theoretical dynamic model of the linear rack guide is established, and the theoretical torque is calculated using velocity and acceleration commands. The approximate theoretical dynamic model of the linear rack guide can be simplified as follows: (10) In the formula, This is the theoretical torque of the guide rail; The theoretical inertial torque of the guide rail; The theoretical Coulomb torque for the guide rail; This refers to the viscous friction torque in the guide rail theory.

[0046] The theoretical inertial torque of the guide rail can be calculated from the acceleration command and the mass of each component. The translational motion of the robot and its base is driven by the output of the guide rail motor. Therefore, at the current sampling moment... According to Newton's second law: (11) In the formula, It is a translational inertial force; The total mass of the moving parts; The acceleration command is after delay compensation; For the mass of the robot body; For the end-effector load mass; For the quality of the robot's base.

[0047] Based on equation (11), the inertial force is converted into an inertial torque, which can be calculated by equation (12): (12) In the formula, The radius of the output shaft gear of the linear rack and pinion motor.

[0048] Furthermore, the theoretical Coulomb torque and the theoretical viscous friction torque of the guide rail can be obtained in the following ways: The theoretical Coulomb torque of the guide rail is determined based on the time-domain aligned speed command data and the Coulomb friction coefficient; the theoretical viscous friction torque of the guide rail is determined based on the time-domain aligned speed command data and the viscous friction coefficient.

[0049] Specifically, the theoretical Coulomb torque and the theoretical viscous friction torque of the guide rail can be calculated using equations (13) to (14): (13) (14) Thus, the theoretical torque of the linear rack guide can be obtained from equations (10) to (14). .

[0050] Based on the above, this embodiment provides a method for dynamic collision detection and response that considers load and operating speed. During operation, collision detection is performed based on torque feedback data within a sliding window, theoretical guide rail torque, and a set dynamic threshold function to obtain the detection result.

[0051] When a linear rack and pinion guide moves, torque feedback is subject to measurement errors and disturbances. Therefore, the collision trigger duration needs to be set appropriately based on measured data. The time is typically 150~300ms. Considering that the theoretical torque is inaccurate during the online friction identification stage, the operation process is divided into an online friction identification stage and a normal operation stage. The dynamic threshold function includes a first dynamic threshold function for the online friction identification stage and a second dynamic threshold function for the normal operation stage.

[0052] The first dynamic threshold function is constructed based on the collision detection level, the base threshold, the temporally aligned velocity command data, the maximum allowable speed, and the upper limit of the variable threshold. Specifically, the first dynamic threshold function is constructed as follows: (15) In the formula, This is the first dynamic threshold function during the online friction force identification stage; This represents the collision detection level, with an amplitude of 0~100. Basic threshold; This is the upper limit of the variable threshold.

[0053] The second dynamic threshold function is constructed based on the collision detection level, the current weight of the moving part, the total weight of the moving part during the online friction identification phase, the temporally aligned velocity command data, the basic threshold, and the upper limit of the variable threshold.

[0054] Specifically, the second dynamic threshold function is constructed as follows: (16) In the formula, This is the second dynamic threshold function during normal operation. This represents the collision detection level, with an amplitude of 0~100. The total weight of moving parts during the online friction force identification phase; The current weight of the moving part; Basic threshold; This is the upper limit of the variable threshold.

[0055] In this embodiment, the dynamic threshold function incorporates the effects of load and operating speed. This design is reasonable because as speed or load increases, the system becomes more sensitive to external disturbances, and the amplitude of sensor measurement noise also increases accordingly.

[0056] Based on this, the steps described above for performing collision detection based on torque feedback data within the sliding window, theoretical guide rail torque, and a set dynamic threshold function can be implemented in the following way: Determine the current stage. If the current stage is the online friction identification stage, obtain the maximum and minimum values ​​of the torque feedback data within the sliding window, calculate the first difference between the maximum and minimum values, and compare the first difference with the first dynamic threshold function to perform collision detection. If the current operation is in the normal operation phase, the torque feedback data and the theoretical torque of the guide rail within the sliding window are obtained, the second difference between the torque feedback data and the theoretical torque of the guide rail is calculated, and the second difference is compared with the second dynamic threshold function to perform collision detection.

[0057] In this embodiment, the maximum and minimum values ​​of the torque feedback data are obtained based on historical torque feedback, and the duration and threshold function are combined to determine whether a collision has occurred. During the first stage of operation, i.e., the online friction identification stage, when performing mean filtering on the torque feedback data, the maximum value of the torque feedback data is obtained by traversing within a sliding window. and minimum value And calculate the difference between the two. : (17) when The system starts counting with each control cycle, and counts continue until the collision duration reaches the set time. A collision can be determined to have been triggered. Furthermore, considering the possibility of a sudden, massive impact, when... At that time, a collision can be directly determined.

[0058] During the second stage of operation, i.e., the normal operation stage, the theoretical torque of the guide rail is calculated using equations (10) to (14), and the difference between the theoretical torque of the guide rail and the torque feedback data is obtained. : (18) Similarly, when The system starts counting with each control cycle, and counts continue until the collision duration reaches the set time. A collision can be determined to have been triggered. Furthermore, considering the possibility of a sudden, massive impact, when... At that time, a collision can be directly determined.

[0059] When a collision is detected, a collision is triggered, and the guide rail and robot should execute a safety response strategy to enhance the safety of the interaction. The safety response strategies are to stop movement and bounce back upon collision. The former involves issuing a stop command upon collision, causing the guide rail and robot to come to an abrupt stop; the latter involves causing the guide rail to move slightly in the opposite direction of the collision to actively evacuate from the collision source.

[0060] The collision detection scheme provided in this embodiment theoretically models the linear rack and pinion guide rail based on motion state switching and online friction force identification. On this basis, a dynamic threshold function considering load and operating speed is designed, combined with time-delay compensated commands and torque feedback data to achieve collision detection of the guide rail and execute a safety response strategy. The proposed scheme has the advantages of strong robustness and low false trigger rate, while also taking into account the response speed and communication latency of the control system, thus improving the safety of the human-machine interaction process at a lower cost.

[0061] Combination Figure 5 The diagram illustrates the accuracy of motion state switching recognition in the solution provided by this invention under T-shaped velocity planning. The blue line represents the planned velocity command curve with a constant speed segment, and the orange line represents the real-time status of the velocity command recognition based on the motion state switching logic defined in this invention. As shown in the diagram, the motion state switching is accurate (where 1 represents acceleration / deceleration; 2 represents constant speed (including stationary); and 3 represents the transition between acceleration / deceleration and constant speed).

[0062] It should be noted that the implementation of the solution provided in this embodiment is not limited to systems where linear rack and pinion guides serve as the external axis of a robot. It can also be used for other translational mechanisms to carry movable loads. Other mechanisms can be linear shafts driven by ball screws.

[0063] In summary, the collision detection method provided in this embodiment proposes a theoretical modeling method for linear rack and pinion guides based on motion state switching and online friction force identification. First, this method introduces a delay compensation mechanism to align the velocity and acceleration commands with the feedback data in the time domain. Second, based on actual physical characteristics, the guide rail motion is divided into different states, and corresponding motion state machines are designed. Third, the guide rail executes a reciprocating motion trajectory, and combined with online data acquisition from the state machine, the Coulomb-viscosity friction coefficient is identified. Finally, based on Newton's second law, an approximate theoretical dynamic model of the linear rack and pinion guide is established. This method effectively avoids the high false trigger rate problem caused by the control system's response speed limitations and communication delays, and achieves approximate modeling based on the physical characteristics of the guide rail motion and motion state switching, significantly reducing the modeling difficulty while ensuring model accuracy.

[0064] On the other hand, a dynamic collision detection and response method considering load and operating speed is proposed. This method divides system operation into two stages: online friction force identification and normal operation, and designs dynamic threshold functions that vary with load and operating speed for each stage. Combining guide rail torque feedback and theoretical model predictions, collision detection during guide rail movement is achieved, triggering a safety response. This method effectively balances the sensitivity and false trigger rate of collision detection, avoiding the failure problem of fixed thresholds under varying operating conditions. Simultaneously, the integrated safety response mechanism ensures a rapid closed loop from collision perception to action execution, significantly improving the system's safety and reliability.

[0065] Based on the same inventive concept, please refer to Figure 6 This invention also provides a robot linear rack and pinion guide collision detection device, which includes: The setting module is used to set response delay data and communication delay data for the control system of the robot linear rack and pinion guide. The alignment operation module is used to perform time-domain alignment operations between instruction data and feedback data based on the communication delay data; The design module is used to design a motion state machine by combining the response delay data and the time-domain aligned instruction data. The control and acquisition module is used to control the linear rack and pinion guide to perform reciprocating motion at different speeds, and to acquire torque feedback data based on the switching of the running state corresponding to the motion state machine. The modeling module is used to perform online friction force identification based on the torque feedback data in order to perform theoretical modeling and obtain the theoretical torque of the guide rail; The detection module is used to perform collision detection during operation based on torque feedback data within a sliding window, theoretical guide rail torque, and a set dynamic threshold function, and obtain the detection results.

[0066] The robot linear rack and pinion guide collision detection device provided in this embodiment can be used to execute the robot linear rack and pinion guide collision detection method under any of the above embodiments. For details not covered in this embodiment, please refer to the corresponding descriptions in the above embodiments. This embodiment will not elaborate further here.

[0067] Furthermore, embodiments of the present invention also provide an electronic device, which may be such as a computer device, a server, a robot controller, etc. The electronic device includes a processor and a memory. The memory stores computer-executable instructions that can be executed by the processor. The processor executes the computer-executable instructions to implement the robot linear rack and pinion guide collision detection method in any of the above embodiments.

[0068] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0069] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0070] Furthermore, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0071] It should be noted that if the functionality is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0072] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0073] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A collision detection method for a robot linear rack and pinion guideway, characterized in that, The method includes: Response delay data and communication delay data are set for the response speed and communication delay of the control system of the robot linear rack and pinion guide respectively; Based on the communication delay data, perform a time-domain alignment operation between the instruction data and the feedback data; Design a motion state machine by combining the response delay data and the time-domain aligned instruction data; The linear rack and pinion guide is controlled to perform reciprocating motion at different speeds, and torque feedback data is collected based on the switching of the running state corresponding to the motion state machine. Based on the torque feedback data, online friction force identification is performed to conduct theoretical modeling and obtain the theoretical torque of the guide rail; During operation, collision detection is performed based on torque feedback data within the sliding window, theoretical torque of the guide rail, and the set dynamic threshold function to obtain the detection results.

2. The collision detection method for a robot linear rack and pinion guideway according to claim 1, characterized in that, The steps of designing a motion state machine by combining the response delay data and the time-domain aligned instruction data include: The linear rack and pinion guide is divided into multiple motion states based on its actual physical characteristics; Multiple counters are initialized, and a motion state machine is set up. Each counter is used to quantize different motion states and their switching states, and the motion state machine is used to maintain the switching between the motion states based on the count values ​​of each counter.

3. The collision detection method for a robot linear rack and pinion guideway according to claim 1, characterized in that, The step of controlling the linear rack guide to perform reciprocating motion at different speeds and collecting torque feedback data based on the switching of the operating state corresponding to the motion state machine includes: Based on the torque variation law of the linear rack and pinion guide when switching between different operating states, the reciprocating motion trajectory at different speeds is designed; The linear rack and pinion guide rail is controlled to perform reciprocating motion according to the reciprocating motion trajectory at different speeds; During the reciprocating motion, torque feedback data is collected based on the motion state switching status maintained by the motion state machine, and the torque feedback data is subjected to mean filtering.

4. The collision detection method for a robot linear rack and pinion guideway according to claim 1, characterized in that, The step of performing online friction force identification based on the torque feedback data to perform theoretical modeling and obtain the theoretical torque of the guide rail includes: Based on the torque feedback data, the Coulomb friction coefficient and the viscous friction coefficient are determined using the least squares method. The theoretical Coulomb torque of the guide rail is determined based on the Coulomb friction coefficient, and the theoretical viscous friction torque of the guide rail is determined based on the viscous friction coefficient. The theoretical torque of the guide rail is obtained by combining the theoretical inertial torque, the theoretical Coulomb torque, and the theoretical viscous friction torque of the guide rail.

5. The collision detection method for a robot linear rack and pinion guideway according to claim 4, characterized in that, The instruction data includes speed instruction data; The steps of determining the theoretical Coulomb torque of the guide rail based on the Coulomb friction coefficient and determining the theoretical viscous friction torque of the guide rail based on the viscous friction coefficient include: The theoretical Coulomb torque of the guide rail is determined based on the time-domain aligned speed command data and the Coulomb friction coefficient. The theoretical viscous friction torque of the guide rail is determined based on the time-domain aligned speed command data and the viscous friction coefficient.

6. The collision detection method for a robot linear rack and pinion guideway according to claim 1, characterized in that, The operation process is divided into an online friction force identification stage and a normal operation stage. The dynamic threshold function includes a first dynamic threshold function for the online friction force identification stage and a second dynamic threshold function for the normal operation stage. The first dynamic threshold function is constructed based on the collision detection level, the basic threshold, the temporally aligned speed command data, the maximum allowable speed, and the upper limit of the variable threshold. The second dynamic threshold function is constructed based on the collision detection level, the current weight of the moving part, the total weight of the moving part in the online friction identification stage, the temporally aligned speed command data, the basic threshold, and the upper limit of the variable threshold.

7. The collision detection method for a robot linear rack and pinion guideway according to claim 6, characterized in that, The step of performing collision detection based on torque feedback data within the sliding window, theoretical guide rail torque, and a set dynamic threshold function includes: Determine the current stage. If the current stage is the online friction identification stage, obtain the maximum and minimum values ​​of the torque feedback data within the sliding window, calculate the first difference between the maximum and minimum values, and compare the first difference with the first dynamic threshold function to perform collision detection. If the current operation is in the normal operation phase, the torque feedback data and the theoretical torque of the guide rail within the sliding window are obtained, the second difference between the torque feedback data and the theoretical torque of the guide rail is calculated, and the second difference is compared with the second dynamic threshold function to perform collision detection.

8. The collision detection method for a robot linear rack and pinion guideway according to claim 7, characterized in that, The step of comparing the second difference with the second dynamic threshold function to perform collision detection includes: Detect whether the second difference is greater than a set multiple of the second dynamic threshold function. If it is greater than the set multiple of the second dynamic threshold function, then determine that a collision has been triggered. If the value is not greater than a set multiple of the second dynamic threshold function, then it is determined whether the value is greater than the second dynamic threshold function and the duration reaches the set time. If the value is greater than the second dynamic threshold function and the duration reaches the set time, then a collision is determined to be triggered.

9. A collision detection device for a robot linear rack and pinion guideway, characterized in that, The device includes: The setting module is used to set response delay data and communication delay data for the control system of the robot linear rack and pinion guide. The alignment operation module is used to perform time-domain alignment operations between instruction data and feedback data based on the communication delay data; The design module is used to design a motion state machine by combining the response delay data and the time-domain aligned instruction data. The control and acquisition module is used to control the linear rack and pinion guide to perform reciprocating motion at different speeds, and to acquire torque feedback data based on the switching of the running state corresponding to the motion state machine. The modeling module is used to perform online friction force identification based on the torque feedback data in order to perform theoretical modeling and obtain the theoretical torque of the guide rail; The detection module is used to perform collision detection during operation based on torque feedback data within a sliding window, theoretical guide rail torque, and a set dynamic threshold function, and obtain the detection results.

10. An electronic device, characterized in that, The method includes a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method of any one of claims 1 to 8.

Images