Robot reducer transmission precision prediction method based on multi-source information fusion

Abstract

The present application relates to the technical field of robot transmission system detection, and more particularly to a robot reducer transmission precision prediction method based on multi-source information fusion. The method comprises the following steps: determining a continuous working condition collection interval according to a single-tooth complete meshing stroke of a robot reducer preset; collecting gear meshing vibration time domain values and output shaft rotation angle time domain values in the continuous working condition collection interval; taking the time sequence of each meshing stroke physical point as a reference, aligning the meshing vibration time domain values and the rotation angle time domain values in the same continuous working condition collection interval in time sequence, and generating meshing time sequence coupling data sets. The present application differentiates the single-tooth meshing point of the robot reducer to obtain incremental values, calculates transmission errors in sections and steps, and iteratively collects them, so as to realize global accurate quantification of the meshing transmission state of the robot reducer, and finally improve the precision of the robot reducer transmission precision prediction result.

Description

Technical Field

[0001] This invention relates to the field of robot transmission system testing technology, and in particular to a method for predicting the transmission accuracy of robot reducers based on multi-source information fusion. Background Technology

[0002] Robot reducers are core precision transmission components in the joints of intelligent robots. During operation, the internal paired gear pairs continuously perform single-tooth meshing actions. The teeth sequentially complete the entire process of meshing engagement, stable contact, and disengagement. The contact forces at different positions on the tooth surface and the rotational state of the shaft system are dynamically changing. The single-tooth meshing condition directly determines the overall transmission performance of the reducer. Existing technologies for predicting transmission accuracy in this type of reducer do not define a specific data collection range based on the actual meshing stroke of a single tooth. The acquired transmission-related data is mixed with irrelevant operating condition information. Different types of transmission monitoring data lack a unified reference standard for meshing points, making it difficult to establish a stable correspondence between data points. The acquired data is difficult to accurately reflect the actual meshing operation, resulting in a rather general transmission accuracy judgment that fails to accurately reflect the actual transmission status throughout the entire single-tooth meshing process. Summary of the Invention

[0003] Therefore, it is necessary to provide a method for predicting the transmission accuracy of robot reducers based on multi-source information fusion to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion is provided, the method comprising the following steps: Step S1: Determine the continuous operating condition acquisition interval based on the preset single-tooth complete meshing stroke of the robot reducer; acquire the time-domain values ​​of gear meshing vibration and the time-domain values ​​of output shaft rotation angle within the continuous operating condition acquisition interval; Step S2: Based on the time sequence of physical points of each meshing stroke, time-domain values ​​of meshing vibration and time-domain values ​​of rotation angle within the same continuous working condition acquisition interval are time-series aligned to generate a meshing time-series coupled dataset. Step S3: Perform differential operation on the meshing time-coupling dataset to calculate the real-time increment of vibration amplitude and the real-time offset increment of rotation angle corresponding to a single meshing point. Step S4: Based on the continuous meshing and traveling sequence of gears, and using the real-time increment of vibration amplitude and the real-time offset increment of rotation angle as prediction references, predict the tooth surface meshing transmission error and shaft rotation transmission error step by step, and obtain the real-time transmission accuracy value of the robot reducer through iterative calculation.

[0005] The beneficial effects of the present invention are as follows: I. This invention defines the continuous operating condition acquisition interval by using the critical coordinates of the start and end of gear tooth meshing. Vibration and time-series angle data are acquired only within the angular range corresponding to the complete meshing of a single tooth. Simultaneously, each meshing entity point is marked with a specific entry time node, and the two types of acquired values ​​at the same time node are grouped into the same spatial coordinate group to achieve time-series alignment. This operation ensures that all acquired data strictly corresponds to the actual meshing operation period and physical contact position of the gear teeth. The arrangement order of various time-domain values ​​perfectly matches the circumferential rotation rhythm of the gear, allowing the acquired and processed data to accurately correspond to the actual transmission operation state of the reducer. This ensures that the data itself has a high degree of operating condition adaptability and also establishes a stable positional correspondence between various types of raw data.

[0006] Second, this invention utilizes the difference between adjacent meshing points to obtain a specific incremental value at a fixed point. Then, combining the actual contact state and transmission structure characteristics of different meshing areas of the outer, middle, and inner rings of the gear, it separately calculates two types of transmission error points. Finally, it iteratively accumulates and summarizes the errors along the meshing path. This fixed-point difference calculation combined with partitioned differential calculation allows for data processing based on the actual operating performance of different meshing positions of the gear teeth. Based on the data calculation results of the entire single-tooth meshing process, it directly derives the corresponding real-time transmission accuracy value, ensuring that the final transmission accuracy result accurately reflects the true transmission operating state of the reducer at each position throughout the entire single-tooth meshing process. Attached Figure Description

[0007] Figure 1 This is a flowchart illustrating the steps of a method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion. Figure 2 This is a schematic diagram of the single-tooth meshing stroke structure of a robot reducer in one embodiment; Figure 3 This is a schematic diagram of the gear pitch circle path in one embodiment; Figure 4 This is a schematic diagram of the reducer output shaft rotation structure in one embodiment; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0008] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0009] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0010] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0011] To achieve the above objectives, please refer to Figures 1 to 4 A method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion, the method comprising the following steps: Preferably, step S1: determine the continuous working condition acquisition interval based on the preset single-tooth complete meshing stroke of the robot reducer; and acquire the time-domain values ​​of gear meshing vibration and the time-domain values ​​of output shaft rotation angle within the continuous working condition acquisition interval. Optionally, in step S1, determining the continuous operating condition data acquisition interval based on the preset single-tooth complete meshing stroke of the robot reducer is specifically as follows: The spatial coordinates of the working surfaces of the tooth tip and tooth root in the paired meshing gear pair inside the reducer of the positioning robot; Mark the critical coordinate positions where the driving gear teeth smoothly enter the tooth groove of the driven gear in sequence, and mark the critical coordinate positions where the teeth completely disengage from the tooth groove contact area. Calculate the complete angular coverage area occupied by the gear when it completes a circular rotation between two critical coordinate points, and add positioning coordinate markers at the beginning and end of this angular range. Based on the complete angular coverage formed after adding positioning coordinate markers, a dedicated data acquisition range is defined to adapt to the complete meshing rotation process of a single tooth.

[0012] In one embodiment, parameter positioning is performed on the standard involute spur gear pair of the robot reducer. A high-precision spatial coordinate acquisition device collects the coordinates of the working surfaces of the paired meshing gear pair inside the reducer. The coordinate acquisition range covers the rotation range of the entire gear tooth profile, accurately locks the three-dimensional spatial coordinate points of the working surfaces of the gear pair tooth tip and tooth root, records the maximum radial coordinate data of the outer edge of the working surface of the tooth tip and the minimum radial coordinate data of the inner edge of the working surface of the tooth root, and uses the gear rotation center as the reference origin. All collected coordinate data are uniformly calibrated and stored.

[0013] In another embodiment, the robot reducer maintains a standard meshing state at a rated constant speed, and the gear polar coordinate measurement system implements a unified benchmark calibration for the meshing critical points. The system identifies the critical coordinate position of the starting meshing point when the driving gear tooth cuts into the driven gear tooth groove. This coordinate position corresponds to the spatial point where the tooth tip first makes contact with the tooth groove wall. At the same time, it identifies the critical coordinate position of the ending meshing point when the driving gear tooth disengages from the driven gear tooth groove. This coordinate position corresponds to the spatial point where the tooth root completely separates from the tooth groove wall. The two sets of critical coordinate points correspond to the starting and ending boundaries of the single-tooth meshing process, respectively.

[0014] It should be noted that the gear center polar angle is used as a measurement standard to calculate the difference in polar angle between the critical point of meshing initiation and the critical point of meshing termination. This difference in polar angle constitutes the angular span of the gear's circumferential rotation corresponding to a complete single-tooth meshing. The angular span of a single-tooth meshing in the robot reducer is maintained within the range of 22° to 25°. Dedicated positioning coordinate markers are set at the beginning and end points of the angular span, and the coordinate parameters of the markers are completely consistent with the coordinate parameters of the corresponding critical points.

[0015] Two sets of critical coordinates with positioning markers constitute boundary constraints, defining a closed angular coverage area for gear rotation. This area covers the entire rotational stroke of a single gear tooth during engagement, contact transmission, and disengagement. This angular coverage area is defined as the continuous operating condition acquisition range corresponding to the complete meshing condition of a single tooth. The boundary coordinate parameters of the area remain continuously locked, providing a fixed operating condition acquisition boundary range for the synchronous acquisition of the time-domain values ​​of gear meshing vibration and the time-domain values ​​of the output shaft rotation angle.

[0016] See Figure 2 In another embodiment, Figure 2This diagram illustrates the single-tooth meshing stroke of a robot reducer, clearly showing the meshing state of the driving and driven gears with the gear's rotation center as the reference. The fan-shaped area in the diagram defines the complete angular stroke of a single tooth from engagement, steady-state contact to disengagement. The black dots at both ends mark the critical coordinate points for the start and end of meshing. In this embodiment, the three-dimensional coordinates of the tooth tip and root working surfaces are locked using a high-precision spatial coordinate acquisition device to determine the critical points for the start and end of meshing. The 22°-25° angular span between these two points represents the complete meshing stroke of a single tooth, providing a physical boundary for subsequently defining continuous operating condition acquisition intervals and intuitively reflecting the spatial positional changes during the single-tooth meshing process.

[0017] Optionally, the specific steps for acquiring the time-domain values ​​of gear meshing vibration in step S1 are as follows: On the complete trajectory of the robot reducer gear as it completes its full circular rotation, multiple sets of fixed coordinate acquisition points with consistent spacing are divided. Within the defined angular acquisition range, complete the acquisition of vibration data corresponding to each fixed coordinate acquisition point; Based on the continuous changes in coordinate position generated during the continuous rotation of the gear, the vibration data is arranged and combined in an orderly manner to form the time-domain numerical value of gear meshing vibration.

[0018] In one embodiment, a unified spatial positioning system is established using the center point of gear rotation as a coordinate reference. The overall rotational motion of the gear covers a complete 360° angular range. Following the actual trajectory of the gear participating in power transmission, multiple fixed acquisition positions are arranged sequentially. The angular interval between two adjacent acquisition positions is uniformly set to 0.5°. Sufficient acquisition positions are arranged within the complete rotational path of the gear, and all acquisition positions are aligned radially with the meshing contact surface of the gear pitch circle.

[0019] The pre-defined angular acquisition range for single-tooth meshing, spanning 22° to 25°, was retrieved. All fixed acquisition positions within this range were selected from the overall rotational trajectory, resulting in 44 to 50 effective acquisition positions. The vibration sensing component was then fitted onto the outer side of the reducer housing in the area corresponding to the acquisition position, with its sensing contact surface directly facing the gear meshing transmission area. The component's operating frequency was set to 20,000 times per second, and the sensing end of the component maintained a fixed perpendicular alignment with the gear meshing working surface. As the gear passed each effective acquisition position, the vibration sensing component directly captured the mechanical vibration signal generated during gear meshing at that position. The signal acquisition time was consistent with the time it took for a single set of teeth to completely traverse the meshing range.

[0020] In another embodiment, the reducer gear maintains a constant operating speed to complete a continuous circular rotation. Fixed acquisition positions pass through the meshing area in sequence as the gear rotates. The signal recording carrier sequentially collects the original vibration information generated at different spatial positions. Each set of vibration information is synchronously bound to the spatial coordinate information and angular position information of the corresponding acquisition position. The information collection process proceeds sequentially according to the actual rotation process of the gear. All kinds of original vibration information are centrally collected and stored according to the position changes formed by the gear's movement.

[0021] The signal recording carrier systematically organizes and arranges all the original vibration information stored in the collection according to the order of the positions formed by the physical rotation of the gears. It completes the continuous splicing and combination of vibration information collected at different operating periods and different spatial positions. All the information that has been sorted and arranged is arranged and combined according to the actual rotation sequence of the gears. After the arrangement and combination, the time domain values ​​of gear meshing vibration are formed with a continuous temporal arrangement and corresponding spatial positions. Each value content and the corresponding acquisition position coordinates and gear rotation sequence form a fixed correspondence, which constitutes the time domain values ​​of gear meshing vibration.

[0022] See Figure 3 In another embodiment, Figure 3 This is a schematic diagram of the gear pitch circle path. Using the gear's rotation center as a reference, equally spaced point markers are arranged along the gear's meshing rotation trajectory, with an angular interval of 0.5° between adjacent points. In this embodiment, a sufficient number of points are arranged within the complete gear rotation path. Points located within the 22°-25° single-tooth meshing angle range are selected as effective acquisition positions, with 44-50 sets of points. A vibration sensing component is fitted onto the outer side of the reducer housing corresponding to the acquisition position area. The component captures the meshing vibration signal at each point at a frequency of 20,000 times / second. These points are evenly distributed along the gear pitch circle, visually presenting the distribution of the meshing points.

[0023] Optionally, the acquisition of the time-domain value of the output shaft rotation angle in step S1 is specifically as follows: Throughout the entire process of continuous angular position change of the meshing contact surface of the gears, various angular orientation related data generated by the shaft structure at the power output end of the reducer are simultaneously recorded. Within the uniformly defined angular acquisition range, at any meshing coordinate position along the gear's rotation path, the angular value generated by the corresponding position of the output shaft is synchronously matched and recorded. Based on the physical arrangement of the coaxial series transmission inside the reducer, all the recorded angular values ​​are integrated and arranged to form the time domain value of the output shaft rotation angle.

[0024] In one embodiment, a high-precision angle sensing component is installed on the end face of the output shaft of the reducer. The angle sensing component is fixedly installed coaxially with the output shaft, and the component follows the output shaft to complete the circumferential rotation. The angle sensing component uses the gear rotation center as a unified coordinate reference and continuously captures the angular orientation change information of the output shaft during the continuous angular displacement of the gear meshing contact surface. It dynamically collects data throughout the entire process, following the running rhythm of the gear meshing rotation. The data collection action is completely synchronized with the gear meshing angular change, adapting to the continuous operation of the reducer.

[0025] The system retrieves a pre-defined single-tooth meshing angle acquisition range of 22° to 25° and locks all 44 to 50 sets of meshing coordinate points of the gear within this fixed angle range. During gear operation, it traverses all meshing coordinate points within the range. For each meshing coordinate point the gear passes through, the angle sensing component records the real-time angular orientation value of the output shaft. Each set of meshing coordinate points is matched with a unique corresponding output shaft angle data. The acquisition timing and spatial position of all angle data are kept at the same benchmark as the acquisition timing and spatial position of gear meshing vibration data, achieving synchronous acquisition of the two types of data.

[0026] In another embodiment, the angle sensing component has a built-in fixed data refresh frequency, with the overall refresh frequency set at 20,000 times / second. This adapts to the angular change rate under the rated operating speed of the gear, preventing data omissions and duplicate data collection. All collected output shaft angular values ​​are bound to the corresponding gear meshing coordinate point parameters and acquisition timing parameters. All raw angle data are batch-collected and stored according to the gear's angular rotation progress.

[0027] Based on the physical layout characteristics of the coaxial series transmission of the robot reducer, the original angular values ​​of all collected output shafts are subjected to sequential normalization. Following the spatial distribution order and rotational timing order of the gear meshing points, the discretely distributed angular values ​​are continuously arranged and integrated, eliminating gaps in the discrete point data. The normalized and integrated angular data forms a complete data sequence with continuous timing and one-to-one positional correspondence. This standardized data sequence is defined as the time-domain value of the output shaft rotation angle.

[0028] See Figure 4 In another embodiment, Figure 4This is a schematic diagram of the reducer's output shaft rotation structure, showing the coaxial series layout of the internal gear pair and the output shaft. A high-precision angle sensing component mounted on the output shaft end face rotates synchronously with the shaft. In this embodiment, the component's refresh rate is set to 20,000 times / second to adapt to the angular variation rate under the gear's rated speed. When the gear meshing contact surface generates continuous angular displacement, the component synchronously captures the output shaft angular orientation value corresponding to each meshing point. Each set of meshing points is matched with a unique corresponding angle data, and the data acquisition timing, spatial position, and gear meshing vibration data maintain a unified benchmark, reflecting the synchronous acquisition system of angle and vibration data, and intuitively presenting the structural basis of the output shaft's movement with gear meshing.

[0029] Preferably, step S2: using the time sequence of physical points of each meshing stroke as a reference, the time-domain values ​​of meshing vibration and the time-domain values ​​of rotation angle within the same continuous working condition acquisition interval are time-series aligned to generate a meshing time-series coupled dataset. Optionally, step S2 involves time-series alignment of the meshing vibration time-domain values ​​and the rotation angle time-domain values ​​within the same continuous operating condition acquisition interval, specifically as follows: Mark the data entry time node corresponding to each physical meshing coordinate point on the surface of the locking gear; The meshing vibration time-domain values ​​and rotation angle time-domain values ​​obtained within the same fixed input time node are uniformly classified into the same spatial coordinate group; Adjust the actual coordinate positions of the vibration and angle acquisition values ​​according to the entire angular range that can be covered by the complete circumference of the gear. Based on the spatial coordinates corresponding to the meshing of the gear bodies, the time-domain values ​​of meshing vibration and rotation angle are aligned in time sequence.

[0030] In one embodiment, time-series calibration is performed on the acquired data within the 22° to 25° angular acquisition range of single-tooth meshing. A unified spatiotemporal coordinate system is established with the gear rotation center as the reference, and a dedicated data entry time node is bound to each of the 44 to 50 sets of physical meshing coordinate points within the range. The time scale precision of all data entry time nodes is uniformly set to 0.0001s, and the time nodes are rigidly bound to the angular position parameters of each meshing point of the gear. Each physical meshing coordinate point corresponds to a unique fixed time scale.

[0031] The system iterates through all data entry time points within the continuous operating condition acquisition interval, and performs classification processing on the time-domain values ​​of gear meshing vibration and output shaft rotation angle acquired at the same fixed time scale. Two types of acquired values ​​generated at the same time point are assigned to the same spatial coordinate group, with the group code consistent with the corresponding meshing entity coordinate point code. This prevents data from different time sequences or points from overlapping or mixing. All data groups strictly adhere to the collection standard of single point, single time sequence, and dual data, completing the initial grouping and classification of the entire acquired data.

[0032] In another embodiment, using the 360° complete circumferential angular coverage of the gear as the adjustment benchmark, and combining the angular distribution position of the single-tooth meshing interval, coordinate alignment adjustments are performed on the vibration and angle acquisition values ​​within each group. All coordinate adjustments are based on the physical coordinates of the gear meshing entity, correcting the spatial offset deviations of the two types of data. This ensures that the spatial coordinate parameters of each group of dual-source data perfectly match the actual physical coordinate positions of the corresponding meshing points. Coordinate alignment calibration is then performed sequentially for all data groups across the entire domain.

[0033] Based on the completion of coordinate calibration and grouping of all data, and relying on the precise spatial coordinate arrangement rules corresponding to gear meshing, a unified temporal arrangement standard for dual-source data acquisition across the entire domain is established. All meshing vibration time-domain values ​​and rotation angle time-domain values ​​are aligned according to the circumferential distribution order of meshing points and the order of data entry time nodes, achieving four-dimensional unification of vibration data, angle data, meshing spatial coordinates, and acquisition timing parameters. This ultimately completes the standardized temporal alignment processing of all dual-source data within the continuous operating condition acquisition range.

[0034] Of particular importance, the generation of the meshing time-series coupling dataset in step S2 includes: Using the coordinate points of the meshing entities evenly distributed along the circumferential direction of the gear as the overall layout reference, the meshing vibration time domain value and the rotation angle time domain value that have been fixed in coordinate position are combined and matched one-to-one. Each set of matched meshing vibration time-domain values ​​and rotation angle time-domain values ​​corresponds to the contact coordinate position of an independent gear meshing entity inside the mechanical equipment. Following the entire operating range that the complete circumference of the gear can cover, the time-domain values ​​of meshing vibration and the time-domain values ​​of rotation angle are combined and grouped, and after fixing the group, a meshing time-series coupled dataset is formed.

[0035] In one embodiment, the uniformly distributed meshing entity coordinate points of the robot reducer gears are used as a unified layout reference. Within the 22° to 25° angular acquisition range for single-tooth meshing, there are 44 to 50 sets of standardized meshing coordinate points. All points have undergone prior coordinate calibration and timing alignment, and their spatial parameters remain fixed. The system retrieves all gear meshing vibration time-domain values ​​and output shaft rotation angle time-domain values ​​with fixed coordinate positions. Using a single meshing entity coordinate point as the smallest processing unit, fixed-point matching and binding are performed on the two types of time-domain values. All value matching operations strictly correspond to the actual spatial point parameters of the gear meshing.

[0036] The time-domain values ​​of gear meshing vibration and rotation angle, after being aligned according to the time sequence, are precisely matched one-to-one to establish a fixed pairing structure with dual data at a single coordinate point. Each pair of gear meshing vibration and rotation angle time-domain values ​​uniquely corresponds to the contact coordinate position of an independent gear meshing entity inside the reducer, and the time sequence parameters, coordinate parameters, and calibration parameters of each pair of data completely coincide with those of the corresponding meshing point.

[0037] In one embodiment, the entire operating range covered by the 360° complete circumference rotation of the gear is used as the processing boundary. All meshing entity coordinate points within the single-tooth meshing interval are traversed, and the time-domain values ​​of each pair are systematically combined and grouped one by one. The grouping operation strictly follows the physical arrangement sequence of the gear's circumferential meshing points. Following the angular travel sequence from the gear's engagement position to disengagement position, the sequential arrangement of each group of dual-source data is completed. All grouping units are stacked according to the spatial position sequence of the actual gear transmission, forming a data arrangement structure with continuous spatial and temporal attributes.

[0038] After the entire data is grouped, the combined layout structure and point correspondence of all paired data are locked, and the binding relationship between the time-domain values ​​of gear meshing vibration, the time-domain values ​​of rotation angle, and the coordinate points of the meshing entity in each group is solidified. The standardized dataset with the fixed layout structure is defined as the meshing time-series coupled dataset. All data units within the dataset maintain a state of independent points, continuous time sequence, and unified parameter correspondence. All data units are arranged in a regular manner according to the physical order of gear circumferential meshing. The parameter benchmarks within the dataset are completely consistent with the parameter benchmarks of the previous working condition acquisition interval, point calibration, and time sequence alignment.

[0039] Preferably, step S3: perform differential operation on the meshing time-coupling dataset, and calculate the real-time increment of vibration amplitude and the real-time offset increment of rotation angle corresponding to a single meshing point point; Of particular importance, the differential operation performed on the meshing timing coupling dataset in step S3 includes: According to the predetermined arrangement of the coordinate points of the meshing entities distributed along the circumferential direction of the gear, the values ​​that have been integrated and stored in the meshing timing coupling dataset are retrieved in sequence. Select the matching combination values ​​corresponding to two meshing points whose circumferential spatial coordinate positions of the gear are close to each other, and perform standard numerical difference calculation processing. For the numerical content distributed in different circumferential coordinate positions within the meshing timing coupling dataset, it is aligned with the actual distribution position of the meshing coordinates of the gear entity. The data processing range is divided into the outer meshing area of ​​the tooth surface, the middle meshing area of ​​the tooth surface, and the inner meshing area of ​​the tooth surface. The differential processing sequence extends from the outer meshing area of ​​the gear teeth to the inner meshing area. The grouping data corresponding to different meshing areas are divided into processing sequences, and the differential processing of the grouping data within different meshing areas is completed in sequence.

[0040] In one embodiment, the inherent arrangement order of the circumferential meshing coordinate points of the robot reducer gears is used as the retrieval reference. All data units within the meshing timing coupling dataset correspond to 44 to 50 sets of meshing coordinate points within the 22° to 25° single-tooth meshing range of the gear. All data units are standardized and integrated according to the physical rotation angle position of the gear. Following the fixed arrangement order of the gear circumferentially from the meshing start point to the meshing end point, the vibration time-domain values ​​and rotation angle time-domain values ​​combined units integrated within the dataset are extracted sequentially. All data retrieval actions are strictly matched with the circumferential coordinate orientation of the meshing points, and the data retrieval order is completely synchronized with the actual meshing transmission order of the gears.

[0041] The algorithm selects the combination of adjacent meshing points with the smallest circumferential coordinate spacing of the gears, maintaining a fixed 0.5° angular interval between adjacent points. It then extracts the dual-source combination values ​​corresponding to each pair of adjacent meshing points and performs standardized difference calculations. The calculation process uses the gear's rotational direction as a reference, employing a fixed logic of subtracting the values ​​of subsequent points from those of preceding points. For all data units distributed across different circumferential coordinate orientations within the dataset, a full-domain coverage calculation is performed based on the actual distribution of the gear's meshing coordinates, ensuring that all combination values ​​corresponding to all meshing points are included in the difference processing range.

[0042] In another embodiment, the gear meshing area is divided into fixed intervals using the radial dimension of the gear pitch circle as a rigid dividing criterion, and a unified radial offset parameter is set to complete the three-zone definition. The annular range with a radial distance of 0.3 mm from the addendum circle to the pitch circle is defined as the outer meshing area of ​​the tooth surface; the annular range with a radial offset of 0.3 mm above and below the pitch circle is defined as the middle meshing area of ​​the tooth surface; and the annular range with a radial distance of 0.3 mm from the pitch circle to the root circle is defined as the inner meshing area of ​​the tooth surface. All meshing coordinate points within a single tooth meshing interval are automatically assigned to their corresponding meshing areas based on their own radial coordinate parameters, and the coupling data corresponding to each area are grouped and independently categorized.

[0043] A fixed differential calculation path is set, and the calculation process strictly progresses from the outer meshing area to the inner meshing area. A fixed processing order is set for the grouped data corresponding to the three types of meshing areas. The grouped data of the outer meshing area is processed first, followed by the grouped data of the middle meshing area, and the grouped data of the inner meshing area is processed last. The combined values ​​of all adjacent points within each area are processed sequentially according to the circumferential arrangement. After the global partition differential calculation is completed, the complete differential result corresponding to each meshing point is retained, completing the global standardized differential processing of the meshing time-series coupled dataset.

[0044] Optionally, the real-time increment of vibration amplitude corresponding to a single meshing point calculated point by point in step S3 includes: The original vibration amplitude values ​​inside two meshing points that are close to each other in the circumferential coordinate position of the gear are extracted sequentially. Select the specific value corresponding to the rear coordinate point of the actual travel direction of the gear's overall circumferential rotation, subtract the specific value corresponding to the front coordinate point, and obtain the vibration amplitude result after calculation; The vibration amplitude result is defined as the real-time increment of vibration amplitude specifically used for the current meshing entity coordinate point.

[0045] In this embodiment, based on the fixed arrangement order of the coordinate points of the gear circumferential meshing entities, data units within the 22° to 25° angular interval of single-tooth meshing are sequentially retrieved from the meshing timing coupling dataset. Within this interval, 44 sets of meshing entity points are evenly distributed, with an angular interval of 0.5° between adjacent points. Following a predetermined processing order from the outer meshing area to the inner meshing area of ​​the gear, adjacent point combinations with closely spaced circumferential coordinates are selected, and the original acquired values ​​of gear meshing vibration amplitude stored within each set of points are extracted sequentially. The specific values ​​used in this paragraph are merely exemplary data and can be adaptively adjusted according to the actual operating conditions of the reducer, gear parameters, and running status. Taking two adjacent meshing points at 22.0° and 22.5° within the interval as examples, the original vibration amplitude acquisition data corresponding to the two points are extracted respectively. All extracted vibration amplitude data are bound to the corresponding angular coordinate parameters and meshing position information, and the extraction order perfectly matches the physical arrangement order of the gear circumferentially.

[0046] The relative positions of points are defined by the fixed unidirectional circumferential rotation direction of the robot reducer gears. Using the conventional counter-clockwise transmission direction of the gears as a reference, the front and rear point attributes are distinguished along the circumferential travel path of the gears. For the adjacent point combination of 22.0° and 22.5°, the 22.5° meshing point is classified as a rear coordinate point, and the 22.0° meshing point as a front coordinate point. The original vibration amplitude value of 0.82g corresponding to the rear 22.5° point and the original vibration amplitude value of 0.76g corresponding to the front 22.0° point are retrieved. Using fixed calculation rules, numerical calculations are performed. Substituting the data, the difference between 0.82g and 0.76g is obtained, resulting in a vibration amplitude difference of 0.06g for this group of points.

[0047] It should be noted that, following a unified computational logic, calculations were performed on each group of adjacent points across the entire domain. For all adjacent point groups, such as 22.5° and 23.0°, and 23.0° and 23.5°, the calculation method of subtracting the value of the preceding point from the value of the following point was repeatedly executed. The vibration amplitude difference obtained from the calculation of each group of points was matched to the coordinate points of each preceding meshing entity. The 22.0° meshing point was matched with a difference of 0.06g, and the remaining meshing points were matched with their corresponding calculated differences in turn. All meshing points underwent independent single calculations, with no data omissions or duplicate matching. The entire domain point calculation process maintained a unified standard.

[0048] All vibration amplitude differences obtained through differential calculations have been fixed as parameters. Each fixed set of difference values ​​is uniformly designated as a real-time vibration amplitude increment specific to the coordinate point of the corresponding meshing entity. The 0.06g value corresponding to the 22.0° meshing point and the calculated difference value corresponding to the 22.5° meshing point are both used as independent real-time vibration amplitude increment parameters for their respective points. All real-time vibration amplitude increments are sequentially collected and organized according to the arrangement of the gear's circumferential meshing points from 22° to 25°. The data parameter benchmarks are aligned with the previous time series, and the parameter system for dataset grouping and partitioned differential processing remains consistent, forming a complete set of single-point vibration amplitude real-time increment data.

[0049] Optionally, the real-time offset increment of the rotation angle corresponding to a single meshing point, calculated point by point in step S3, includes: Retrieve the original value of the rotation angle of the output shaft of the synchronous reducer from two sets of meshing entity points whose spatial coordinates are close to each other. The overall numerical calculation is completed by subtracting the original values ​​of the rotation angles corresponding to the two spatial coordinate points before and after the calculation. The resulting value is defined as the real-time offset increment of the rotation angle corresponding to the current meshing entity coordinate point.

[0050] In this embodiment, the rotation angle increment is calculated based on the 22° to 25° angular acquisition range of single-tooth meshing. 44 sets of meshing entity points are evenly distributed within this range, with a constant 0.5° circumferential angular interval between adjacent meshing points. Following a fixed processing order from the outer to the inner region of the gear circumferential meshing points, two sets of adjacent meshing entity points with matching spatial coordinates are selected sequentially. The original acquisition values ​​of the reducer output shaft rotation angle, centrally bound to the two sets of points in the meshing timing coupling data set, are retrieved. The specific angle values ​​listed in this embodiment are exemplary parameters and can be adapted and adjusted according to the actual transmission conditions of the reducer and the gear assembly parameters. All original angle values ​​are synchronously bound to the angular coordinates and acquisition timing information of the corresponding meshing points.

[0051] Based on the fixed unidirectional circular rotation direction of the gear, the front and rear positions of adjacent meshing points are defined. Along the gear's rotation path, the point behind the rotation is the rear coordinate point, and the point in front of the rotation is the front coordinate point. Taking two sets of adjacent meshing points at 23.0° and 23.5° as examples, the 23.0° point is the front coordinate point, and the 23.5° point is the rear coordinate point. The original value of the output shaft rotation angle corresponding to the 23.0° point is 35.126°, and the original value of the output shaft rotation angle corresponding to the 23.5° point is 35.158°. Both sets of angle values ​​are the original measured angle data collected at the moment of meshing at the corresponding meshing points.

[0052] It should be noted that a fixed calculation method of subtracting the angle value of the front coordinate point from the angle value of the rear coordinate point is uniformly adopted to perform the difference calculation for each group of adjacent meshing point combinations. For the 23.0° and 23.5° point combination, the measured values ​​are substituted to complete the calculation: 35.158° minus 35.126° yields an angle difference of 0.032°. Following the same calculation rule, batch calculations are performed on all adjacent point groups, such as 22.0° and 22.5°, 22.5° and 23.0°, and 23.5° and 24.0°. Each point combination is calculated independently, ensuring that the calculation standard remains consistent across the entire area, preventing duplicate calculations and data omissions.

[0053] The angle difference generated by the calculation of adjacent points in each group is fixedly matched with the corresponding coordinate points of the meshing entity on the front side, completing the parameter-specific binding process. The 23.0° meshing point is bound to an angle difference of 0.032°, and the remaining points are bound to the corresponding calculated angle differences in sequence. All bound angle differences are uniformly defined as the real-time offset increment of the rotation angle specific to the coordinate points of the corresponding meshing entity, to form a complete dataset of rotation angle offset increments for meshing points.

[0054] Optionally, after calculating the real-time increment of vibration amplitude and the real-time offset increment of rotation angle corresponding to a single meshing point in step S3, the following steps are also included: Based on the preset point order arrangement of the gear tooth meshing area of ​​the reducer, the real-time increment of vibration amplitude and the real-time offset increment of rotation angle obtained by differential calculation are matched to the corresponding order mark position of the gear tooth meshing point. Based on the sequential arrangement formed by the circumferential meshing of the gear teeth, the real-time increment of vibration amplitude and the real-time offset increment of rotation angle are aligned and arranged, so that the increment value generated at each meshing point corresponds to a unique sequential marker to complete the alignment arrangement.

[0055] In one embodiment, incremental data embedding is performed according to a preset point arrangement rule in the gear meshing area of ​​the reducer. 44 sets of equidistant meshing points are arranged within the 22° to 25° meshing range of a single tooth, with an angular interval of 0.5° between adjacent points. All meshing points are assigned a unique sequence identifier according to the circumferential meshing sequence of the gears, and these identifiers are sequentially encoded using a natural number sequence from 1 to 44. The encoding and point-to-point values ​​used in this embodiment are exemplary parameters and can be adapted and adjusted according to the actual gear specifications and meshing range of the reducer. All sequence identifiers are permanently bound to the spatial coordinates and acquisition timing parameters of the corresponding meshing points, forming a fixed point encoding and arrangement system.

[0056] The real-time increments of vibration amplitude and rotation angle offsets calculated from all meshing points are retrieved and matched one-to-one according to the unique sequence identifiers of the gear meshing points. Sequence code 1 corresponds to the 22.0° meshing point, sequence code 2 corresponds to the 22.5° meshing point, sequence code 3 corresponds to the 23.0° meshing point, and so on, to complete the coding matching of all points. Taking the 23.0° meshing point with sequence code 3 as an example, the real-time increment of vibration amplitude calculated for this point is 0.06g, and the real-time offset increment of rotation angle is 0.032°. These two sets of increment values ​​are individually matched to the corresponding sequence identifier positions for this point.

[0057] In another embodiment, the overall data arrangement logic is determined based on the physical arrangement of the gear teeth's circumferential meshing. The data arrangement direction conforms to the gear's meshing path from the outer meshing area to the inner meshing area, strictly following the ascending order of sequence marker 1 to sequence marker 44 to complete the incremental data sequence arrangement. The real-time incremental vibration amplitude and rotation angle offset of all points do not involve cross-point swapping, misalignment, or order reversal. The arrangement position of each set of incremental data perfectly matches the circumferential physical distribution position of the corresponding meshing point.

[0058] After completing the data normalization of the entire area, the binding relationship between all meshing points, sequence identifiers, and dual-dimensional incremental values ​​is locked, realizing the fixed alignment of the incremental value of each meshing point with its unique sequence identifier. All meshing points corresponding to sequence codes 1 to 44 are respectively equipped with their own calculated real-time increments of vibration amplitude and rotation angle offset, forming a corresponding incremental arrangement data sequence.

[0059] Preferably, step S4: based on the continuous meshing and traveling sequence of gears, using the real-time increment of vibration amplitude and the real-time offset increment of rotation angle as prediction references, the tooth surface meshing transmission error and shaft rotation transmission error are predicted step by step, and the real-time transmission accuracy value of the robot reducer is obtained by iterative calculation.

[0060] Optionally, the step-by-step prediction of tooth surface meshing transmission error and shaft rotation transmission error in step S4 is as follows: Identify and mark the meshing positions of the outer, middle, and inner rings of the gears in the robot reducer; The vibration amplitude increments of the meshing points in each zone are retrieved in real time and matched with the squeezing contact state of different areas of the gear tooth surface to complete the point-by-point calculation of the meshing transmission error of the tooth surface. The real-time offset increment of the rotation angle corresponding to the meshing point of the same partition is retrieved and matched with the angular offset characteristics of the fixed shaft transmission structure of the reducer to complete the point-by-point calculation of the rotation transmission error of the shaft system.

[0061] In one embodiment, the meshing area is divided into zones based on the radial spatial coordinates of the gear tooth surface. The gear pitch circle is used as a fixed reference to divide the meshing positions into three zones: the outer ring meshing area (0.3 mm radially from the addendum circle to the pitch circle), the middle meshing area (0.3 mm radially above and below the pitch circle), and the inner ring meshing area (0.3 mm radially from the pitch circle to the root circle). Forty-four meshing points are arranged within the 22° to 25° meshing interval of a single tooth, with an angular interval of 0.5° between adjacent points. All meshing points are assigned to their corresponding zones based on their radial coordinate parameters, completing the zoning and marking of all meshing points. The radial and angular interval parameters used in this embodiment are example parameters and can be adjusted based on the actual gear operating parameters of the reducer. All zone boundary parameters are permanently stored.

[0062] In another embodiment, the real-time increment of vibration amplitude at each meshing point in the partition is retrieved, and the point-by-point calculation of the tooth surface meshing transmission error is completed according to the fixed conversion coefficient of each partition. Each meshing area has a preset exclusive fixed conversion coefficient. The conversion coefficient for the outer ring meshing area of ​​the gear is set to 1.02, the conversion coefficient for the middle meshing area of ​​the gear is set to 1.00, and the conversion coefficient for the inner ring meshing area of ​​the gear is set to 0.98. The conversion coefficient corresponds one-to-one with the tooth surface extrusion contact state. Taking the 22.5° meshing point on the outer ring of the gear, the 23.5° meshing point in the middle, and the 24.5° meshing point on the inner ring as examples, the real-time increments of the vibration amplitude at the three points are 0.06g, 0.04g, and 0.05g, respectively. The single-point tooth surface meshing transmission error is calculated by multiplying the real-time increment of the single-point vibration amplitude by the corresponding partition conversion coefficient, resulting in tooth surface meshing transmission error values ​​of 0.0612g, 0.0400g, and 0.0490g, respectively. All meshing points are solved independently one by one according to this calculation logic.

[0063] It should be noted that by retrieving the real-time offset increment of the rotation angle corresponding to the same meshing point, and combining it with the angular offset correction parameter of the fixed-axis transmission structure, the point-by-point calculation of the shaft system rotation transmission error is completed. The angular correction parameter for the outer ring meshing area of ​​the gear is set to 1.01, the angular correction parameter for the middle meshing area of ​​the gear is set to 1.00, and the angular correction parameter for the inner ring meshing area of ​​the gear is set to 0.99. Corresponding to the aforementioned three meshing points, the real-time offset increments of the rotation angle are 0.032°, 0.025°, and 0.028°, respectively. The single-point shaft system rotation transmission error is calculated by multiplying the single-point rotation angle real-time offset increment by the corresponding angular correction parameter, yielding shaft system rotation transmission error values ​​of 0.0323°, 0.0250°, and 0.0277°, respectively.

[0064] Optionally, in step S4, the real-time transmission accuracy value of the robot reducer is obtained through iterative calculation as follows: The tooth surface meshing transmission error and shaft rotation transmission error obtained by solving different meshing partition points along the continuous meshing path of the gear circumference are iteratively accumulated point by point. Based on the meshing process sequence from the outer ring to the inner ring of the gear, the positional error of the entire domain is summarized layer by layer. The transmission accuracy is predicted based on the error values ​​after layer-by-layer summarization, and the real-time transmission accuracy value corresponding to the complete meshing cycle of a single tooth of the robot reducer is output.

[0065] In one embodiment, the iterative calculation benchmark is the circumferential continuous meshing path of the gears, which strictly follows the fixed order of the outer ring meshing area, the middle meshing area, and the inner ring meshing area. Forty-four meshing points within the 22° to 25° meshing range of a single tooth are evenly arranged at 0.5° angular intervals. The tooth surface meshing transmission error and shaft rotation transmission error have been solved point-by-point at each point, and all error values ​​are retained to four decimal places. The error values ​​and accumulation process used in this embodiment are exemplary data and can be adjusted according to the actual assembly parameters and transmission conditions of the reducer. All iterative calculations follow unified accumulation rules and accuracy standards.

[0066] The iterative accumulation operation uses a single meshing point as the smallest computational unit, performing point-by-point superposition processing on the two types of error data along the circumferential meshing direction of the gear. Taking the points in the outer ring region of the gear as an example, the tooth surface meshing transmission error at the 22.0° meshing point is 0.0612g, and the shaft rotation transmission error is 0.0323°; the tooth surface meshing transmission error at the 22.5° meshing point is 0.0585g, and the shaft rotation transmission error is 0.0318°. The error value of the first point is used as the initial accumulation base, and the error values ​​of subsequent adjacent points are sequentially superimposed on the accumulation base. After the error iterative accumulation of all outer ring meshing points is completed sequentially, the total error value of the outer ring region of the gear is formed. The whole-domain single-point iteration process maintains the operation form of point-by-point progressive and continuous accumulation, without skipping operations or repeated accumulation operations.

[0067] In another embodiment, after completing the iterative accumulation of single-zone points, a layer-by-layer summary of the overall error is performed according to the meshing processing order from the outer ring to the middle and from the middle to the inner ring of the gear. The middle meshing area and the inner ring meshing area of ​​the gear are calculated separately using the same point-by-point iterative accumulation method to determine the total tooth surface meshing transmission error and the total shaft system rotational transmission error for each zone. The accumulated error values ​​of the outer ring area and the middle area are superimposed to generate the summed error value for the outer and middle areas. This value is then superimposed and integrated with the accumulated error value of the inner ring area to finally obtain the overall total tooth surface meshing transmission error and the overall total shaft system rotational transmission error for the complete meshing range of a single tooth.

[0068] After the total error across the entire domain is aggregated layer by layer, the transmission accuracy is determined and output using the two types of total total error values ​​as the core benchmark. The aggregated error values ​​are uniformly retained to four decimal places, conforming to the accuracy standards of all previous data calculations. The data benchmark and the parameter system for partition calibration, point-by-point error calculation, and incremental calculation are kept consistent. The final aggregated total error value corresponds to the transmission state of a single tooth of the robot reducer during a complete meshing cycle, outputting the real-time transmission accuracy value of the reducer corresponding to the single tooth meshing cycle, thus completing the quantitative prediction process for a single complete reducer transmission accuracy.

[0069] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0070] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion, characterized in that, Includes the following steps: Step S1: Determine the continuous working condition acquisition interval based on the preset single-tooth complete meshing stroke of the robot reducer; The time-domain values ​​of gear meshing vibration and the time-domain values ​​of output shaft rotation angle are collected during the continuous operating condition acquisition interval. Step S2: Based on the time sequence of physical points of each meshing stroke, time-domain values ​​of meshing vibration and time-domain values ​​of rotation angle within the same continuous working condition acquisition interval are time-series aligned to generate a meshing time-series coupled dataset. Step S3: Perform differential operation on the meshing time-coupling dataset to calculate the real-time increment of vibration amplitude and the real-time offset increment of rotation angle corresponding to a single meshing point. Step S4: Based on the continuous meshing and traveling sequence of gears, and using the real-time increment of vibration amplitude and the real-time offset increment of rotation angle as prediction references, predict the tooth surface meshing transmission error and shaft rotation transmission error step by step, and obtain the real-time transmission accuracy value of the robot reducer through iterative calculation.

2. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, In step S1, the continuous operating condition data acquisition interval is determined based on the preset single-tooth complete meshing stroke of the robot reducer as follows: The spatial coordinates of the working surfaces of the tooth tip and tooth root in the paired meshing gear pair inside the reducer of the positioning robot; Mark the critical coordinate positions where the driving gear teeth smoothly enter the tooth groove of the driven gear in sequence, and mark the critical coordinate positions where the teeth completely disengage from the tooth groove contact area. Calculate the complete angular coverage area occupied by the gear when it completes a circular rotation between two critical coordinate points, and add positioning coordinate markers at the beginning and end of this angular range. Based on the complete angular coverage formed after adding positioning coordinate markers, a dedicated data acquisition range is defined to adapt to the complete meshing rotation process of a single tooth.

3. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, The specific steps for collecting the time-domain values ​​of gear meshing vibration in step S1 are as follows: On the complete trajectory of the robot reducer gear as it completes its full circular rotation, multiple sets of fixed coordinate acquisition points with consistent spacing are divided. Within the defined angular acquisition range, complete the acquisition of vibration data corresponding to each fixed coordinate acquisition point; Based on the continuous changes in coordinate position generated during the continuous rotation of the gear, the vibration data is arranged and combined in an orderly manner to form the time-domain numerical value of gear meshing vibration.

4. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, The specific steps for acquiring the time-domain value of the output shaft rotation angle in step S1 are as follows: Throughout the entire process of continuous angular position change of the meshing contact surface of the gears, various angular orientation related data generated by the shaft structure at the power output end of the reducer are simultaneously recorded. Within the uniformly defined angular acquisition range, at any meshing coordinate position along the gear's rotation path, the angular value generated by the corresponding position of the output shaft is synchronously matched and recorded. Based on the physical arrangement of the coaxial series transmission inside the reducer, all the recorded angular values ​​are integrated and arranged to form the time domain value of the output shaft rotation angle.

5. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, Step S2 involves aligning the time-domain values ​​of meshing vibration and rotation angle within the same continuous operating condition acquisition interval in terms of timing. Mark the data entry time node corresponding to each physical meshing coordinate point on the surface of the locking gear; The meshing vibration time-domain values ​​and rotation angle time-domain values ​​obtained within the same fixed input time node are uniformly classified into the same spatial coordinate group; Adjust the actual coordinate positions of the vibration and angle acquisition values ​​according to the entire angular range that can be covered by the complete circumference of the gear. Based on the spatial coordinates corresponding to the meshing of the gear bodies, the time-domain values ​​of meshing vibration and rotation angle are aligned in time sequence.

6. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, The real-time increment of vibration amplitude corresponding to a single meshing point calculated point by point in step S3 includes: The original vibration amplitude values ​​inside two meshing points that are close to each other in the circumferential coordinate position of the gear are extracted sequentially. Select the specific value corresponding to the rear coordinate point of the actual travel direction of the gear's overall circumferential rotation, subtract the specific value corresponding to the front coordinate point, and obtain the vibration amplitude result after calculation; The vibration amplitude result is defined as the real-time increment of vibration amplitude specifically used for the current meshing entity coordinate point.

7. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, The real-time offset increment of the rotation angle corresponding to a single meshing point, calculated point by point in step S3, includes: Retrieve the original value of the rotation angle of the output shaft of the synchronous reducer from two sets of meshing entity points whose spatial coordinates are close to each other. The overall numerical calculation is completed by subtracting the original values ​​of the rotation angles corresponding to the two spatial coordinate points before and after the calculation. The resulting value is defined as the real-time offset increment of the rotation angle corresponding to the current meshing entity coordinate point.

8. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, After step S3 calculates the real-time increment of vibration amplitude and the real-time offset increment of rotation angle corresponding to a single meshing point, it also includes: Based on the preset point order arrangement of the gear tooth meshing area of ​​the reducer, the real-time increment of vibration amplitude and the real-time offset increment of rotation angle obtained by differential calculation are matched to the corresponding order mark position of the gear tooth meshing point. Based on the sequential arrangement formed by the circumferential meshing of the gear teeth, the real-time increment of vibration amplitude and the real-time offset increment of rotation angle are aligned and arranged, so that the increment value generated at each meshing point corresponds to a unique sequential marker to complete the alignment arrangement.

9. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 1, characterized in that, The step-by-step prediction of tooth surface meshing transmission error and shaft rotation transmission error in step S4 is as follows: Identify and mark the meshing positions of the outer, middle, and inner rings of the gears in the robot reducer; The vibration amplitude increments of the meshing points in each zone are retrieved in real time and matched with the squeezing contact state of different areas of the gear tooth surface to complete the point-by-point calculation of the meshing transmission error of the tooth surface. The real-time offset increment of the rotation angle corresponding to the meshing point of the same partition is retrieved and matched with the angular offset characteristics of the fixed shaft transmission structure of the reducer to complete the point-by-point calculation of the rotation transmission error of the shaft system.

10. The method for predicting the transmission accuracy of a robot reducer based on multi-source information fusion according to claim 9, characterized in that, In step S4, the real-time transmission accuracy value of the robot reducer is obtained through iterative calculation as follows: The tooth surface meshing transmission error and shaft rotation transmission error obtained by solving different meshing partition points along the continuous meshing path of the gear are iteratively accumulated point by point. Based on the meshing process sequence from the outer ring to the inner ring of the gear, the positional error of the entire domain is summarized layer by layer. Based on the error values ​​after layer-by-layer summarization, the transmission accuracy is predicted, and the real-time transmission accuracy value corresponding to the complete meshing cycle of a single tooth of the robot reducer is output.

Images