Injection mold man-machine interaction method and system based on real-time feedback

Abstract

The invention discloses an injection mold man-machine interaction method and system based on real-time feedback, and particularly relates to the field of man-machine interaction.The method comprises the steps that space operation data of an operator are collected through a posture integrated recognition sensor worn by the operator, and the space operation data comprise the angular speed, the angular acceleration, the displacement speed and the posture angle; forming an action input vector set based on the spatial operation data; and the action input vector set is subjected to inertial fusion and Kalman filtering processing to form three-dimensional attitude trajectory data, and the three-dimensional attitude trajectory data is calibrated as a feature operation trajectory template. Space operation data of an operator is converted into a posture input vector in real time, a tactile feedback signal is generated in combination with a physical response state of multiple components of the mold, and an interaction path capable of realizing closed-loop iteration is constructed; the core problem that in the background technology, an operator cannot intuitively perceive the real motion state of the multi-axis mold at the moment of motion is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of human-computer interaction, and more particularly to a human-computer interaction method and system for an injection mold based on real-time feedback. Background Art

[0002] During the commissioning of large multi-axis injection molds, operators need to precisely control the linkage displacement of multiple structural components (such as templates, slides, and tie rods) to ensure accurate mold closure and synchronous movement.

[0003] However, current control methods rely primarily on buttons or remote controls to trigger actions, and on screen interfaces or buzzer prompts to determine execution results. This results in high mold structural inertia and high response delays, making it prone to severe operational feedback lags. This also creates a disconnect between the refresh rate of the visual interface and the actual mold status, preventing operators from instantly understanding whether the mold has completed its intended action at critical moments.

[0004] Especially when debugging multiple axes simultaneously, it is even more difficult to capture the deviation between the physical state of each axis and the real-time feedback. This "perception lag" is particularly fatal during high-speed mold trials or fine-tuning. The core problem that ultimately arises is: the operator cannot intuitively perceive the true motion state of each axis of the mold at the moment of action. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a human-computer interaction method and system for an injection mold based on real-time feedback. By converting the operator's spatial operation data into a posture input vector in real time, and combining it with the physical response state of multiple components of the mold to generate a tactile feedback signal, a closed-loop iterative interaction path is constructed to solve the core problem raised in the background technology that "the operator cannot intuitively perceive the true motion state of the multi-axis mold at the moment of action."

[0006] To achieve the above-mentioned object, the present invention provides the following technical solutions: a human-computer interaction method for an injection mold based on real-time feedback, comprising a posture integrated recognition sensor;

[0007] The operator's spatial operation data is collected through the gesture integrated recognition sensor worn by the operator. The spatial operation data includes angular velocity, angular acceleration, displacement velocity and gesture angle, and an action input vector set is formed based on the spatial operation data.

[0008] The action input vector set is processed by inertial fusion and Kalman filtering to form three-dimensional posture trajectory data, and is calibrated as a characteristic operation trajectory template;

[0009] Dynamically match the three-dimensional posture trajectory data with the operation behavior control parameter mapping table. The operation behavior parameters of the operation behavior control parameter mapping table include the template drive path, the slider opening and closing positions, and the structural control instructions of the pull rod synchronization process, and output the corresponding structural response execution command sequence;

[0010] Continuous data acquisition is performed on three physical parameters of each structure in the injection mold: spatial displacement, motion rate, and clamping force. The acquired results are combined by component type and aligned based on timestamps to obtain a structural state feedback matrix.

[0011] The structural state feedback matrix is ​​converted into a tactile feedback signal model through a tactile mapping function. The posture integrated recognition sensor worn by the operator is driven by a multi-channel drive unit to output a multi-dimensional tactile signal including vibration direction, frequency amplitude, and surface temperature.

[0012] The operator completes dynamic regulation of posture control based on multi-dimensional tactile signals, synchronously reconstructs the action input vector and performs continuous iterative feedback to form human-computer interaction.

[0013] In a preferred embodiment, normalization processing and semantic integration operations are performed on the raw sensory data contained in the structural state feedback matrix to generate a structural state fusion vector block, wherein the raw sensory data includes spatial displacement data, posture angle data, clamping force data, thermal state data, and cooling flow rate data;

[0014] Dynamically jointly modeling the structural state fusion vector block with the 3D posture trajectory data, which is obtained by posture solution of the action input vector set, is used to establish the corresponding characteristic relationship between the operator's action intention and the mold response state;

[0015] A structural state perception map is constructed based on the corresponding feature relationship. The structural state perception map presents the linkage path between the response state of each substructure of the injection mold and the operator's posture input in the form of multi-dimensional spatial logical aggregation.

[0016] In a preferred embodiment, the linkage relationship between the response state of each substructure of the injection mold presented in the structural state perception map and the operator's posture input is calculated according to the spatial position, time rhythm and behavioral intention. Figure 3 A response pattern set is constructed in the class dimension and converted into prompt elements. The prompt elements include structural position guidance, inertial offset trend and error risk block, and are synchronously fed back to the operator's perception path through a graphical interface path and a tactile signal channel to assist the operator in real-time posture adjustment and path judgment.

[0017] In a preferred embodiment, a temporal fusion is performed on the multiple rounds of gesture input trajectories and mold response feedback sequences formed during the real-time interaction process to construct an operation behavior causal chain based on the feedback history. The operation behavior causal chain includes the action start state, transition state, and completion state corresponding to different feedback results. The trigger relationship between each state is calibrated to form a path trigger condition set.

[0018] The causal chain of the operation behavior is converted into a response state graph structure according to the path mapping relationship between the posture input trajectory and the injection mold response feedback sequence. The response state graph structure contains three types of state transition forms: jump nodes, fallback paths and nested response branches.

[0019] In a preferred embodiment, the response amplitude value and delay data in the structural state feedback matrix are introduced into the response state diagram structure as the quantitative weight of the operation behavior constraint condition to construct the operation behavior restriction map;

[0020] The behavior path restriction conditions, false touch rejection paths and state fault tolerance intervals generated in the operation behavior restriction map are applied to the control interface where the operator inputs instructions, realizing a path-based control strategy based on feedback status, a real-time correction mechanism and structural cascade fault tolerance.

[0021] In a preferred embodiment, a difference analysis is performed on the continuous input adjustment behavior formed by the operator during the posture control process, and fine-tuning behavior events with feedback result change characteristics are extracted. The difference between the posture input change and the structural response corresponding to each event is calibrated;

[0022] The extracted fine-tuning behavior events are clustered according to the feedback difference characteristics and operation intention trends, and a set of behavior correction response functions that reflect the control correction rules are constructed;

[0023] The behavior correction response function set is input into the control rule adjustment path for dynamically adjusting the structure execution parameters, and the angle limit, action execution rate and response scheduling strategy are parameterized according to the function content;

[0024] The operator's historical input behavior sequence and corresponding feedback change data are integrated into the control rule adjustment path to construct an empirical operation preference model. The parameter adjustment boundary is dynamically optimized in combination with the behavior deviation trend to form a control interval integration mechanism with real-time feedback.

[0025] In a preferred embodiment, the extreme value variation intervals of three parameters related to spatial displacement, thermal state and clamping force in the structural state feedback matrix are extracted to construct a corresponding risk trigger critical parameter set, covering three risk conditions: spatial position limit, temperature upper limit and clamping force mutation threshold;

[0026] Perform time series analysis on the continuously collected structural state feedback matrix to detect in real time any sudden increase in any dimension parameter that exceeds the trigger critical parameter set, and mark it as a potential risk trigger candidate event;

[0027] The feature vectors of potential risk trigger candidate events are compared with the sample set of historical high-risk events, the risk level is determined based on the similarity weight and time evolution trend, and the corresponding suppression path plan is matched according to the level result.

[0028] In a preferred embodiment, a multi-channel pressing mechanism is synchronously activated according to the pressing path plan. The multi-channel pressing mechanism includes a control instruction channel blocking mechanism, a tactile prompt mechanism based on direction reinforcement, and an interface information flash blocking mechanism to achieve operation interruption feedback linkage when the mold is in a dangerous state;

[0029] All confirmed risk behavior samples and their corresponding trigger conditions, response paths, and suppression result data are sent to the risk behavior record set for subsequent strategy evolution and model retraining.

[0030] A human-computer interaction system for injection molds based on real-time feedback, including a data acquisition module, a posture reconstruction module, an instruction mapping module, a state perception module, a tactile generation module, and a feedback closed-loop module;

[0031] The data acquisition module is used to collect the operator's spatial operation data through the gesture integrated recognition sensor worn by the operator. The spatial operation data includes angular velocity, angular acceleration, displacement velocity and gesture angle, and forms an action input vector set based on the spatial operation data;

[0032] The posture reconstruction module is used to process the action input vector set through inertial fusion and Kalman filtering to form three-dimensional posture trajectory data, and calibrate it into a feature operation trajectory template;

[0033] The instruction mapping module is used to dynamically match the 3D posture trajectory data with the operation behavior control parameter mapping table. The operation behavior parameters of the operation behavior control parameter mapping table include the template drive path, the slider opening and closing positions, and the structural control instructions of the pull rod synchronization process, and output the corresponding structural response execution command sequence;

[0034] The state perception module is used to continuously collect data on three physical parameters of each structure in the injection mold: spatial displacement, motion rate, and clamping force. The collected results are combined by component type and aligned based on timestamps to obtain a structural state feedback matrix.

[0035] The tactile generation module is used to convert the structural state feedback matrix into a tactile feedback signal model through a tactile mapping function. The multi-channel drive unit drives the posture integrated recognition sensor worn by the operator to output a multi-dimensional tactile signal including vibration direction, frequency amplitude, and surface temperature.

[0036] The feedback closed-loop module is used to complete dynamic regulation of posture control based on multi-dimensional tactile signals by the operator, synchronously reconstruct the action input vector and perform continuous iterative feedback to form human-computer interaction.

[0037] Technical effects and advantages of the present invention:

[0038] 1. By converting the operator's spatial operation data into posture trajectories in real time and combining it with the mold's physical response to generate tactile feedback signals, the operator can sense the actual motion status of each axis of the mold structure at the moment of action, effectively solving the problems of "response lag and unknown status" in traditional mold debugging;

[0039] 2. By constructing a structural state fusion vector block that fuses the mold substructure state and the operator trajectory projection, joint modeling between multiple physical quantity states (position, clamping force, temperature, etc.) and posture input is achieved, improving the human-machine perception matching ability of the mold response state in space;

[0040] 3. By identifying the behavioral causal relationship in multiple rounds of operation feedback and constructing a response state graph structure, the system introduces the structural response amplitude and delay as constraint weights to achieve predictive control of the posture input path and automatic avoidance of illegal paths, ensuring stability and behavioral rationality during complex debugging processes.

[0041] 4. Events in the operator's fine-tuning behavior that differ significantly from the structural response are modeled as feedback correction functions. The mold control rule parameter range is dynamically adjusted based on historical behavior trends, enabling the system to self-repair and reconstruct habitual strategies, thereby improving operational flexibility and intelligence.

[0042] 5. Determine the risk level based on high-order disturbances and timing surge behaviors in the feedback matrix, build a matching pressing path, and jointly execute a synchronous blocking mechanism for tactile, interface, and control channels to achieve precise pressing and immediate operator avoidance in the critical state of the mold. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the main process of interactive input and structural feedback of the present invention.

[0044] Figure 2 This is a flow chart of the posture control and reinjection adjustment of the present invention.

[0045] Figure 3 This is a risk identification and suppression linkage flow chart of the present invention.

[0046] Figure 4 Schematic diagram of the system module of the present invention.

[0047] Figure 5 The figure is a flow chart of the method steps of the present invention. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0049] Refer to the instruction manual Figure 1-5 , an embodiment of the present invention provides a human-computer interaction method for an injection mold based on real-time feedback, comprising a gesture integrated recognition sensor;

[0050] The operator's spatial operation data is collected through the gesture integrated recognition sensor worn by the operator. The spatial operation data includes angular velocity, angular acceleration, displacement velocity and gesture angle, and an action input vector set is formed based on the spatial operation data.

[0051] The action input vector set is processed by inertial fusion and Kalman filtering to form three-dimensional posture trajectory data, and is calibrated as a characteristic operation trajectory template;

[0052] Dynamically match the three-dimensional posture trajectory data with the operation behavior control parameter mapping table. The operation behavior parameters of the operation behavior control parameter mapping table include the template drive path, the slider opening and closing positions, and the structural control instructions of the pull rod synchronization process, and output the corresponding structural response execution command sequence;

[0053] Continuous data acquisition is performed on three physical parameters of each structure in the injection mold: spatial displacement, motion rate, and clamping force. The acquired results are combined by component type and aligned based on timestamps to obtain a structural state feedback matrix.

[0054] The structural state feedback matrix is ​​converted into a tactile feedback signal model through a tactile mapping function. The posture integrated recognition sensor worn by the operator is driven by a multi-channel drive unit to output a multi-dimensional tactile signal including vibration direction, frequency amplitude, and surface temperature.

[0055] The operator completes dynamic control of posture control based on multi-dimensional tactile signals, synchronously reconstructs the action input vector and performs continuous iterative feedback to form human-computer interaction;

[0056] It should be further explained that the operator's spatial operation data, including angular velocity, angular acceleration, displacement velocity and posture angle, is collected through the posture integrated recognition sensor to construct the action input vector set. After trajectory reconstruction and control parameter mapping, the control execution command sequence corresponding to the injection mold structure is generated; It represents the sequence of structural control execution commands derived from the operator's spatial motion, which is used to drive the coordinated motion of the template, slider, pull rod and other components in the injection mold; Formula 1 is:

[0057]

[0058] where θ i (t) is the attitude angle change function in the direction of the i-th axis; ω i (t),α i (t) are the angular velocity and angular acceleration functions of the i-th axis respectively; ψ i (t) is the instantaneous curvature function of the attitude angle, which is used to represent the nonlinear perturbation of the operation path in the unit attitude space; μ i Define parameters for the motion stability domain, which are used to calibrate the tolerance domain of the perturbation operation; φ i (·) is the trajectory driving function, which is used to construct the velocity-angle coupling behavior corresponding to the input trajectory; ζ i (·) is the input suppression function, which controls the response weight based on the intensity of the nonlinear perturbation; δ i is the structural behavior correlation coefficient, which is used to indicate the degree of dependence of the input on the action of the template / slider / pull rod;

[0059] The first second-order derivative in Equation 1 represents the instantaneous angular acceleration, which is used to respond to rapid behavioral changes; the second integral represents the cumulative driving force of the historical input action; in practical applications, the trajectory driving function is used to represent the dynamic excitation term composed of the product of velocity and acceleration; i Change the curvature ψ i (t) combined with the perturbation tolerance, reflects the strategy of suppressing input fluctuations; δ i Assign behavioral input to the corresponding mold structure and establish a "directional association mapping" between input and structural behavior.

[0060] performing normalization and semantic integration operations on raw sensory data contained in a structural state feedback matrix to generate a structural state fusion vector block, wherein the raw sensory data includes spatial displacement data, posture angle data, clamping force data, thermal state data, and cooling flow rate data;

[0061] Dynamically jointly modeling the structural state fusion vector block with the 3D posture trajectory data, which is obtained by posture solution of the action input vector set, is used to establish the corresponding characteristic relationship between the operator's action intention and the mold response state;

[0062] A structural state perception map is constructed based on the corresponding feature relationship. The structural state perception map presents the linkage path between the response state of each substructure of the injection mold and the operator's posture input in the form of multi-dimensional spatial logical aggregation.

[0063] The linkage relationship between the response state of each substructure of the injection mold presented in the structural state perception map and the operator's posture input is analyzed according to the spatial position, time rhythm and behavioral intention. Figure 3 A response pattern set is constructed in the class dimension and converted into prompt elements. The prompt elements include structural position guidance, inertial offset trend and error risk block, and are synchronously fed back to the operator's perception path through a graphical interface path and a tactile signal channel to assist the operator in real-time posture adjustment and path judgment.

[0064] It should be further explained that in the chain of structural state fusion and spatial cognitive modeling, the physical states of each substructure of the mold are integrated and combined with the operator's posture trajectory to generate a spatial structure perception map; It represents the fusion tensor formed by feature alignment and cross-modeling of the multi-dimensional state data of each substructure of the mold and the operator's posture input information in a unified time and space. It is used to characterize the response linkage relationship between humans and machines and is expressed by Equation 2 as follows:

[0065]

[0066] in is the state tensor of the jth structural component, which contains information such as spatial displacement, clamping force, and thermal diffusion rate; is the operator posture input projection tensor corresponding to the j-th structure; j is the response coupling strength factor of the local structure; ρ j is the structural thermal state disturbance coefficient; τ j is the structural response hysteresis time constant; η j is the nonlinear tension density distribution function of the unit posture feedback to the structure;

[0067] In Equation 2, the first term is the main term of the thermal response, which uses the local disturbance and coupling strength to make a nonlinear ratio of the hysteresis tension gradient; the second term is the spatial cross-coupling tensor, which represents the linkage strength by the cross product of the state change direction and the input direction; the overall structural fusion model generates the cognitive map basic tensor with the coupling term of "structural state change + posture input dynamics".

[0068] The operator performs temporal fusion on multiple rounds of gesture input trajectories and mold response feedback sequences formed during real-time interaction, constructing a causal chain of operation behaviors based on feedback history. The causal chain includes the action start state, transition state, and completion state corresponding to different feedback results, and performs conditional calibration on the trigger relationship between each state to form a path trigger condition set.

[0069] The causal chain of the operation behavior is converted into a response state graph structure according to the path mapping relationship between the posture input trajectory and the injection mold response feedback sequence. The response state graph structure contains three types of state transition forms: jump nodes, fallback paths and nested response branches.

[0070] The response amplitude value and delay data in the structural state feedback matrix are introduced into the response state diagram structure as the quantitative weight of the operation behavior constraint conditions to construct the operation behavior restriction map;

[0071] Apply the behavior path restriction conditions, false touch rejection paths, and state fault tolerance intervals generated in the operation behavior restriction map to the control interface where the operator inputs commands, thus realizing a path-based control strategy based on feedback status, a real-time error correction mechanism, and structural cascade fault tolerance.

[0072] It should be further explained that in the path mapping and state constraint generation chain, the state transition rules in the feedback behavior sequence are extracted to construct a graph structure, and the feedback amplitude and delay are added to form a path constraint model; Equation 3 is:

[0073]

[0074] where Δθ k is the attitude jump amplitude of the node in the kth row; Δt k The shortest reaction interval corresponding to it; σ k is the behavioral fuzziness of the transition state between nodes; γ k is the feedback strength index generated by the posture change between nodes; ∈ k is the behavior tolerance limit parameter; δs k / δτ k represents the local perturbation density of the feedback response on the state trajectory over time; The set of path state mapping graph structures finally constructed; It represents the set union of the path substructures constructed for all interaction behavior nodes k, that is, the path contribution item corresponding to each node k is regarded as an independent path element, and finally a complete response state path graph is formed through the union operation;

[0075] In Equation 3, each graph node constructs a preliminary state connection through the attitude jump amplitude and response interval time; the dynamic instability of the feedback trajectory is reflected by adding the perturbation density integral between nodes; ln(γ k / ∈ k ) is used as an inhibitory adjustment factor for behavior jumps, controlling the path passing probability of behaviors with too fast state jumps or exceeding limits in the graph structure.

[0076] Perform difference analysis on the operator's continuous input adjustment behavior during the posture control process, extract fine-tuning behavior events with feedback result change characteristics, and calibrate the difference between the posture input change and the structural response corresponding to each event;

[0077] The extracted fine-tuning behavior events are clustered according to the feedback difference characteristics and operation intention trends, and a set of behavior correction response functions that reflect the control correction rules are constructed;

[0078] The behavior correction response function set is input into the control rule adjustment path for dynamically adjusting the structure execution parameters, and the angle limit, action execution rate and response scheduling strategy are parameterized according to the function content;

[0079] The operator's historical input behavior sequence and corresponding feedback change data are integrated into the control rule adjustment path to build an empirical operation preference model. The parameter adjustment boundary is dynamically optimized based on the behavior deviation trend to form a control interval integration mechanism with real-time feedback.

[0080] It should be further explained that in the posture behavior back-injection and control interval adjustment chain, the fine-tuning behavior response difference is modeled as a function set, and the mold control parameter interval is adjusted by function back-injection, which is expressed by Equation 4:

[0081]

[0082] Variable definition; χ m (t) is the deviation response function of the fine-tuning input behavior of the mth segment; ξ m represents the corresponding historical feedback output response trajectory; φ m is the control rule change function corresponding to the fine-tuning section; Adjust the partial derivative mapping on the mold parameter space (such as angle / force limit) for behavior; Ω m The weight function for the impact of feedback fluctuations on strategy correction (non-constant, dynamic fitting); is the final dynamic control strategy adjustment function range; T m The duration interval length of the mth segment fine-tuning behavior;

[0083] Formula 4 uses each fine-tuning behavior and its feedback trajectory differential to form the action response rate; The part is used to describe the gradient response effect of the current behavior on the boundary of the control rule; the sum integral in the formula forms the strategy adjustment area by accumulating the correction amplitude during the behavior time period; the whole forms a continuous reinjection mechanism for the "dynamically adjustable domain" of the control rule space.

[0084] The extreme value variation ranges of three parameters, namely spatial displacement, thermal state and clamping force, in the structural state feedback matrix are extracted, and a corresponding risk trigger critical parameter set is constructed, covering three risk conditions: spatial position limit, temperature upper limit and clamping force mutation threshold.

[0085] Perform time series analysis on the continuously collected structural state feedback matrix to detect in real time any sudden increase in any dimension parameter that exceeds the trigger critical parameter set, and mark it as a potential risk trigger candidate event;

[0086] The feature vectors of potential risk trigger candidate events are compared with the sample set of historical high-risk events, the risk level is determined based on the similarity weight and time evolution trend, and the corresponding suppression path plan is matched according to the level result.

[0087] Synchronously activate the multi-channel pressing mechanism according to the pressing path plan. The multi-channel pressing mechanism includes a control instruction channel blocking mechanism, a high-intensity tactile prompt mechanism based on direction enhancement, and an interface information flash blocking mechanism to achieve operation interruption feedback linkage when the mold is in a dangerous state.

[0088] All confirmed high-risk behavior samples and their corresponding trigger conditions, response paths, and suppression result data are sent to the high-risk behavior record collection for subsequent strategy evolution and model retraining;

[0089] It should be further explained that the over-limit events are extracted from the continuous structural state changes, the risk level is evaluated and the suppression mechanism is mapped, which can be described by Equation 5.

[0090]

[0091] where ρ r is the current value of the risk variable of type r (displacement, temperature, clamping force); is the second-order disturbance intensity (volatility) of the variable; Δτ r is the critical mutation time window detected in the sampling interval; Φ r Trigger the feedback direction sequence function in the path for the corresponding variable; is the risk acceleration factor, which is used to amplify the rapidly evolving trend; It is an indicator function that marks whether the variable exceeds the risk threshold; is the intensity of suppression strategy activation mapped to the final risk level; R represents the total number of risk variables involved in risk identification and suppression judgment;

[0092] In Equation 5, the second-order disturbance is used to extract whether the current variable is in a high-order fluctuation state; Emphasize the drastic jump trend of the response path in a short period of time; the indicator function only includes the total risk value for the variable exceeding the limit; the overall output suppression intensity Used to initiate tactile / visual / control channel blocking actions.

[0093] A human-computer interaction system for injection molds based on real-time feedback, including a data acquisition module, a posture reconstruction module, an instruction mapping module, a state perception module, a tactile generation module, and a feedback closed-loop module;

[0094] The data acquisition module is used to collect the operator's spatial operation data through the gesture integrated recognition sensor worn by the operator. The spatial operation data includes angular velocity, angular acceleration, displacement velocity and gesture angle, and forms an action input vector set based on the spatial operation data;

[0095] The posture reconstruction module is used to process the action input vector set through inertial fusion and Kalman filtering to form three-dimensional posture trajectory data, and calibrate it into a feature operation trajectory template;

[0096] The instruction mapping module is used to dynamically match the 3D posture trajectory data with the operation behavior control parameter mapping table. The operation behavior parameters of the operation behavior control parameter mapping table include the template drive path, the slider opening and closing positions, and the structural control instructions of the pull rod synchronization process, and output the corresponding structural response execution command sequence;

[0097] The state perception module is used to continuously collect data on three physical parameters of each structure in the injection mold: spatial displacement, motion rate, and clamping force. The collected results are combined by component type and aligned based on timestamps to obtain a structural state feedback matrix.

[0098] The tactile generation module is used to convert the structural state feedback matrix into a tactile feedback signal model through a tactile mapping function. The multi-channel drive unit drives the posture integrated recognition sensor worn by the operator to output a multi-dimensional tactile signal including vibration direction, frequency amplitude, and surface temperature.

[0099] The feedback closed-loop module is used to complete dynamic regulation of posture control based on multi-dimensional tactile signals by the operator, synchronously reconstruct the action input vector and perform continuous iterative feedback to form human-computer interaction.

[0100] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A human-computer interaction method for an injection mold based on real-time feedback, comprising a gesture integrated recognition sensor, characterized in that: The operator's spatial operation data is collected through the gesture integrated recognition sensor worn by the operator. The spatial operation data includes angular velocity, angular acceleration, displacement velocity and gesture angle, and an action input vector set is formed based on the spatial operation data. The action input vector set is processed by inertial fusion and Kalman filtering to form three-dimensional posture trajectory data, and is calibrated as a characteristic operation trajectory template; Dynamically match the three-dimensional posture trajectory data with the operation behavior control parameter mapping table. The operation behavior parameters of the operation behavior control parameter mapping table include the template drive path, the slider opening and closing positions, and the structural control instructions of the pull rod synchronization process, and output the corresponding structural response execution command sequence; Continuous data acquisition is performed on three physical parameters of each structure in the injection mold: spatial displacement, motion rate, and clamping force. The acquired results are combined by component type and aligned based on timestamps to obtain a structural state feedback matrix. The structural state feedback matrix is ​​converted into a tactile feedback signal model through a tactile mapping function. The posture integrated recognition sensor worn by the operator is driven by a multi-channel drive unit to output a multi-dimensional tactile signal including vibration direction, frequency amplitude, and surface temperature. The operator completes dynamic regulation of posture control based on multi-dimensional tactile signals, synchronously reconstructs the action input vector and performs continuous iterative feedback to form human-computer interaction.

2. The method for human-computer interaction of an injection mold based on real-time feedback according to claim 1, characterized in that: performing normalization and semantic integration operations on raw sensory data contained in a structural state feedback matrix to generate a structural state fusion vector block, wherein the raw sensory data includes spatial displacement data, posture angle data, clamping force data, thermal state data, and cooling flow rate data; Dynamically jointly modeling the structural state fusion vector block with the 3D posture trajectory data, which is obtained by posture solution of the action input vector set, is used to establish the corresponding characteristic relationship between the operator's action intention and the mold response state; A structural state perception map is constructed based on the corresponding feature relationship. The structural state perception map presents the linkage path between the response state of each substructure of the injection mold and the operator's posture input in the form of multi-dimensional spatial logical aggregation.

3. The method for human-computer interaction of an injection mold based on real-time feedback according to claim 2, characterized in that: The linkage relationship between the response states of each substructure of the injection mold presented in the structural state perception map and the operator's posture input is constructed into a response pattern set according to three dimensions: spatial position, time rhythm and behavioral intention. The response pattern set is converted into prompt elements, which include structural position guidance, inertia offset trend and error risk blocks. The prompt elements are synchronously fed back to the operator's perception path through a graphical interface path and a tactile signal channel to assist the operator in real-time posture adjustment and path judgment.

4. The method for human-computer interaction of an injection mold based on real-time feedback according to claim 3, characterized in that: The operator performs temporal fusion on multiple rounds of gesture input trajectories and mold response feedback sequences formed during real-time interaction, constructing a causal chain of operation behaviors based on feedback history. The causal chain includes the action start state, transition state, and completion state corresponding to different feedback results, and performs conditional calibration on the trigger relationship between each state to form a path trigger condition set. The causal chain of the operation behavior is converted into a response state graph structure according to the path mapping relationship between the posture input trajectory and the injection mold response feedback sequence. The response state graph structure contains three types of state transition forms: jump nodes, fallback paths and nested response branches.

5. The method for human-computer interaction of an injection mold based on real-time feedback according to claim 4, characterized in that: The response amplitude value and delay data in the structural state feedback matrix are introduced into the response state diagram structure as the quantitative weight of the operation behavior constraint conditions to construct the operation behavior restriction map; The behavior path restriction conditions, false touch rejection paths and state fault tolerance intervals generated in the operation behavior restriction map are applied to the control interface where the operator inputs instructions, realizing a path-based control strategy based on feedback status, a real-time correction mechanism and structural cascade fault tolerance.

6. The method for human-computer interaction of an injection mold based on real-time feedback according to claim 5, characterized in that: Perform difference analysis on the operator's continuous input adjustment behavior during the posture control process, extract fine-tuning behavior events with feedback result change characteristics, and calibrate the difference between the posture input change and the structural response corresponding to each event; The extracted fine-tuning behavior events are clustered according to the feedback difference characteristics and operation intention trends, and a set of behavior correction response functions that reflect the control correction rules are constructed; The behavior correction response function set is input into the control rule adjustment path for dynamically adjusting the structure execution parameters, and the angle limit, action execution rate and response scheduling strategy are parameterized according to the function content; The operator's historical input behavior sequence and corresponding feedback change data are integrated into the control rule adjustment path to construct an empirical operation preference model. The parameter adjustment boundary is dynamically optimized in combination with the behavior deviation trend to form a control interval integration mechanism with real-time feedback.

7. The method for human-computer interaction of an injection mold based on real-time feedback according to claim 6, characterized in that: The extreme value variation ranges of three parameters, namely spatial displacement, thermal state and clamping force, in the structural state feedback matrix are extracted, and a corresponding risk trigger critical parameter set is constructed, covering three risk conditions: spatial position limit, temperature upper limit and clamping force mutation threshold. Perform time series analysis on the continuously collected structural state feedback matrix to detect in real time any sudden increase in any dimension parameter that exceeds the trigger critical parameter set, and mark it as a potential risk trigger candidate event; The feature vectors of potential risk trigger candidate events are compared with the sample set of historical high-risk events, the risk level is determined based on the similarity weight and time evolution trend, and the corresponding suppression path plan is matched according to the level result.

8. The method for human-computer interaction of an injection mold based on real-time feedback according to claim 7, characterized in that: Synchronously activate the multi-channel pressing mechanism according to the pressing path plan. The multi-channel pressing mechanism includes a control instruction channel blocking mechanism, a tactile prompt mechanism based on direction reinforcement, and an interface information flash blocking mechanism to achieve operation interruption feedback linkage when the mold is in a dangerous state. All confirmed risk behavior samples and their corresponding trigger conditions, response paths, and suppression result data are sent to the risk behavior record set for subsequent strategy evolution and model retraining.

9. A human-computer interaction system for injection molds based on real-time feedback, comprising a data acquisition module, a posture reconstruction module, a command mapping module, a state perception module, a tactile generation module, and a feedback closed-loop module, characterized in that: The data acquisition module is used to collect the operator's spatial operation data through the gesture integrated recognition sensor worn by the operator. The spatial operation data includes angular velocity, angular acceleration, displacement velocity and gesture angle, and forms an action input vector set based on the spatial operation data; The posture reconstruction module is used to process the action input vector set through inertial fusion and Kalman filtering to form three-dimensional posture trajectory data, and calibrate it into a feature operation trajectory template; The instruction mapping module is used to dynamically match the 3D posture trajectory data with the operation behavior control parameter mapping table. The operation behavior parameters of the operation behavior control parameter mapping table include the template drive path, the slider opening and closing positions, and the structural control instructions of the pull rod synchronization process, and output the corresponding structural response execution command sequence; The state perception module is used to continuously collect data on three physical parameters of each structure in the injection mold: spatial displacement, motion rate, and clamping force. The collected results are combined by component type and aligned based on timestamps to obtain a structural state feedback matrix. The tactile generation module is used to convert the structural state feedback matrix into a tactile feedback signal model through a tactile mapping function. The multi-channel drive unit drives the posture integrated recognition sensor worn by the operator to output a multi-dimensional tactile signal including vibration direction, frequency amplitude, and surface temperature. The feedback closed-loop module is used to complete dynamic regulation of posture control based on multi-dimensional tactile signals by the operator, synchronously reconstruct the action input vector and perform continuous iterative feedback to form human-computer interaction.

Images