An automated chip placement device and a method for optimizing the material handling path.

Abstract

This invention relates to the field of microelectronics manufacturing equipment technology, and discloses an automatic placement device and a material picking path optimization method. The device includes: a three-axis fixed material picking system configured to pick up materials using a nozzle and place them onto a target; the nozzle maintains a constant picking position throughout the placement process, performing only up-and-down movements in the Z-axis direction; a modular stage system configured to distribute materials into various modules on the stage, each module capable of actively moving the target material to a fixed picking position below the nozzle according to the placement task; an intelligent path planning system configured to dynamically generate material sequence and path scheduling plans based on the placement task; and a visual positioning coordinate compensation system configured to acquire images of the fixed picking position and perform real-time identification and deviation correction of the actual center point of the material. This invention can significantly reduce equipment structural complexity and path uncertainty, and improve system responsiveness and compatibility.

Description

Technical Field

[0001] This invention relates to the field of microelectronics manufacturing equipment technology, and in particular to an automatic chip mounting device and a method for optimizing the material handling path. Background Technology

[0002] Automated chip placement technology plays a crucial role in chip micro-assembly processes. Placement devices use nozzles to pick up chips from the material placement area and mount them to the target location. Currently, commonly used automated placement nozzle devices typically operate by having the nozzle frequently move back and forth between the material placement area and the placement position. The different positions of the materials result in varying path lengths the nozzle must travel for each pickup. This approach suffers from complex path planning, long nozzle movement paths, and long equipment uptime, severely limiting the improvement of chip placement efficiency. Furthermore, long-distance reciprocating motion accelerates wear on the nozzle's mechanical components, reducing equipment stability and lifespan.

[0003] To address these issues, some existing technologies attempt to optimize path planning algorithms or increase the operating speed of mechanical structures. However, existing algorithms often rely on static path planning, making it difficult to quickly handle the frequent changes in material position during actual placement processes, and the path optimization effect is limited. Furthermore, the lifting speed of mechanical structures has physical limits, and high-speed movement increases equipment vibration and positioning errors, reducing placement accuracy. Therefore, while existing methods have made some optimizations, they still struggle to simultaneously meet the requirements of speed, accuracy, and stability in practical applications. Summary of the Invention

[0004] To address the problems of excessively long pick-up paths, insufficient positioning accuracy, and low production efficiency in traditional chip micro-assembly equipment, this invention proposes an automatic chip mounting device and a pick-up path optimization method, which can significantly reduce the complexity of equipment structure and path uncertainty, and improve system responsiveness and compatibility.

[0005] The technical solution adopted in this invention is as follows:

[0006] An automatic patch placement device, comprising:

[0007] The three-axis fixed material handling system is configured to pick up materials through a nozzle and attach them to the target; the nozzle maintains a constant picking position throughout the entire attachment process and only performs up-and-down movements in the Z-axis direction;

[0008] The modular stage system is configured to distribute materials to various modules on the stage, and each module can actively move the target material to a fixed pick-up position below the nozzle according to the mounting task.

[0009] The intelligent path planning system is configured to dynamically generate material sequence and path scheduling plans based on the mounting task, and achieve optimal scheduling among multiple modules in the modular platform system through an improved ant algorithm.

[0010] The visual positioning coordinate compensation system is configured to acquire images of a fixed material picking position and perform real-time identification and deviation correction of the actual center point of the material.

[0011] Furthermore, the three-axis fixed material handling system includes a suction nozzle, a gantry support frame structure, and a servo motor drive device. The suction nozzle is mounted on the gantry support frame structure and is driven by the servo motor drive device. The suction nozzle adopts a modular structure design and is equipped with a snap-fit ​​installation interface.

[0012] Furthermore, the modular stage system is based on a two-dimensional matrix layout, with each module corresponding to a standard material position; each module integrates a micro servo motor or linear drive unit, which can drive the corresponding module to perform lateral and longitudinal positioning fine-tuning, for aligning materials with the suction nozzle or realizing positioning compensation.

[0013] Furthermore, in the visual positioning coordinate compensation system, image acquisition is performed on the fixed material picking position, and the actual center point of the material is identified in real time. This includes: before each material picking action, a CCD camera is used to capture a real-time image of the material's position on the platform. After the image is illuminated by a light source, the outline, center point, and feature points of the material are displayed. After the image data is transmitted to the image processing unit, the actual position coordinates of the material are calculated using a sub-pixel edge extraction and pattern matching algorithm.

[0014] Furthermore, in the visual positioning coordinate compensation system, the actual center point of the material is identified and the deviation is corrected in real time, including: based on the preset fixed material picking position coordinates, the deviation between the actual position coordinates of the material and the preset fixed material picking position coordinates is calculated in real time by the image processing unit, and the compensation displacement of the platform is calculated by the coordinate error compensation algorithm; finally, the error correction is completed with minimal movement by the platform micro-motion control command, and the material is transported to the preset position.

[0015] A method for optimizing the material handling path of an automated placement device includes:

[0016] Placement process analysis: Taking the placement process document as input, extract the types, order, quantity and spatial coordinate information of the required materials to establish a structured material requirement sequence;

[0017] Time-space mapping modeling: Constructing a dynamic graph structure model that integrates the temporal sequence and physical spatial coordinates in the mounting task;

[0018] Improved Ant Algorithm Optimization: Based on a dynamic graph structure model, the improved ant algorithm is used to calculate the optimal path solution.

[0019] Furthermore, in the mounting process parsing, the mounting process document is used as input to extract the types, order, quantity, and spatial coordinate information of the required materials, and a structured material requirement sequence is established, including:

[0020] Standardized interface for placement process: It is compatible with multiple placement process file formats through a standardized data parsing interface, and extracts the material number, package type, placement sequence number and corresponding warehouse location coordinates required in each placement task;

[0021] Material Requirements Sequence Construction Logic: After extracting the data, it is arranged according to the mounting sequence number to form a one-dimensional time series structure. Each element It includes the material ID, target coordinates, and required time point, which are used to establish a time-space mapping later.

[0022] Furthermore, in the time-space mapping modeling, the temporal sequence and physical spatial coordinates in the mounting task are constructed into a dynamic graph structure model, including:

[0023] Construction of a temporal-spatial coupled graph model: Constructing a weighted directed graph As a dynamic graph structure model, where V is the set of nodes and E is the set of edges; each node Corresponding to one material requirement item Edges between nodes Indicates from materials arrive The transition path is defined by the weight of the edges, which is the Euclidean distance between the two points in space. Or weighted estimated time cost ;

[0024] Dynamic constraint factor insertion: Dynamic constraints are inserted into the dynamic graph structure model. The dynamic constraints include mounting order dependency constraints, time window constraints, and stage movement boundary and physical obstacle area constraints.

[0025] Real-time path graph updates: Based on a dynamic graph structure model, completed nodes are deleted and the status of nodes to be installed is updated in real time according to the progress of the installation task. Urgent task nodes can be re-inserted during the installation process.

[0026] Furthermore, in the improved ant colony optimization, the path-optimal solution is calculated based on a dynamic graph structure model and combined with the improved ant colony algorithm, including:

[0027] Pheromone initialization and heuristic function setting: Initialize pheromones for all edges of the dynamic graph structure model. Heuristic function Based on the side length distance or time cost, the smaller the heuristic factor, the higher the priority; through weighting coefficients... Controlling the influence of pheromones and heuristic functions;

[0028] Probabilistic path selection: Each ant starts from the initial node and calculates the transition probability:

[0029]

[0030] in, The transition probability is given by k, where k is the ant number and s is a candidate node among all the possible next nodes that the current ant can choose when it is at node i. This means treating each node in the set as a possible next destination in turn, and summing up the probability components of all these candidate points in the denominator to normalize the result; The parameter represents the importance of pheromones; the larger the value, the stronger the influence of pheromones on path selection. β represents the importance of heuristic functions; the larger the value, the stronger the influence of distance or time cost. This represents the pheromone concentration along the edge from node i to node s; This is heuristic information; the larger the value, the closer or faster the journey from i to s, and therefore it is preferred.

[0031] Pheromones Update: After each iteration, update the pheromones for all paths.

[0032]

[0033] in, The pheromone evaporation coefficient; The path feedback strength is set inversely to the total path cost; the shorter the path, the more pheromones are retained.

[0034] Path output: Output path sequence The optimal movement command executed by the stage, where This indicates the target node moved to in step 1; This indicates the location of the last material that needs to be picked up or mounted, and is the endpoint instruction of the entire scheduling path.

[0035] Furthermore, in the improved ant colony optimization, based on the dynamic graph structure model and combined with the improved ant colony algorithm to calculate the optimal path solution, it also includes the introduction of a multi-strategy improvement mechanism: introducing a local search mechanism; adopting an elite ant strategy to retain the pheromone enhancement of the best path in this round; and dynamic pheromone adjustment to suppress early convergence and avoid getting trapped in local optima.

[0036] The beneficial effects of this invention are as follows:

[0037] 1. Fixed Pickup Position Dynamic Coordination System: By fixing the pickup position of the nozzle, the platform actively transports the material to the fixed position, effectively avoiding the problem of extended movement path caused by frequent changes in the pickup position of the nozzle, greatly shortening the pickup path of the nozzle, significantly improving the placement efficiency, and effectively reducing equipment wear and failure rate.

[0038] 2. Intelligent path planning algorithm: By dynamically analyzing the placement program and using an improved ant algorithm to optimize the movement path of the platform, the optimal material conveying path can be obtained in real time, further reducing the platform movement distance and waiting time, and significantly improving the operating efficiency and stability of the entire placement process.

[0039] 2. Modular stage system: Through matrix layout and independent servo drive unit design, the stage system achieves high-efficiency response and precise control. Combined with intelligent identification module to automatically verify and track material information, it can quickly and accurately transport the materials to be mounted to the fixed pick-up position, effectively improving the accuracy of material management and the overall operating efficiency of mounting, and ensuring the smoothness and stability of the mounting process. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of a three-axis fixed material handling system according to Embodiment 1 of the present invention. Detailed Implementation

[0041] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments are now described. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention; that is, the described embodiments are only a part of the embodiments of the invention, not all of them. 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.

[0042] Example 1

[0043] To address the problems of excessively long pick-up paths, insufficient positioning accuracy, and low production efficiency in traditional chip placement equipment during chip micro-assembly, this embodiment provides an automated placement device, including:

[0044] The three-axis fixed material handling system is configured to pick up materials through a nozzle and attach them to the target; the nozzle maintains a constant picking position throughout the entire placement process, and only performs up-and-down movements in the Z-axis direction;

[0045] The modular stage system is configured to distribute materials to various modules on the stage, and each module can actively move the target material to a fixed pick-up position below the nozzle according to the mounting task.

[0046] The intelligent path planning system is configured to dynamically generate material sequence and path scheduling plans based on the mounting task, and achieve optimal scheduling among multiple modules in the modular platform system through an improved ant algorithm.

[0047] The visual positioning coordinate compensation system is configured to acquire images of a fixed material picking position and perform real-time identification and deviation correction of the actual center point of the material.

[0048] It should be noted that the automatic placement device in this embodiment achieves a fixed and minimal material picking path length by fixing the nozzle's pick-up position and dynamically conveying the material to the fixed pick-up point via an actively movable platform. The integration of a three-axis fixed pick-up system and a vision positioning coordinate compensation system ensures high positioning accuracy. Through the intelligent path planning system, the chip placement sequence and material demand sequence can be dynamically analyzed to construct a time-space mapping model. The improved ant algorithm is used to optimize the movement path of the stage in real time. In the modular stage system, each module has an independent servo drive unit, which responds quickly and can realize automatic material verification and tracking, thereby effectively improving the speed, accuracy and stability of the entire chip placement process and meeting the precision requirements of high-end micro-assembly technology. The specific details are as follows.

[0049] (a) Dynamic coordination of fixed material picking positions

[0050] 1. Three-axis fixed material handling system

[0051] In this embodiment, the three-axis fixed material handling system is a key hardware component for fixing the material handling position, and its specific structure is as follows.

[0052] (1) Gantry support frame structure

[0053] The gantry support frame structure, serving as the main framework of the three-axis fixed material handling system, can be made of high-strength aluminum alloy or carbon fiber composite materials. It possesses high rigidity and low thermal deformation characteristics, ensuring high stability of the suction nozzle's material handling position during long-term operation. The main body of the gantry support frame structure is a box-shaped structure with internal reinforcing ribs to further improve overall rigidity and vibration resistance, ensuring that the equipment will not deform during long-term continuous operation.

[0054] (2) High-precision servo drive Z-axis module

[0055] The suction nozzle moves up and down along the Z-axis of the gantry support frame, driven and controlled by a precision linear guide and a servo motor. Figure 1 As shown. Preferably, a high-precision servo motor drive device is used, coupled with a precision ball screw linear direct drive motor, to achieve precise speed and position control. Specifically, the repeatability of the Z-axis motion is designed and controlled within... Within the range, actual operation is stable at Inside.

[0056] (3) Fixed XY axis material picking and positioning module

[0057] In the XY axis direction, the nozzle position remains fixed, eliminating the frequent horizontal reciprocating movement of the nozzle between material areas in traditional methods, thus achieving a constant nozzle picking position. The nozzle positioning module adopts a combination structure of high-precision rigid connectors and fixed brackets to ensure the absolute stability of the XY axis position during long-term operation and avoid position drift.

[0058] (4) Suction nozzle assembly and quick-change structure

[0059] To improve equipment flexibility and maintenance efficiency, the nozzles adopt a modular design with quick-connect snap-on installation interfaces, allowing for rapid nozzle replacement in a very short time. The modular design also supports quick replacement of various nozzle sizes to meet the needs of mounting materials of different sizes, thus improving the system's applicability and versatility.

[0060] 2. Visual positioning coordinate compensation system

[0061] The visual positioning coordinate compensation system is the core component for achieving precise positioning, as detailed below.

[0062] (1) Hardware configuration of visual positioning system

[0063] It can employ a combination of a high-definition industrial-grade CCD camera and a high-performance lens, along with a stable and uniform industrial-grade ring or strip light source. Specifically, the CCD camera has a resolution of no less than 5 megapixels, and the lens focal length and field of view are carefully designed according to the actual material size and positioning accuracy requirements to ensure image clarity and positioning accuracy.

[0064] (2) Real-time image acquisition and processing methods

[0065] Before each material handling operation, a CCD camera captures a real-time image of the material's position on the platform. After illumination, the image clearly shows the material's outline, center point, and feature points. The image data is then transmitted to the image processing unit, where sub-pixel edge extraction and pattern matching algorithms (such as template matching, morphological algorithms, or deep learning localization networks) are used to quickly and accurately calculate the material's actual position coordinates.

[0066] (3) Coordinate error calculation and position compensation algorithm

[0067] The system internally presets fixed material handling position coordinates. The image processing unit calculates the minute deviation between the actual material coordinates and the preset coordinates in real time, and accurately calculates the compensation displacement of the platform through coordinate error compensation algorithms (such as PID compensation or adaptive fuzzy control). Finally, through the platform's micro-motion control commands, the error correction is quickly completed with minimal movement, accurately conveying the material to the preset position.

[0068] 3. Coordinated control mechanism for the active movement of the suction nozzle and the stage

[0069] To achieve precise coordination between nozzle fixation and stage active movement, this embodiment employs a collaborative control mechanism:

[0070] (1) Real-time communication and instruction coordination mechanism

[0071] The stage and the vision positioning system communicate via a real-time industrial communication bus (such as EtherCAT or CANopen protocol) to achieve high-speed data transmission and real-time position adjustment. The vision positioning system feeds back material deviation data to the stage control system in real time, and the stage control system generates compensation actions in real time based on the vision data, ensuring a high degree of coordination and real-time cooperation between the two.

[0072] (2) Collaborative control closed-loop feedback structure

[0073] The motion control system of the platform detects the current position in real time through a built-in encoder and feeds it back to the vision positioning and main control system, forming a closed-loop control structure. After each fine-tuning action, the vision positioning system verifies the material position again in real time to confirm the effect of the compensation action, further reducing positioning errors and ensuring the accuracy and stability of the entire material handling process.

[0074] (3) Adaptive error compensation control strategy

[0075] Considering potential external environmental disturbances (such as equipment vibration and thermal drift) in actual operating conditions, an adaptive mechanism is added to the system control strategy. Through statistical analysis of historical positioning errors and trend prediction algorithms (such as Kalman filters), the error compensation parameters are automatically optimized to improve the compensation effect and system stability during long-term operation.

[0076] 4. Analysis of Actual Operation Results and Technical Indicators

[0077] Through actual operation testing and verification, this device has achieved the following technical indicators during continuous operation on a microelectronic chip assembly production line:

[0078] The long-term stability of the suction nozzle's fixed material-picking position reaches within;

[0079] The actual positioning accuracy of the visual positioning compensation system is stably controlled within ~ Inside;

[0080] The system effectively reduces the nozzle travel distance by more than 80%, significantly reducing mechanical wear;

[0081] The positioning compensation process has a response time of less than 50ms, ensuring the requirements for high-speed continuous placement.

[0082] Compared to traditional nozzle movement solutions, the overall material handling path optimization improves efficiency by more than 35%, and the overall system production efficiency increases by more than 30%.

[0083] (II) Modular Platform System

[0084] The modular stage system is one of the key hardware components of this device for achieving fixed nozzle pick-up and optimized path collaborative placement. Through a matrix-layout programmable stage structure, independent high-speed servo drive units, and an intelligent recognition module, this system achieves high-precision, high-response, and intelligent material movement and management capabilities. Compared to traditional integrated moving material tray designs, this modular system offers greater flexibility, maintainability, and scalability, making it particularly suitable for high-density chip or multi-type device parallel placement scenarios. The following sections will elaborate on its structural layout design, drive control system, intelligent recognition and status tracking, and system collaboration mechanisms.

[0085] 1. Matrix modular structure layout design

[0086] (1) Modular array arrangement structure

[0087] The entire stage system adopts a two-dimensional matrix layout, with each module unit corresponding to a standard material position. The module size can be customized according to the material packaging size and stage space utilization (e.g., 20×20 mm or 40×40 mm modules), supporting a mixed arrangement of various mounting specifications.

[0088] (2) Standardized modular unit construction

[0089] Each module unit includes the following core components:

[0090] Miniature tray structure: used to fix chips, components, films and other mounting materials, with anti-static and dustproof design.

[0091] Module housing structure: It adopts lightweight materials and encapsulates standard interfaces, possessing mechanical strength and anti-interference capabilities.

[0092] Quick-connect power and communication connectors: support plug-and-play and hot-swappable capabilities, facilitating module replacement and maintenance.

[0093] Upper-level system identification QR code or RFID interface area: used for unique identification.

[0094] (3) Scalable chain architecture design

[0095] The stage supports expansion in any horizontal or vertical dimension, suitable for the capacity requirements of different production lines or placement tasks. Modules can be dynamically added or removed without stopping the system, improving system flexibility.

[0096] 2. Independent high-speed servo drive control system

[0097] (1) Module-level drive motor configuration

[0098] Each module integrates a miniature servo motor or linear drive unit, enabling rapid lateral and longitudinal positioning and fine-tuning within a small space (typical displacement range 1~). Positioning accuracy ), used for aligning materials with the suction nozzle or for achieving positioning compensation.

[0099] (2) High-speed response and closed-loop control capability

[0100] Each servo unit integrates an encoder, enabling sub-millisecond closed-loop position feedback control. Typical response time is controlled within the range of 8~10ms, meeting the requirements of thousands of pick-and-place cycles per minute.

[0101] (3) Platform collaborative drive control architecture

[0102] All module-level drivers are uniformly scheduled by the main control system, supporting centralized / distributed control architectures. High-speed synchronous transmission of drive commands and status data is achieved based on bus-type communication protocols (such as CANopen, RS485, or EtherCAT).

[0103] (4) Security control and conflict prediction mechanism

[0104] The system features a predictive collision detection mechanism that can pre-lock relevant modules based on the placement path, avoiding path interference or collision risks when multiple nozzles operate concurrently. If a sudden deviation is detected (such as abnormal resistance or sudden current jump), it immediately enters a safety stop mode and sends an alarm.

[0105] 3. Intelligent identification and material status tracking module

[0106] (1) Design of integrated recognition module

[0107] Each module integrates a QR code / barcode reader or RFID reading antenna on its bottom or side to read the unique identification code of the material tray or chip package, enabling accurate acquisition of material type, batch, and location information.

[0108] (2) State awareness and verification mechanism

[0109] The platform module can sense its current material status in real time (whether it is empty, whether it has been grabbed, or whether it is being repeatedly placed) and synchronize it to the main control system via a status table. In conjunction with the placement task list system, it verifies whether the current material matches the task target, preventing placement errors or material mixing.

[0110] (3) Construction of a full-process traceability chain

[0111] All materials are uniquely numbered before entering the platform system, and placement records are automatically registered after placement, forming a complete digital traceability chain from material loading to placement and packaging. This chain can interface with the MES (Manufacturing Execution System) to achieve visualized quality management of the manufacturing process.

[0112] 4. Coordinated scheduling mechanism between the stage system and the suction nozzle

[0113] (1) Position calibration and mission prediction

[0114] Once the main control system detects the location of the next material to be mounted, it controls the stage module to move the material to the designated nozzle pick-up position in advance, and simultaneously notifies the nozzle that it is about to pick up the material.

[0115] (2) Task scheduling and dynamic response

[0116] The main control system dynamically schedules the response order of multiple modules based on path optimization results. Module actions are independent of each other and have priority control capabilities, enabling a balanced mode of partial first action and overall coordination, thereby improving system response speed.

[0117] (3) Closed-loop linkage of nozzle status feedback

[0118] After the suction nozzle completes one suction cycle, it provides real-time feedback on the current status to the platform. Based on this feedback, the platform dynamically adjusts the material replenishment rhythm and decides whether to activate the next module and advance the task rhythm, thus forming a synchronized scheduling of the entire system's rhythm.

[0119] 5. System performance and parameter indicators

[0120] Through integration testing on an actual chip micro-assembly production line, the modular stage system achieved the following technical performance indicators:

[0121] Module response time: Average response time is 8~10mm, maximum not exceeding ;

[0122] Positioning accuracy: Module displacement error is controlled within... scope;

[0123] System flexibility: Module hot-swap time does not exceed 5 seconds, enabling uninterrupted maintenance;

[0124] The placement error rate decreased by more than 90%, and the overall cycle time efficiency increased by more than 30%.

[0125] All mounting record information can be tracked, statistically analyzed, and exported, supporting data-driven quality management.

[0126] Example 2

[0127] This embodiment is based on embodiment 1:

[0128] To address the problems of irregular material pick-up paths, slow response times, and low overall system efficiency in traditional chip placement equipment, this embodiment provides a method for optimizing the material pick-up path of an automated chip placement device, including:

[0129] Placement process analysis: Taking the placement process document as input, extract the types, order, quantity and spatial coordinate information of the required materials to establish a structured material requirement sequence;

[0130] Time-space mapping modeling: Constructing a dynamic graph structure model that integrates the temporal sequence and physical spatial coordinates in the mounting task;

[0131] Improved Ant Algorithm Optimization: Based on a dynamic graph structure model, the improved ant algorithm is used to calculate the optimal path solution.

[0132] This path optimization method possesses the capabilities of process logic-driven, adaptive optimization, and real-time updates, which can significantly reduce the ineffective movement of the stage and improve the overall placement cycle time and operational efficiency. The following sections elaborate on this method from four aspects: placement program analysis, time-space mapping modeling, improved ant colony optimization, and operational performance.

[0133] 1. Placement process analysis

[0134] This path optimization method takes the mounting process document as input, extracts the types, order, quantity and spatial coordinate information of the required materials, and establishes a structured material requirement sequence to provide basic data for subsequent path model construction.

[0135] (1) Standardized interface design for mounting process

[0136] Through standardized data parsing interfaces (such as CSV, IPC-2581 format parsing modules), the system is compatible with various surface mount process file formats, extracting the material number, package type, mounting sequence number, and corresponding warehouse location coordinates required for each mounting task.

[0137] (2) Material Requirements Sequence Construction Logic

[0138] After extracting the data, the system arranges it according to the mounting sequence number to form a one-dimensional time series structure. Each of them It includes the material ID, target coordinates, and required time point, which are used to establish a time-space mapping later.

[0139] 2. Temporal-spatial mapping modeling

[0140] To achieve efficient modeling and solvability of path planning, this method constructs a dynamic graph structure model of the "time sequence" and "physical spatial coordinates" in the mounting task.

[0141] (1) Construction of temporal-spatial coupled graph model

[0142] Construct a weighted directed graph Where V is the set of nodes and E is the set of edges; each node Corresponding to one material requirement item Edges between nodes Indicates from materials arrive The transition path is defined by the weight of the edges, which is the Euclidean distance between the two points in space. or weighted estimated time cost .

[0143] (2) Insert dynamic constraint factors

[0144] The following constraints are introduced into the graphical model:

[0145] Placement sequence dependency constraint (some materials must be placed first);

[0146] Time window constraints (certain materials must be completed within a specified time period);

[0147] Platform movement boundaries and physical obstacle area restrictions (avoiding conflict paths).

[0148] (3) Real-time update mechanism for path map

[0149] The path diagram is a dynamic structure that can delete completed nodes and update the status of nodes to be placed in real time as the placement task progresses; it also supports re-inserting emergency task nodes during the placement process, achieving a highly flexible response mechanism.

[0150] 3. Improved Antminer Algorithm Optimization

[0151] Based on the graph model, the path search algorithm uses the Improved AntColony Optimization (IACO) algorithm to calculate the optimal path solution.

[0152] (1) Pheromone initialization and heuristic function design

[0153] All edges are initialized with pheromone values. .

[0154] Heuristic function Based on the side length distance or time cost, the smaller the heuristic factor, the higher the priority;

[0155] Weighting coefficient To control the influence of pheromones and heuristic functions, the following values ​​are often used: .

[0156] (2) Probability path selection rules

[0157] Each ant starts from the initial node and calculates the transition probability according to the following formula:

[0158]

[0159] in, The transition probability is given by k, where k is the ant number and s is a candidate node among all the possible next nodes that the current ant can choose when it is at node i. This means treating each node in the set as a possible next destination in turn, and summing up the probability components of all these candidate points in the denominator to normalize the result; The parameter represents the importance of pheromones; the larger the value, the stronger the influence of pheromones on path selection. β represents the importance of heuristic functions; the larger the value, the stronger the influence of distance or time cost. This represents the pheromone concentration along the edge from node i to node s; This is heuristic information; the larger the value, the closer or faster the journey from i to s, and therefore it is preferred.

[0160] (3) Pheromone update mechanism

[0161] After each iteration, all path pheromones are updated as follows:

[0162]

[0163] in, The pheromone evaporation coefficient, The path feedback strength is set inversely to the total path cost; the shorter the path, the more pheromones are retained.

[0164] (4) Introduction of multi-strategy improvement mechanism

[0165] Introduce local search mechanisms, such as 2-opt fine-tuning of path segments.

[0166] The elite ant strategy is adopted: retain the pheromone reinforcement of the best path in this round.

[0167] Dynamic pheromone adjustment: suppresses early convergence and avoids getting trapped in local optima.

[0168] (5) Path output and dynamic replanning capability

[0169] Final path sequence The optimal movement command executed by the stage, where This indicates the target node moved to in step 1; This indicates the location of the last material to be picked up or placed, and is the endpoint instruction of the entire scheduling path. If path anomalies occur during the placement process due to external intervention, the algorithm can re-invoke the subgraph reconstruction and path replanning process within milliseconds, ensuring system robustness and real-time performance.

[0170] 4. Analysis of Operational Results and Performance Improvement

[0171] After practical testing on multiple surface mount technology (SMT) production lines, this path optimization method demonstrates significant superiority in the following metrics:

[0172] Path length reduction: Compared to the traditional zigzag traversal strategy, the total path length is reduced by an average of 25%.

[0173] Improved response speed: Path generation and update time is less than 100ms, meeting the requirements of high-speed placement.

[0174] Task throughput optimization: The number of placement tasks completed per unit time has increased by approximately 20% to 30%.

[0175] Strong compatibility with dynamic tasks: Supports task queueing and emergency scheduling, improving process flexibility and engineering adaptability.

[0176] Example 3

[0177] This embodiment is based on embodiment 1:

[0178] This embodiment provides an automatic placement device, which modifies the original placement system in a modular way, as detailed below.

[0179] Existing SMT System: On the packaging production line of a high-end chip micro-assembly company, a traditional fully automated SMT machine was used for the positioning and placement of high-density BGA chips. This machine employed a multi-nozzle scanning structure, requiring the nozzles to frequently reciprocate from a large material tray to pick up components sequentially. Due to the large area of ​​the tray and the dispersed locations of the material storage areas, the nozzles' pick-up paths varied significantly each time, resulting in inconsistent movement paths, longer average pick-up times, and impacting the placement cycle time. Furthermore, the trays were prone to slight displacement during high-speed placement, causing unstable nozzle pick-up accuracy, further leading to placement misalignment, repeated pick-ups, and even placement failures. During peak operation, the production line achieved a placement yield of less than 95%, resulting in low equipment utilization, frequent manual intervention, and high production costs.

[0180] The modular modification of the original surface mount system in this embodiment is as follows:

[0181] Modify the nozzle system structure: The original nozzle XY moving structure is replaced with the XYZ three-axis fixed material picking system of Example 1. The material picking position of the nozzle remains constant throughout the entire placement process, and it only performs up and down movements in the Z-axis direction, completely eliminating horizontal reciprocating motion. The structure is simple and the response is fast.

[0182] A modular platform system has been added: the original static material trays have been replaced with a modular platform system. Materials are distributed into each module unit, and each module can actively move the target material to a fixed pick-up position below the nozzle according to the placement task. All modules integrate independent drivers and recognition modules to achieve precise positioning and dynamic material verification, preventing incorrect or missed placement.

[0183] Integrated intelligent path planning system: The system dynamically generates material sequence and path scheduling plans based on the mounting process documents. An improved ant colony algorithm is used to achieve optimal scheduling among multiple modules, reducing interference from the movement of non-target modules and improving scheduling efficiency.

[0184] A visual positioning coordinate compensation system is introduced: To ensure high-precision placement, a high-resolution CCD camera and sub-pixel positioning algorithm are configured at a fixed material picking position to identify and correct the actual center point of the material in real time, thereby improving the accuracy of material picking.

[0185] After the upgrade, the system showed significant improvements in actual operation: the average material pick-up time of the nozzle decreased from 1.25 seconds to 0.72 seconds, and the placement cycle time increased by about 30%; the fluctuation of the nozzle path length approached zero, and the path controllability was greatly enhanced; the placement offset rate was reduced by more than 85%, and the equipment operation was more stable; through modular design, some modules can be quickly replaced when the stage fails, reducing maintenance time by more than 50%.

[0186] Beyond actual production deployments, this system solution has also been applied in multiple R&D prototype platforms. For example, in fields such as heterogeneous integrated chip mounting and wearable flexible electronic module assembly, where the requirements for mounting speed and precision of multi-size devices are higher, traditional systems, due to their complex paths and imprecise control, struggle to meet the demands of rapid trial production and small-batch customization. The fixed nozzle + intelligent scheduling + modular stage approach proposed in this invention significantly reduces equipment structural complexity and path uncertainty, while improving system responsiveness and compatibility.

[0187] Example 4

[0188] This embodiment is based on embodiment 2:

[0189] This embodiment provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the material handling path optimization method for the automatic placement device of Embodiment 2. The computer program can be in the form of source code, object code, executable file, or some intermediate form.

[0190] Example 5

[0191] This embodiment is based on embodiment 2:

[0192] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the material handling path optimization method for the automatic placement device of Embodiment 2. The computer program can be in the form of source code, object code, executable file, or some intermediate form. The storage medium includes any entity or device capable of carrying computer program code, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content contained in the storage medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the storage medium does not include electrical carrier signals and telecommunication signals.

[0193] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

[0194] It should be noted that, for the sake of simplicity, the foregoing method embodiments are described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

Claims

1. An automatic patching device, characterized in that, The application relates to a three-axis fixed material taking system configured to grab materials through a suction nozzle and attach the materials to targets; the suction nozzle keeps a constant material taking position during the whole attaching process and only performs up-down action in the Z-axis direction; A modular object table system configured to distribute materials into each module of the object table, and each module can actively move target materials to a fixed material taking position below the suction nozzle according to an attaching task; An intelligent path planning system configured to dynamically generate a material sequence and a path scheduling plan according to the attaching task, and realize optimal scheduling among multiple modules in the modular object table system through an improved ant algorithm, including: constructing a dynamic graph structure model of the time sequence and the physical space coordinates in the attaching task, and combining the improved ant algorithm to calculate the optimal path solution; A visual positioning coordinate compensation system configured to collect images of the fixed material taking position and identify and correct the actual center point of the material in real time; The three-axis fixed material taking system comprises a suction nozzle, a gantry support frame structure and a servo motor driving device, the suction nozzle is arranged on the gantry support frame structure and is driven through the servo motor driving device; the suction nozzle adopts a modular structure design and is provided with a clamping type mounting interface; The modular object table system is arranged in a two-dimensional matrix form, and each module corresponds to a standard material position; each module is integrated with a micro servo motor or a linear driving unit and can drive the corresponding module to perform horizontal and vertical positioning fine adjustment for aligning the material with the suction nozzle or realizing positioning compensation; In the visual positioning coordinate compensation system, images of the fixed material taking position are collected and the actual center point of the material is identified in real time, including: before each material taking action, a CCD camera is used to shoot a position image of the material on the object table in real time, and the image displays the outline, center point and feature point of the material after being illuminated by a light source; after the image data is transmitted to an image processing unit, the actual position coordinates of the material are calculated through a sub-pixel edge extraction and pattern matching algorithm; In the visual positioning coordinate compensation system, the actual center point of the material is identified in real time and deviation correction is performed, including: based on the preset fixed material taking position coordinates, the deviation between the actual position coordinates of the material and the preset fixed material taking position coordinates is calculated in real time through the image processing unit, and the compensation displacement amount of the object table is calculated through a coordinate error compensation algorithm; finally, the error correction is completed through the object table micro motion control instruction with minimum action, and the material is transported to the preset position. The material taking path optimization method comprises the following steps:

2. The method for optimizing the material taking path of the automatic patching device, applied to the automatic patching device as claimed in claim 1, characterized in that, Attaching program analysis: taking an attaching process file as input, extracting the type, sequence, quantity and space coordinate information of the required materials, and establishing a structured material demand sequence; Time sequence-space mapping modeling: constructing a dynamic graph structure model of the time sequence and the physical space coordinates in the attaching task; Improved ant algorithm optimization: based on the dynamic graph structure model, the optimal path solution is calculated through the improved ant algorithm; In the time sequence-space mapping modeling, the time sequence and the physical space coordinates in the attaching task are constructed into a dynamic graph structure model, including: ​ Temporal-spatial coupled graph model construction: construct a weighted directed graph As a dynamic graph structure model, V is a node set, E is an edge set; each node corresponds to a material demand item , the edge between the nodes represents the transfer path from material to , and the weight of the edge is the Euclidean distance in space or the estimated time cost after weighting ; Dynamic constraint factor insertion: inserting dynamic constraints in the dynamic graph structure model, the dynamic constraints including mounting sequence dependency constraints, time window constraints, and stage movement boundary and physical obstacle area constraints; Path graph real-time update: based on the dynamic graph structure model, real-time deleting completed nodes and updating pending node states according to the progress of the mounting task, and re-inserting emergency task nodes in the mounting process; In the improved ant algorithm optimization, based on the dynamic graph structure model, the improved ant algorithm is used for path optimal solution calculation, and further includes a multi-strategy improvement mechanism: introducing a local search mechanism; adopting an elite ant strategy to retain the best path pheromone information of the current round for reinforcement; and dynamically adjusting the pheromone to inhibit early convergence and avoid falling into a local optimum.

3. The method of claim 2, wherein, In the mounting program analysis, taking the mounting process file as input, the types, sequence, quantity and spatial coordinate information of the required materials are extracted, and a structured material demand sequence is established, including: Mounting program standardized interface: through a standardized data analysis interface, multiple mounting process file formats are compatible, and the material number, package type, mounting sequence number and corresponding warehouse coordinate required in each mounting task are extracted; Material demand sequence construction logic: after data extraction, arrange according to the mounting sequence number, form a one-dimensional time sequence structure wherein each element contains material ID, target coordinate position and required time point, for subsequent establishment of time-space mapping.

4. The method of claim 2, wherein, In the improved ant algorithm optimization, based on the dynamic graph structure model, the improved ant algorithm is used for path optimal solution calculation, including: Pheromone initialization and heuristic function setting: all edges of the dynamic graph structure model are initialized with pheromone ; heuristic function According to the edge length distance or time cost setting, the smaller the heuristic factor, the higher the priority; through the weight coefficient Control the degree of influence of pheromone and heuristic function; Probabilistic path selection: each ant starts from the initial node, and calculates the transition probability: wherein, is the transition probability, k is the ant number; s is a candidate node in all the selectable next nodes of the current ant at node i; indicates that each node in the set is taken in turn as a possible next destination, and the probability components of all these candidate points are added up in the denominator for normalization; indicates the importance of the pheromone parameter, the greater the value, the stronger the influence of the pheromone on path selection; β indicates the importance of the heuristic function parameter, the greater the value, the stronger the influence of the distance or time cost; indicates the pheromone concentration on the edge from node i to node s; is the heuristic information, the greater the value, the closer or faster from i to s, and therefore is preferentially selected; Pheromone update: after each iteration, update all path pheromones: wherein, is the pheromone evaporation coefficient; is the path feedback strength, which is inversely proportional to the total cost of the path, the shorter the path, the more pheromone is retained; Path output: output path sequence Optimal movement instruction executed for the object table, wherein Indicates the target node moved to in Step 1; Indicates the last material position requiring picking or placing, which is the end instruction of the entire scheduling path.

Images