Multi-production-line mixed-flow vehicle tracking and queue optimization control system and method

Abstract

The invention provides a multi-production-line mixed-flow vehicle tracking and queue optimization control system and a multi-production-line mixed-flow vehicle tracking and queue optimization control method. The system comprises a real-time station tracking module, a production queue optimization module and an MES (Manufacturing Execution System) integration module. The real-time station tracking module is fused with an RFID positioning unit and an encoder positioning unit, and high-precision continuous positioning of the vehicle is realized through Kalman filtering; the production queue optimization module is used for dynamically scheduling vehicles in multi-production-line mixed-flow production based on a reinforcement learning algorithm, and carrying out conflict resolution by adopting a space-time resource reservation model; and the MES integration module interacts with the MES in real time through an OPC UA protocol to realize closed-loop control of a station state, an equipment state and a scheduling instruction. The method can improve the real-time tracking precision of the vehicle, improves the queue scheduling efficiency and the system collaboration, and is suitable for multi-production-line mixed-flow production scenes such as automobile manufacturing.

Description

Technical Field

[0001] The present invention relates to the fields of vehicle production management and intelligent manufacturing, and in particular to a vehicle tracking and queue optimization control system and method for mixed flow of multiple production lines. Background Art

[0002] In multi-line mixed production scenarios, such as automobile manufacturing, when vehicles are mixed across different lines and models, it is necessary to achieve real-time accurate positioning of vehicles and efficient queue scheduling. However, existing technologies have the following shortcomings:

[0003] 1. Insufficient real-time tracking accuracy: Traditional barcode scanning, single-point RFID recognition, or single encoder solutions all suffer from data lag, susceptibility to interference, and inability to continuously locate. Specifically, RFID technology is widely used in logistics tracking, but it is susceptible to metal interference in complex industrial scenarios, and the single-point trigger mode cannot provide continuous position information. Encoder technology is used to measure conveyor line displacement, but when used independently, it cannot be associated with vehicle identity information and requires manual data binding. Visual recognition technology is significantly affected by lighting and occlusion, has high computational complexity, and cannot meet real-time requirements.

[0004] 2. Inefficient queue scheduling: Existing MES scheduling, based on fixed rules, uses static scheduling algorithms that are unable to dynamically respond to production disturbances (such as equipment failures and interrupted orders). Some independent optimization algorithms, including genetic algorithms or reinforcement learning, have been used to optimize queues, but these methods lack deep integration with real-time workstation data, resulting in delayed decision-making.

[0005] 3. Poor system coordination: There is a large delay in data interaction between MES and underlying equipment, making it difficult to form real-time closed-loop optimization. Summary of the Invention

[0006] In response to the defects in the existing technology, the purpose of the present invention is to provide a vehicle tracking and queue optimization control system and method for mixed production lines. Through the deep integration technology of RFID+encoder+MES, precise tracking of vehicle workstations and dynamic optimization of production queues can be achieved, thereby improving the flexibility and efficiency of mixed production.

[0007] In order to achieve the above technical effects, the present invention adopts the following technical solutions:

[0008] According to a first aspect of the present invention, a quasi-vehicle tracking and queue optimization control system for mixed production lines is provided, comprising a real-time workstation tracking module, a production queue optimization module, and an MES integration module:

[0009] The real-time workstation tracking module includes an RFID positioning unit, an encoder positioning unit, and a data fusion unit. The RFID positioning unit is used to obtain vehicle identification information and RFID positioning data through a high-temperature ceramic packaged RFID tag and a multi-antenna array reader, and adopts dynamic frequency band switching to achieve anti-interference positioning. The encoder positioning unit includes a main encoder and a redundant incremental encoder, which are used to measure the displacement data of the conveyor line and output the displacement increment ΔS. The data fusion unit is used to fuse the RFID positioning data and the encoder displacement data based on the Kalman filter algorithm to generate the real-time position coordinates of the vehicle.

[0010] The production queue optimization module includes a dynamic scheduling engine and a conflict resolution unit. The dynamic scheduling engine uses a reinforcement learning algorithm to calculate queue adjustment instructions using order priority, equipment status, and workstation load data provided by the MES. The conflict resolution unit allocates spatiotemporal windows to the intersection of multiple production lines based on a spatiotemporal resource reservation model to avoid vehicle path conflicts.

[0011] The MES integration module includes an OPC UA protocol interface and a visual dashboard. The OPC UA protocol interface is used to achieve real-time data interaction with the MES system; the visual dashboard is used to display vehicle trajectory, scheduling status and equipment efficiency indicators.

[0012] Optionally, the RFID tag is embedded and installed at the bottom of the vehicle at 15±5 cm from the ground. The tag storage capacity is 1KB, and the stored information includes the vehicle identification code, order number and vehicle model code.

[0013] Optionally, the multi-antenna array reader deploys four groups of directional antennas at each key node with a horizontal coverage angle of 90°, and activates each antenna group in turn according to a polling cycle of 10ms.

[0014] Optionally, the dynamic frequency band switching detects environmental noise based on received signal strength indication (RSSI) and automatically switches between the 865-868 MHz and 902-928 MHz frequency bands.

[0015] Optionally, the resolution of the main encoder in the encoder positioning unit is 0.1 mm / pulse, the sampling frequency is 1 kHz, and the redundant incremental encoder automatically switches to work when the main encoder fails.

[0016] Optionally, the state equation of the Kalman filter is: k =x k-1 +ΔS+w k , z k =Hx k +v k ; Among them, w k is the process noise, v kis the observation noise, ΔS is the displacement increment, x k is the displacement estimate, z k is the RFID observation value.

[0017] Optionally, the state space of the reinforcement learning algorithm includes order priority, equipment status and workstation load matrix, the action space includes acceleration instructions, deceleration instructions and path switching instructions, and the reward function is: R = α⋅T 节省 −β⋅C 拥堵 ; Among them, α, β are weight coefficients, T 节省 Save time for estimation; C 拥堵 Penalty for path conflict.

[0018] Optionally, the visual dashboard includes: a real-time trajectory map, which is displayed in green, yellow and red according to the vehicle status; a scheduling efficiency panel, which displays the cycle time deviation rate and equipment overall efficiency, and a red warning when the deviation rate exceeds 15%.

[0019] According to a second aspect of the present invention, a method for tracking and optimizing a vehicle queue for a mixed flow of multiple production lines is provided, which is implemented using the above-mentioned vehicle tracking and optimizing a vehicle queue for a mixed flow of multiple production lines, and is characterized by comprising the following steps:

[0020] Step S1. Vehicle entry and binding: When a vehicle enters the first station, a multi-antenna array RFID reader reads the vehicle identification code, order number, and model code from the high-temperature ceramic packaged RFID tag installed on the bottom of the vehicle, records the initial displacement value, and establishes a mapping relationship between the vehicle information and the initial displacement value;

[0021] Step S2. Continuous displacement acquisition: During the vehicle transportation process, the main encoder outputs a pulse signal in real time to calculate the displacement increment. The redundant incremental encoder automatically switches to work when the main encoder fails.

[0022] Step S3. Position correction and data fusion: When the vehicle passes through a key node, the vehicle's absolute position is acquired through an RFID reader. Combined with the encoder displacement data, the position is fused and calculated using a Kalman filter algorithm to generate the vehicle's real-time coordinates. If the RFID is not triggered and the encoder's cumulative error exceeds 100mm, an audible and visual alarm is triggered.

[0023] Step S4. Dynamic optimization of the production queue: Input the real-time vehicle position, order priority, equipment status, and workstation load matrix into the reinforcement learning scheduling engine to calculate acceleration, deceleration, or path switching instructions;

[0024] Step S5. Conflict resolution and instruction issuance: In areas where multiple production lines intersect, potential route conflicts are detected based on the spatiotemporal resource reservation model. Spatiotemporal windows are allocated to vehicles, and conflict-free dispatch instructions are generated and issued to the MES system via the OPC UA protocol.

[0025] Step S6. Closed-loop feedback and parameter adjustment: Collect the instruction execution status and the actual vehicle arrival time, adjust the reward function weight parameters according to the execution results, and automatically block the faulty workstation and enable the preset detour path when a equipment failure is detected.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. The vehicle tracking and queue optimization control system for mixed production lines provided by the present invention integrates RFID and encoder data to improve positioning accuracy and support complex path tracking of mixed production lines.

[0028] 2. The vehicle tracking and queue optimization control system for mixed production lines provided by the present invention shortens the scheduling response time to seconds, improves production capacity utilization, and significantly increases order delivery rate.

[0029] 3. The multi-line mixed-flow vehicle tracking and queue optimization control system provided by the present invention can be seamlessly connected to the mainstream MES system through the OPC UA protocol, reducing deployment costs.

[0030] 4. The multi-antenna RFID array and filtering algorithm used in the vehicle tracking and queue optimization control system for mixed production lines provided by the present invention can effectively suppress interference from the industrial environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0032] Figure 1 This is a module composition diagram of the vehicle tracking and queue optimization control system for mixed production of multiple production lines described in the first embodiment;

[0033] Figure 2 This is a flowchart of the steps of the vehicle tracking and queue optimization control method for mixed flow of multiple production lines described in the second embodiment. DETAILED DESCRIPTION

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Generally, the components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0035] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for protection, but merely represents selected embodiments of the present application. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments in the present application without creative work are within the scope of protection of the present application.

[0036] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined or explained in subsequent figures. In addition, all directional indications in this application (such as up, down, left, right, front, back, bottom...) are only used to explain the relative position relationship, movement, etc. between the components under a specific posture (as shown in the figures). If the specific posture changes, the directional indication will also change accordingly. Furthermore, the descriptions of "first", "second", etc. in the application are for descriptive purposes only and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features.

[0037] First embodiment

[0038] like Figure 1 As shown, this embodiment provides a vehicle tracking and queue optimization control system for mixed production lines, including a real-time workstation tracking module, a production queue optimization module, and an MES integration module, wherein:

[0039] 1. The real-time workstation tracking module is used to achieve high-precision continuous positioning at the workstation level during vehicle transportation. It includes an RFID positioning unit, an encoder positioning unit, and a data fusion unit. Through the collaborative positioning technology of RFID trigger correction and encoder displacement compensation, multi-source data fusion is achieved, improving positioning accuracy while supporting tracking of complex paths in mixed-flow production lines. Among them:

[0040] (1) The RFID positioning unit is used to obtain vehicle identification information through RFID tags and multi-antenna array readers, and perform anti-interference positioning based on dynamic frequency band switching and time division multiple access polling technology.

[0041] In the RFID positioning unit, a high-temperature ceramic packaged tag (operating temperature -40°C to 150°C) is embedded in the vehicle's underbody, 15 ± 5 cm above the ground to reduce interference from metal reflections. The tag has a 1KB storage capacity and stores information including the vehicle identification number (VIN, 17 digits), order number (12 digits), and vehicle model code (3 digits).

[0042] At the same time, multi-antenna arrays are deployed at key nodes of the production line, with four sets of directional antennas at each node, a horizontal coverage angle of 90°, activated sequentially according to a 10ms polling sequence, and using the time division multiple access (TDMA) protocol.

[0043] The RFID positioning unit can also perform anti-interference positioning based on dynamic frequency band switching technology, detect environmental noise in real time based on the received signal strength indication (RSSI) and dynamically switch between the 865–868MHz / 902–928MHz frequency bands.

[0044] At the same time, the deployment spacing D can be calculated based on the conveyor line speed v (m / s), where the deployment spacing satisfies D≤20 / v. When v≤0.5m / s, D=20m.

[0045] (2) The encoder positioning unit measures conveyor line displacement and outputs incremental displacement. Its hardware configuration includes redundant designs for both a main encoder and an incremental encoder. The main encoder is absolute (resolution 0.1 mm / pulse), directly coupled to the drive shaft, and has a sampling frequency of 1 kHz. The backup encoder is incremental and automatically switches to the backup encoder if the main encoder fails. The mean time between failures (MTBF) is greater than 100,000 hours.

[0046] The unit calculates the displacement increment of the conveyor line according to the following displacement calculation formula: ΔS = N×K, where K is 0.1mm and N is the pulse count.

[0047] (3) The data fusion unit is used to fuse the RFID positioning data and the encoder displacement data based on the Kalman filter to generate the real-time position coordinates of the vehicle. The state equation implemented by the Kalman filter includes the observation state equation and the prediction state equation. The observation state equation is: k =x k-1 +ΔS+w k , the predicted state equation is: k =Hx k +v k ; where x k is the displacement estimate, z k is the RFID observation value; w k is the process noise, v k is the observation noise.

[0048] Blind zone processing: Rely on encoder displacement when RFID is not triggered; trigger sound and light alarm when cumulative error >100mm.

[0049] 2. The production queue optimization module is used to adjust vehicle scheduling strategies in real time during mixed-flow production on multiple production lines. It includes a dynamic scheduling engine and a conflict resolution unit.

[0050] The dynamic scheduling engine, based on the deep reinforcement learning (DRL) algorithm, receives input data such as MES order priority, equipment status, and workstation load matrix in real time, and outputs queue adjustment instructions including acceleration, deceleration, and path switching. The state space of the reinforcement learning algorithm includes order priority, equipment status, and workstation load matrix, and the action space includes acceleration instructions, deceleration instructions, and path switching instructions. Its reward function is: R = α⋅T 节省 −β⋅C 拥堵 ; Among them, α, β are weight coefficients, T 节省 Save time for estimation; C 拥堵 Penalty for path conflict.

[0051] The conflict resolution unit allocates spatiotemporal windows using a spatiotemporal resource reservation model for the intersection area, and triggers rescheduling when a path conflict (overlapping paths and overlapping time windows) is detected.

[0052] 3. The MES integration module is used to achieve two-way real-time interaction between the system and MES, including the OPC UA protocol interface and visual dashboard, including:

[0053] The OPC UA protocol interface enables real-time data exchange with MES systems, including workstation status and equipment fault alarms. Seamless integration with mainstream MES systems through the OPC UA protocol enhances system compatibility and reduces deployment costs.

[0054] The visual dashboard displays vehicle trajectories, dispatch status, and equipment efficiency metrics, including a real-time track map and a dispatch efficiency dashboard. The real-time track map displays color-coded vehicle status (green: normal, yellow: delay >30s, red: fault). The dispatch efficiency dashboard displays information such as CT (cycle time) deviation rate and OEE (overall equipment effectiveness). Its alert rule automatically turns red when the CT deviation rate exceeds 15%.

[0055] Second embodiment

[0056] like Figure 2 As shown, this embodiment provides a vehicle tracking and queue optimization control method for mixed flow of multiple production lines, which is implemented using the precise vehicle tracking and queue optimization control system for mixed flow of multiple production lines described in the first embodiment, and mainly includes the following steps: The vehicle enters the production line: The RFID tag is recognized by the first station reader, and the system binds the vehicle information to the initial position of the encoder. Real-time tracking: The encoder continuously outputs the displacement, the RFID reader corrects the position at the key node, and the fusion algorithm generates real-time coordinates. Queue optimization: The dynamic scheduling engine calculates the optimal queue every 5 seconds and issues instructions (such as acceleration / deceleration, path switching) through the MES. Closed-loop feedback: The MES updates the scheduling parameters according to the execution results of the workstation, forming a "perception-decision-execution-feedback" closed loop. Specifically including:

[0057] Step S1. Vehicle entry and binding on the production line: When the vehicle carrier enters the first station, the vehicle identification code, order number and model code in the high-temperature resistant ceramic package RFID tag installed on the bottom of the carrier are read by the RFID reader with a multi-antenna array. If multiple tags are detected at the same time, the time-slot ALOHA algorithm is used for signal separation.

[0058] Step S2. The encoder (resolution: 0.1mm / pulse) records the initial displacement value of the conveyor line. The system binds the vehicle information to the initial position of the encoder and establishes a mapping relationship between the vehicle information and the initial displacement value. If the RFID reader does not recognize it, an audible and visual alarm is triggered and the conveyor line is paused, waiting for manual intervention.

[0059] Step S3. Continuous Displacement Acquisition: During vehicle transport, the primary absolute encoder outputs pulse signals in real time, and the displacement increment is calculated according to the following formula: ΔS = N × K, where K is 0.1 mm and N is the pulse count. A redundant incremental encoder automatically switches to the primary encoder if it fails.

[0060] Step S4. Position correction and data fusion: When the vehicle passes through key nodes (station entrance / exit, diversion point), the absolute position of the vehicle is obtained through the RFID reader. Combined with the encoder displacement data, the position is fused and calculated using the Kalman filter algorithm to generate the real-time coordinates of the vehicle and output the vehicle's continuous trajectory (positioning accuracy is ±50mm). When the RFID is not triggered and the encoder cumulative error exceeds 100mm, an audible and visual alarm is triggered. The state equation of the Kalman filter algorithm is: k =x k-1 +ΔS+w k , z k =Hx k +v k ; Among them, w k is the process noise, v k is the observation noise, ΔS is the displacement increment, x k is the displacement estimate, z k is the RFID observation value.

[0061] Step S5. Dynamic optimization of the production queue: Input the real-time vehicle position, order priority, equipment status, and workstation load matrix into the reinforcement learning algorithm scheduling engine to calculate acceleration, deceleration, or path switching instructions; the reward function of the reinforcement learning algorithm is: R = α⋅T 节省 −β⋅C 拥堵 ; Among them, α, β are weight coefficients, T 节省 Save time for estimation; C 拥堵 Penalty for path conflict.

[0062] Step S6. Conflict resolution and instruction issuance: Before a vehicle enters the multi-line intersection area, the system predicts the entry time and path occupancy of each vehicle based on the spatiotemporal resource reservation model, and detects whether there is a potential conflict of overlapping paths and time windows. If a conflict exists, the spatiotemporal windows are reallocated based on vehicle priority, order urgency, and workstation load, and conflict-free scheduling instructions such as acceleration, deceleration, waiting, or path switching are generated. These instructions are issued to the MES system via the OPC UA protocol every 5 seconds. If a conflict is detected again during operation, rescheduling is immediately triggered and the instructions are updated to ensure safe and efficient passage through the intersection area.

[0063] Step S7. Closed-loop feedback and parameter adjustment: After the dispatch instruction is executed, the PLC (Programmable Logic Controller) returns the instruction execution status information to the system, including the flag of successful execution or failure, and the corresponding equipment fault code, such as "E102" indicating conveyor line overload. At the same time, the scanner at the relevant station scans the vehicle to obtain the actual arrival time T actual and the vehicle planned arrival time T in the scheduling plan plan Compare the actual arrival time and the planned arrival time to form the arrival time deviation information. According to the difference between the actual arrival time and the planned arrival time, the system dynamically adjusts the reward function weight in the reinforcement learning scheduling model to optimize the subsequent scheduling strategy. The adjustment formula is: β new =β old +0.01×(T actual −T plan ), where β old is the penalty weight before adjustment, β new is the adjusted penalty weight. When the actual arrival time is later than the planned time, the system increases the penalty weight appropriately to increase sensitivity to delays; otherwise, the penalty weight is reduced.

[0064] When the PLC returns a fault code indicating a fault, the system immediately triggers a fault response mechanism. First, the system automatically blocks the faulty workstation, preventing subsequent vehicles from entering the faulty area. Then, based on pre-defined detour plans, it activates alternate routes, such as activating an AGV (Automated Guided Vehicle) backup channel, to bypass the faulty workstation, ensuring uninterrupted production flow.

[0065] The above describes the specific embodiments of the present invention. Based on the above description, relevant personnel can make various changes and modifications without departing from the scope of the technical concept of this invention.

Claims

1. A vehicle tracking and queue optimization control system for mixed production lines, characterized in that: Including real-time workstation tracking module, production queue optimization module and MES integration module: The real-time workstation tracking module includes an RFID positioning unit, an encoder positioning unit, and a data fusion unit. The RFID positioning unit is used to obtain vehicle identification information and RFID positioning data through a high-temperature ceramic packaged RFID tag and a multi-antenna array reader, and adopts dynamic frequency band switching to achieve anti-interference positioning. The encoder positioning unit includes a main encoder and a redundant incremental encoder, which are used to measure the displacement data of the conveyor line and output the displacement increment ΔS. The data fusion unit is used to fuse the RFID positioning data and the encoder displacement data based on the Kalman filter algorithm to generate the real-time position coordinates of the vehicle. The production queue optimization module includes a dynamic scheduling engine and a conflict resolution unit. The dynamic scheduling engine uses a reinforcement learning algorithm to calculate queue adjustment instructions using order priority, equipment status, and workstation load data provided by the MES. The conflict resolution unit allocates spatiotemporal windows to the intersection of multiple production lines based on a spatiotemporal resource reservation model to avoid vehicle path conflicts. The MES integration module includes an OPC UA protocol interface and a visual dashboard. The OPC UA protocol interface is used to achieve real-time data interaction with the MES system; the visual dashboard is used to display vehicle trajectory, scheduling status and equipment efficiency indicators.

2. The vehicle tracking and queue optimization control system for mixed production lines according to claim 1 is characterized in that: The RFID tag is embedded and installed at the bottom of the vehicle at 15±5cm from the ground. The tag has a storage capacity of 1KB and stores information including the vehicle identification code, order number and vehicle model code.

3. The vehicle tracking and queue optimization control system for mixed production lines according to claim 1, characterized in that: The multi-antenna array reader deploys four groups of directional antennas at each key node with a horizontal coverage angle of 90°, and activates each antenna group in turn according to a polling cycle of 10ms.

4. The vehicle tracking and queue optimization control system for mixed production lines according to claim 1, characterized in that: The dynamic frequency band switching detects ambient noise based on the received signal strength indication (RSSI) and automatically switches between the 865-868MHz and 902-928MHz frequency bands.

5. The vehicle tracking and queue optimization control system for mixed production lines according to claim 1, characterized in that: The main encoder in the encoder positioning unit has a resolution of 0.1 mm / pulse and a sampling frequency of 1 kHz. The redundant incremental encoder automatically switches to work when the main encoder fails.

6. The vehicle tracking and queue optimization control system for mixed production lines according to claim 1, characterized in that: The state equation of the Kalman filter is: k =x k-1 +ΔS+w k , z k =Hx k +v k ; Among them, w k is the process noise, v k is the observation noise, ΔS is the displacement increment, x k is the displacement estimate, z k is the RFID observation value.

7. The vehicle tracking and queue optimization control system for mixed production lines according to claim 1, characterized in that: The state space of the reinforcement learning algorithm includes order priority, equipment status and workstation load matrix, and the action space includes acceleration instructions, deceleration instructions and path switching instructions. The reward function is: R = α⋅T 节省 −β⋅C 拥堵 ; Among them, α, β are weight coefficients, T 节省 Save time for estimation; C 拥堵 Penalty for path conflict.

8. The vehicle tracking and queue optimization control system for mixed production lines according to claim 1, characterized in that: The visual dashboard includes: a real-time trajectory map, which is displayed in green, yellow and red according to the vehicle status; a scheduling efficiency panel, which displays the cycle time deviation rate and equipment overall efficiency, and a red warning when the deviation rate exceeds 15%.

9. A method for tracking and optimizing the platoon of vehicles with mixed production lines, implemented by using the precise tracking and optimizing the platoon of vehicles with mixed production lines as claimed in any one of claims 1 to 8, characterized in that: The following steps are involved: Step S1. Vehicle entry and binding: When a vehicle enters the first station, a multi-antenna array RFID reader reads the vehicle identification code, order number, and model code from the high-temperature ceramic packaged RFID tag installed on the bottom of the vehicle, records the initial displacement value, and establishes a mapping relationship between the vehicle information and the initial displacement value; Step S2. Continuous displacement acquisition: During the vehicle transportation process, the main encoder outputs a pulse signal in real time to calculate the displacement increment. The redundant incremental encoder automatically switches to work when the main encoder fails. Step S3. Position correction and data fusion: When the vehicle passes through a key node, the vehicle's absolute position is acquired through an RFID reader. Combined with the encoder displacement data, the position is fused and calculated using a Kalman filter algorithm to generate the vehicle's real-time coordinates. If the RFID is not triggered and the encoder's cumulative error exceeds 100mm, an audible and visual alarm is triggered. Step S4. Dynamic optimization of the production queue: Input the real-time vehicle position, order priority, equipment status, and workstation load matrix into the reinforcement learning scheduling engine to calculate acceleration, deceleration, or path switching instructions; Step S5. Conflict resolution and instruction issuance: In areas where multiple production lines intersect, potential route conflicts are detected based on the spatiotemporal resource reservation model. Spatiotemporal windows are allocated to vehicles, and conflict-free dispatch instructions are generated and issued to the MES system via the OPC UA protocol. Step S6. Closed-loop feedback and parameter adjustment: Collect the instruction execution status and the actual vehicle arrival time, adjust the reward function weight parameters according to the execution results, and automatically block the faulty workstation and enable the preset detour path when a equipment failure is detected.

Images