A safe navigation warehousing method and system based on a laser communication network

Abstract

This invention discloses a secure navigation warehousing method and system based on a laser communication network, belonging to the field of intelligent robot technology. It employs a three-level network initialization mechanism of "master node-slave node-edge node," deploying nodes based on the principle of "adaptive clustering and partitioning networking + dynamic node collaboration," and completing network performance testing through "full-scenario dynamic testing + AI prediction optimization." Each robot establishes an encrypted communication link with the nearest slave node, completing identity authentication and key negotiation. Warehouse task scheduling instructions are encrypted and transmitted to the robots via the master and slave nodes, achieving dynamic allocation in transit. The robots adjust their driving posture and direction accordingly, optimizing their paths. This invention improves warehousing efficiency through efficient task scheduling and precise path optimization, ensures data transmission security through encrypted communication links and identity authentication mechanisms, and ensures network reliability through real-time monitoring of communication link status and dynamic replacement of interrupted nodes.

Description

Technical Field

[0001] This invention relates to the field of intelligent robot technology, and in particular to a safe navigation warehousing method and system based on laser communication networks. Background Technology

[0002] In the field of smart warehousing, the widespread application of automated equipment such as unmanned forklifts, AGVs and AMR robots has greatly improved the operational efficiency of factories and warehouses, and solved problems such as labor shortages, high labor costs, safety hazards and low work efficiency.

[0003] Referring to patent CN107422735A, a hybrid navigation method combining laser and visual features for trackless AGVs is proposed. This method combines the advantages of laser navigation and visual feature recognition (such as QR code recognition), solving the path guidance problem of AGVs in corridor environments. It features simple construction, low maintenance costs, and no impact on the usage environment. However, in complex warehouse environments, such as those with large variations in ambient light and uneven ground, the accuracy and reliability of navigation may decrease, and visual navigation has limitations in terms of rapid response and accurate environmental recognition.

[0004] Another reference patent, CN214360153U, provides a navigation and positioning system for AMR and AGV forklifts. This system achieves seamless connection between computer signals and the forklift control system through seamless intelligent linking technology for control signals. It is simple to deploy, suitable for various lighting and ground conditions, requires minimal maintenance, and supports 24-hour operation in the dark. However, most of these navigation methods rely on wireless... Network communication poses security risks to information transmission and cannot be used in work environments with special information security requirements.

[0005] In practical applications, especially in complex warehousing environments, the two navigation methods mentioned above still have certain limitations in terms of environmental adaptability, navigation accuracy and real-time performance, and communication security, making it difficult to meet the needs of certain special scenarios.

[0006] Therefore, there is an urgent need for a new navigation method that can provide a secure communication network while ensuring navigation accuracy and real-time performance, so as to adapt to a wider range of application scenarios, especially in situations where information security is a strict requirement.

[0007] To address the above issues, this application proposes a safe navigation warehousing method and system based on a laser communication network, aiming to provide a safe and efficient navigation solution to meet the needs of complex warehousing environments and special working conditions. This solution combines the security of laser communication networks, the high precision and real-time performance of laser navigation, and the flexibility of intelligent automated equipment, providing a more comprehensive, reliable, and safe solution for intelligent warehousing. Summary of the Invention

[0008] This invention provides a secure navigation warehousing method and system based on laser communication networks, which solves the problems of insufficient security and efficiency in existing technologies.

[0009] The present invention solves the technical problem through the following technical solution:

[0010] A safe navigation warehousing method based on a laser communication network employs a safe navigation warehousing system. The system includes several shelves and several robots. Aisles are provided between adjacent shelves arranged side-by-side, and each robot moves within the aisles. The safe navigation warehousing method based on a laser communication network includes the following steps:

[0011] (1) Construct a laser communication network using a three-level networking initialization mechanism of "master node - slave node - edge node"; deploy slave nodes and edge nodes in the warehouse environment based on the principle of "adaptive clustering partition networking + dynamic node collaboration"; conduct networking performance testing of the completed laser communication network using "full-scenario dynamic testing + AI prediction optimization" until the preset communication performance indicators are met.

[0012] (2) Each robot establishes an encrypted communication link with the nearest slave node to complete identity authentication and key negotiation; the scheduling instructions for the warehousing task are encrypted and transmitted to the robot through the master node and slave node, and are dynamically allocated. The robot adjusts its driving posture and direction accordingly based on the scheduling instructions, guide magnetic strip information and QR code identification information to achieve path optimization.

[0013] (3) Monitor the bit error rate, delay and working status of each node in real time of the laser communication link, and dynamically fill in interrupted nodes.

[0014] Furthermore, in step (1), the three-level network initialization mechanism of "master node-slave node-edge node" specifically includes the following steps:

[0015] (11) Deploy one master node in the warehouse environment, wherein the master node performs data forwarding, communication status monitoring, slave node collaborative scheduling, channel resource allocation and fault alarm;

[0016] (12) Deploy at least one slave node in each zone within the warehouse environment. Each slave node provides laser signal coverage to the warehouse area of ​​its assigned zone and establishes a Mesh link with adjacent slave nodes to achieve full laser communication coverage of the warehouse area.

[0017] (13) Deploy edge nodes in the gaps between shelves and at corners in the warehouse environment. The edge nodes perform edge blind spot filling of the laser communication network.

[0018] Furthermore, in step (1), the principle of "adaptive clustering partitioning networking + dynamic node collaboration" specifically includes:

[0019] (1-1) Based on the working density of each robot and the occlusion of the shelf, the communication partition is automatically divided by K-means clustering algorithm. At least one slave node is deployed in each partition. The partition boundary adopts an overlapping coverage design with an overlap width of ≥5m.

[0020] (1-2) Each slave node dynamically adjusts the horizontal and vertical divergence angles based on the occlusion data; the edge node establishes a bidirectional Mesh link with the slave node of its partition, forming a dual redundancy architecture of "partition main coverage + edge blind spot filling";

[0021] (1-3) A new inter-node collaborative perception protocol is added, in which slave nodes of adjacent partitions share signal strength and load status data in real time. When the robot density in a certain partition increases suddenly or temporary occlusion occurs, the slave nodes of adjacent partitions are automatically scheduled to expand the coverage area and the edge nodes to increase the transmission power, so as to ensure that there are no communication blind spots or weak signal points in the robot movement area.

[0022] Furthermore, in step (2), the dynamic allocation of the warehousing tasks and the path optimization of the robot are realized based on the improved genetic algorithm. The improved genetic algorithm includes an upper-level genetic algorithm and a lower-level ant colony algorithm, specifically:

[0023] (2-1) The upper-level genetic algorithm outputs the globally optimal robot-warehouse task allocation scheme and initial path node sequence, wherein the upper-level genetic algorithm includes:

[0024] A two-dimensional encoding is used: one dimension is the mapping between robot number and task number, and the two dimensions are the path node sequence.

[0025] A multi-objective weighted fitness function is adopted, which includes the total length of the robot's planned path and its corresponding weight, the path conflict probability and its corresponding weight, the total time to complete the task and its corresponding weight, and the stability of the laser communication signal and its corresponding weight.

[0026] The selection operator uses a combination of "elite retention + roulette wheel selection", retaining the top 10% of the best individuals, and selecting the remaining individuals based on fitness roulette wheel selection;

[0027] The crossover operator uses "segmented crossover," which only crosses path node sequence segments, preserving the stability of the AGV-task mapping.

[0028] The mutation operator introduces "laser communication constrained mutation", which prioritizes path nodes with laser signal intensity ≥ -50dBm during mutation;

[0029] The genetic algorithm iterates to output the globally optimal robot-warehouse task allocation scheme and the initial path node sequence;

[0030] (2-2) Local path fine-tuning and real-time conflict avoidance are achieved through the lower-level ant colony algorithm.

[0031] Furthermore, the upper-level genetic algorithm described in step (2-2) includes:

[0032] We introduce "laser communication weighted pheromone," where pheromone concentration is positively correlated with laser signal intensity and negatively correlated with path congestion.

[0033] The heuristic function introduces a "positioning accuracy constraint," prioritizing the selection of path nodes within the precisely positioned area provided on-site.

[0034] The system collects the position and speed information of all robots in real time. When the distance between the intersection points of the paths of two robots and the meeting time are less than a preset threshold, the system temporarily adjusts the local path of the robot with lower priority to avoid conflicts in real time.

[0035] Furthermore, the multi-objective weighted fitness function is: Where L is the total length of the robot's planned path, with a weight w1=0.4, C is the path conflict probability, with a weight w2=0.3, T is the total time taken to complete the warehousing task, with a weight w3=0.2, and S is the stability of the laser communication signal, with a weight w4=0.1.

[0036] Furthermore, the weighted pheromone for laser communication is Where ρ is the pheromone evaporation coefficient. Let α be the pheromone increment of the robot from node i to j during this iteration, α be the laser signal weighting coefficient, Sij be the laser signal intensity from node i to j, and Dij be the path congestion degree from node i to j. and These are the weighted pheromones of laser communication between the robot and node i and j during the next iteration and the current iteration, respectively.

[0037] Furthermore, the heuristic function is ,in: Let be the Euclidean distance from node i to j, and β be the positioning accuracy weight. Let be the positioning accuracy from node i to j.

[0038] A laser communication network-based safe navigation warehouse system for performing the above method includes a host computer, a three-level distributed wireless laser communication unit, a guide magnetic strip, a QR code label, and at least one robot.

[0039] The host computer monitors the signal strength, signal occupancy rate, and data transmission rate of the laser communication network in real time, and supports fault alarms and automatic switching of each node;

[0040] The three-level distributed wireless laser communication unit includes a master node, at least one slave node, and at least one edge node. The master node is connected to a host computer, and the deployment locations of the slave nodes and edge nodes meet the requirements for full laser communication coverage and edge blind spot filling in the warehouse area.

[0041] The guide magnetic strip is set on the centerline of each tunnel, and the guide magnetic strip provides a path reference for each robot to assist in stabilizing the tracking.

[0042] The QR code markings are laid on the path of the guide magnetic strips to provide accurate positioning information for each robot;

[0043] Each robot integrates the communication link provided by the three-level distributed wireless laser communication unit, the magnetic strip position signal provided by the guide magnetic strip, and the position coordinates built into the QR code to perform warehousing tasks.

[0044] The advantages and effects of this invention are:

[0045] Efficient task scheduling and precise path optimization reduce robot travel time and path length, improving the processing speed and efficiency of warehousing tasks. A dynamic fault-filling mechanism ensures the continuity and stability of the communication network, preventing robot downtime and task delays due to communication failures, further improving warehousing efficiency. Encrypted communication links and authentication mechanisms guarantee data transmission security, preventing data leakage and unauthorized access, and enhancing the information security level of the warehousing system. Real-time monitoring of the error rate, latency, and working status of each node in the laser communication link, along with dynamic fault-filling, ensures the reliability of the communication network, reducing the risk of robot collisions and cargo damage due to communication failures, and improving the physical security level of the warehousing system. Attached Figure Description

[0046] Figure 1 This is a layout diagram of the present invention in a warehouse environment.

[0047] In the diagram: 1-13 are wireless laser communication modules; 14 and 25 are robots; 15 is a guide magnetic strip; 16 is a QR code label; 17-24 are shelves in the warehouse environment; 26, 27, and 28 are charging piles. Detailed Implementation

[0048] The present invention will be further described below with reference to the embodiments, but the present invention is not limited to these embodiments.

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] The robot referred to in this invention ( Figure 1 Items 14 and 25 in the diagram refer to AGV (Automated Guided Vehicle), a type of wheeled mobile robot with automatic navigation, driving, and loading / unloading capabilities. In this invention, the AGV is responsible for cargo handling and includes a navigation module and a communication module. The navigation module integrates a magnetic stripe detection sensor, a high-definition industrial camera (resolution ≥ 2 megapixels), and a laser signal receiving module to achieve multi-source fusion navigation combining magnetic stripe tracking, QR code recognition, and laser communication. The communication module has one wireless laser communication module at the front and one at the rear of the robot. The gimbal supports ±30° pitch and ±180° horizontal coverage. A PID attitude adjustment algorithm tracks the signal from the fixed-point laser communication unit in real time to ensure stable communication. The two wireless laser communication modules are connected to the AGV's control system, providing dual-link redundancy and backup to prevent operational interruptions due to a single communication unit failure. In this invention, the AGV intelligent robot serves as the terminal node, and its dual-axis gimbal-type laser communication unit serves as the terminal of the laser communication network. It is responsible for bidirectional data transmission with the slave node, uploading information such as the AGV intelligent robot's position, speed, and load status, and receiving scheduling instructions and path planning data issued by the host computer.

[0051] like Figure 1 As shown, a safe navigation warehousing system based on a laser communication network includes several shelves ( Figure 1 (17-24) and several robots ( Figure 1 14 and 25 in the middle), two adjacent shelves placed side by side ( Figure 1 The system includes a tunnel between 17 and 24, where each robot moves. It also includes a host computer, a three-level distributed wireless laser communication unit, a guide magnetic strip 15, and a QR code label 16.

[0052] The host computer monitors the signal strength, signal occupancy rate, and data transmission rate of the laser communication network in real time, and supports fault alarms and automatic switching of each node;

[0053] The three-level distributed wireless laser communication unit includes a master node, at least one slave node, and at least one edge node. The master node is connected to a host computer, and the deployment locations of the slave nodes and edge nodes meet the requirements for full laser communication coverage and edge blind spot filling in the warehouse area.

[0054] The guide magnetic strip 15 is set on the centerline of each tunnel, and the guide magnetic strip 15 provides a path reference for each robot and assists in stabilizing the tracking.

[0055] The QR code identifier 16 is laid on the path of the guide magnetic strip 15 to provide accurate positioning information for each robot;

[0056] Each robot integrates the communication link provided by the three-level distributed wireless laser communication unit, the magnetic strip position signal provided by the guide magnetic strip 15, and the position coordinates built into the QR code identifier 16 to perform warehousing tasks. The layout of this invention in a warehousing environment is shown below. Figure 1 As shown.

[0057] A secure navigation and warehousing method based on a laser communication network includes the following steps:

[0058] (1) Construct a laser communication network using a three-level networking initialization mechanism of "master node - slave node - edge node"; deploy slave nodes and edge nodes in the warehouse environment based on the principle of "adaptive clustering partition networking + dynamic node collaboration"; conduct networking performance testing of the completed laser communication network using "full-scenario dynamic testing + AI prediction optimization" until the preset communication performance indicators are met.

[0059] (2) Each robot establishes an encrypted communication link with the nearest slave node to complete identity authentication and key negotiation; the scheduling instructions for the warehousing task are encrypted and transmitted to the robot through the master node and slave node, and are dynamically allocated. The robot adjusts its driving posture and direction accordingly based on the scheduling instructions, the information of the guide magnetic strip 15, and the information of the QR code identifier 16 to achieve path optimization.

[0060] (3) Monitor the bit error rate, delay and working status of each node in real time of the laser communication link, and dynamically fill in interrupted nodes.

[0061] Furthermore, in step (1), the three-level network initialization mechanism of "master node-slave node-edge node" specifically includes the following steps:

[0062] (11) Deploy one master node in the warehouse environment, wherein the master node performs data forwarding, communication status monitoring, slave node collaborative scheduling, channel resource allocation and fault alarm;

[0063] (12) Deploy at least one slave node in each zone within the warehouse environment. Each slave node provides laser signal coverage to the warehouse area of ​​its assigned zone and establishes a Mesh link with adjacent slave nodes to achieve full laser communication coverage of the warehouse area.

[0064] (13) Shelves in the storage environment ( Figure 1 Edge nodes are deployed at gaps and turns in sections 17-24 of the laser communication network, and these edge nodes perform edge blind spot filling.

[0065] The master node, slave nodes, edge nodes, and each robot all possess laser signal transmission, reception, decoding, and data forwarding capabilities. Core technical parameters include: laser wavelength 1550nm, communication bandwidth 10Gbps, transmission distance 0~300m, and divergence angle 0.1°~1° (adjustable). The master node can be... Figure 1 Any one of the wireless laser communication modules in 1-13 (such as Figure 1 In section 1), the master node is directly connected to the host computer and is responsible for coordinating the operation of the entire laser communication network. There are several slave nodes (e.g., ...). Figure 1 In addition to certain wireless laser communication modules of the master node, such as Figure 1 Several of the nodes (2-13 in the text) are fixedly installed on the surrounding walls of the warehouse environment, using the principle of "partition coverage + seamless connection". They support the establishment of Mesh links with adjacent slave nodes to achieve redundant signal coverage. When a slave node fails, communication is automatically switched to an adjacent slave node to avoid communication blind spots, forming a full coverage network of laser signals in the warehouse environment. Edge nodes support the establishment of bidirectional Mesh links with slave nodes in their respective partitions, forming a dual redundancy architecture of "partition main coverage + edge blind spot filling".

[0066] This invention innovatively designs a three-level network initialization mechanism of "master node-slave node-edge node," which differs from the traditional single registration mode. The master control node presets network baseline parameters and issues network commands via laser broadcast. When a slave node joins, it first completes its own position calibration through laser ranging, then completes link negotiation with adjacent slave nodes, automatically assigns a unique channel (avoiding co-channel interference), and completes node ID registration and network initialization. The backup redundant node adopts a "dynamic sleep-wake-up" mechanism. It is normally in a low-power sleep state and only establishes a high-frequency heartbeat connection (heartbeat interval of 30ms) with the two nearest slave nodes to monitor their communication quality, load status, and link stability in real time. A new node fault prediction mechanism is added. The master node predicts node faults 500ms in advance by analyzing parameters such as the signal attenuation rate and data transmission delay of slave nodes, triggering the backup node to wake up in advance and complete link pre-connection, ensuring that the fault switching response time is ≤20ms, which is far superior to the traditional switching efficiency and significantly improves network reliability. This invention adopts a three-level networking architecture of "master node-slave node-edge node". The master node is responsible for global scheduling, the slave node provides partition coverage and supports Mesh links, and the edge node provides edge blind spot coverage, forming a dual redundancy architecture of "partition master coverage + edge blind spot coverage", which solves the signal blind spot problem in traditional warehouse communication.

[0067] Furthermore, in step (1), the principle of "adaptive clustering partitioning networking + dynamic node collaboration" specifically includes:

[0068] (1-1) Based on the operating density of each robot and the shelf ( Figure 1 Based on the occlusion conditions in sections 17-24, communication partitions are automatically divided using the K-means clustering algorithm. At least one slave node is deployed in each partition, and the partition boundaries adopt an overlapping coverage design with an overlap width of ≥5m.

[0069] (1-2) Each slave node dynamically adjusts the horizontal and vertical divergence angles based on the occlusion data; the edge node establishes a bidirectional Mesh link with the slave node of its partition, forming a dual redundancy architecture of "partition main coverage + edge blind spot filling";

[0070] (1-3) A new inter-node collaborative perception protocol is added, in which slave nodes of adjacent partitions share signal strength and load status data in real time. When the robot density in a certain partition increases suddenly or temporary occlusion occurs, the slave nodes of adjacent partitions are automatically scheduled to expand the coverage area and the edge nodes to increase the transmission power, so as to ensure that there are no communication blind spots or weak signal points in the robot movement area.

[0071] This invention innovatively employs an "adaptive clustering partitioning network + dynamic node collaboration" principle to deploy slave nodes and edge nodes, overcoming the limitations of traditional fixed partitioning. In step (1-1), each partition deploys 1-2 slave nodes. In step (1-2), each wireless laser communication module ( Figure 1 (2-13) is equipped with a laser angle adaptive adjustment submodule, which, combined with the occlusion data of the laser radar scanning, can dynamically adjust the horizontal divergence angle (adjustable from 90° to 180°) and the vertical divergence angle (adjustable from -45° to 30°).

[0072] The networking performance testing scheme of "full-scenario dynamic testing + AI prediction optimization" in step (3) breaks through the limitations of traditional single-path testing: First, control the AGV to travel along the preset path for the entire process, test the stability of the communication link under different movement speeds (0.1m / s-1.5m / s) and different operating scenarios (empty / full load, straight / turning), and collect parameters such as signal strength, delay, and bit error rate at the same time; Second, simulate sudden scenarios in the warehouse (temporary obstruction, sudden change in the number of concurrent AGVs, node failure) to test the dynamic response capability of the network; Finally, introduce an AI coverage prediction model, train the model based on the test data, predict potential weak signal areas in the warehouse in advance, and automatically output node installation height adjustment suggestions (adjustment range 0.5m-3m), transmission power optimization value (maximum transmission power ≤5mW, in compliance with IEC 60825-1:2014 Class 1 laser safety level) and edge node supplementary deployment positions without manual intervention; Add multi-AGV concurrent communication stress test (concurrent number ≤20 units) to ensure that the link bit error rate is ≤ With a data transmission latency of ≤5ms, it meets the needs of high-concurrency warehouse operations. At the same time, through AI-optimized models, it reduces network energy consumption by more than 15%.

[0073] The host computer is a distributed collaborative scheduling platform, and its core functions include:

[0074] Laser communication network management: Real-time monitoring of the communication status of master nodes, slave nodes and AGV terminal nodes, displaying parameters such as signal strength, channel occupancy rate and data transmission rate of each node, and supporting node fault alarm and automatic switching;

[0075] Navigation parameter configuration: Supports dynamic adjustment of parameters such as magnetic strip tracking threshold, QR code recognition sensitivity, and laser positioning accuracy, which can adapt to the operational needs of different warehouse layouts;

[0076] Multi-AGV collaborative scheduling: Based on an improved genetic algorithm, tasks are dynamically allocated and paths are optimized, supporting concurrent operations of multiple AGVs, real-time detection of AGV position information, avoiding path conflicts, and scheduling response time ≤10ms.

[0077] Furthermore, based on the genetic algorithm, a dynamic hierarchical genetic-ant colony hybrid algorithm (DL-GA-ACO) that integrates real-time sensing of laser communication is proposed. This algorithm combines the global optimization capability of the genetic algorithm and the local path fine-tuning advantage of the ant colony algorithm, and deeply couples the real-time sensing characteristics of the laser communication network, thus solving the core pain points of traditional algorithms such as "static planning, no real-time feedback, and lagging conflict avoidance".

[0078] This invention designs the entire process of collaborative scheduling of multiple AGVs in warehousing, encompassing "task allocation - path planning - conflict avoidance - dynamic adjustment." It consists of an upper-level genetic algorithm (global task allocation + initial path planning) and a lower-level ant colony algorithm (local path fine-tuning + real-time conflict avoidance). The algorithm's dynamic iteration is achieved through real-time data sensing via a laser communication network. The core process is as follows:

[0079] (I) Upper Layer: Genetic Algorithm (GA) – Global Task Allocation and Initial Path Planning

[0080] 1. Innovative Coding Methods

[0081] A two-dimensional coding system is adopted. The first dimension is the mapping between robot number and task number, and the second dimension is the path node sequence. That is, the first dimension is the mapping of "AGV number - task number" (such as the coding segment 01-05 indicating that AGV No. 1 performs the picking task No. 5), and the second dimension is the "path node sequence" (such as the coding segment 17-03-24 indicating that AGV passes through shelf No. 17 → the coverage area of ​​node No. 3 → shelf No. 24). This solves the problem that traditional single-dimensional coding cannot take into account both task allocation and path planning.

[0082] 2. Fitness function optimization

[0083] Breaking away from the limitations of traditional methods that only focus on "path length," a multi-objective weighted fitness function is constructed. This function includes the total length of the robot's planned path and its corresponding weight, the path conflict probability and its corresponding weight, the total task completion time and its corresponding weight, and the stability of the laser communication signal and its corresponding weight. Where L is the total length of the robot's planned path, with a weight w1=0.4; C is the path conflict probability (calculated based on the real-time position of the AGV using laser communication sensing), with a weight w2=0.3; T is the total time (s) to complete the warehousing task, with a weight w3=0.2; S is the stability of the laser communication signal (values ​​from 0 to 1, with 1 being optimal), with a weight w4=0.1. A higher fitness value indicates better path planning, taking into account the four core objectives of "short path, no conflict, low time consumption, and stable communication."

[0084] 3. Improved genetic manipulation

[0085] Selection operator: A combination of "elite retention + roulette wheel selection" is adopted to retain the top 10% of the best individuals, and the remaining individuals are selected by fitness roulette wheel selection to avoid losing the optimal solution;

[0086] Crossover operator: "Segmented crossover" is adopted, which only crosses path node sequence segments, preserving the stability of AGV-task mapping, and the crossover probability is set to 0.8;

[0087] Mutation operator: Introduce "laser communication constraint mutation", and select path nodes with laser signal intensity ≥ -50dBm during mutation to avoid the path falling into the communication blind zone after mutation. The mutation probability is set to 0.05.

[0088] 4. Output results: Through iterative genetic algorithm (50 iterations and 100 population size), the globally optimal robot-warehouse task allocation scheme and initial path node sequence are output.

[0089] (II) Lower Layer: Ant Colony Algorithm (ACO) – Local Path Fine-tuning and Real-time Conflict Avoidance

[0090] Based on the initial path output by the genetic algorithm, combined with the AGV's position, speed, and laser signal status sensed in real time by the laser communication network, local path fine-tuning and real-time conflict avoidance are performed. The core improvements are:

[0091] 1. Innovative Pheromone Update Rule: A "laser communication weighted pheromone" is introduced to ensure that the ant colony prioritizes local paths with stable communication and no congestion. Pheromone concentration is positively correlated with laser signal intensity and negatively correlated with path congestion.

[0092] Where ρ is the pheromone evaporation coefficient (set to 0.1). Let α be the pheromone increment of the robot from node i to j during this iteration, α be the laser signal weighting coefficient (set to 0.6), Sij be the laser signal intensity from node i to j (normalized to 0~1), and Dij be the path congestion degree from node i to j (calculated based on the number of AGVs sensed by laser communication, 0 for no congestion and 1 for complete congestion). and These are the weighted pheromones of laser communication between the robot and node i and j during the next iteration and the current iteration, respectively.

[0093] 2. Heuristic Function Optimization: The heuristic function introduces a "positioning accuracy constraint," prioritizing path nodes within the precisely positioned area provided by the QR code identifier, thereby improving the AGV's driving accuracy.

[0094] The heuristic function is: ,in: Let β be the Euclidean distance from node i to j, and let β be the positioning accuracy weight (set to 0.5). The positioning accuracy from node i to j (normalized to 0~1, where 1 is the QR code identifier).

[0095] 3. Real-time Conflict Avoidance: The system collects the position and speed data of all AGVs in real time via a laser communication network (delay ≤ 10ms). When the path intersection distance between two AGVs is detected to be ≤ 2m (preset threshold) and the estimated encounter time is ≤ 5s (preset threshold), the ant colony algorithm for local replanning is immediately triggered.

[0096] For AGVs with low priority (sorted by task urgency), their local paths are temporarily adjusted (e.g., they detour around adjacent QR code nodes).

[0097] After adjustment, new path instructions are sent via laser communication, with a response time of ≤10ms, to avoid AGV collisions.

[0098] (III) Dynamic Iteration: Real-time Algorithm Update Based on Laser Communication

[0099] The laser communication network collects position, speed, signal strength, and load status data of all AGVs every 100ms, triggering dynamic algorithm iteration.

[0100] 1. If the actual position of the AGV deviates from the planned path by ≤1mm, maintain the current path;

[0101] 2. If the deviation is >1mm and ≤5mm, only local fine-tuning of the lower-level ant colony algorithm is triggered;

[0102] 3. If the deviation is greater than 5mm, or if an obstacle is detected on the path, immediately restart the upper-level genetic algorithm to generate a new global task allocation and initial path, and then fine-tune it using the ant colony algorithm.

[0103] Furthermore, the guide magnetic strip 15 is made of highly wear-resistant polyurethane material (1.5mm thick, 20mm wide), serving to provide a path reference for the AGV intelligent robot and assist in stable tracking. The guide magnetic strip 15 is positioned on the centerline of the AGV robot's preset travel route, with rounded transitions at turns (radius of curvature ≥ 50cm) to prevent centrifugal force deviation during turns. The AGV intelligent robot collects the magnetic strip position signal through a magnetic strip detection sensor (detection accuracy ±0.1mm), combines it with the position deviation data output by the differential drive module, and achieves precise tracking through a PID correction algorithm (proportional coefficient Kp=0.8, integral coefficient Ki=0.2, derivative coefficient Kd=0.1). The correction response time is ≤2ms, ensuring stable operation of the AGV along the magnetic strip path (travel deviation ≤1mm).

[0104] Furthermore, the QR code labels are made of high-temperature resistant and stain-resistant PET material (50mm×50mm in size), laid on the magnetic strip path, with the spacing adjusted according to the warehousing accuracy requirements (standard spacing 1~2m, high-precision area spacing 0.5m). The QR code labels contain embedded location coordinates, shelf number, path direction, and other data. The AGV intelligent robot scans the QR code using its built-in high-definition industrial camera, combined with an image recognition algorithm (recognition latency ≤3ms, recognition success rate ≥99.9%), to extract the QR code location information and achieve high-precision positioning. The visual recognition method based on a vision sensor eliminates error accumulation, allowing the AGV's repeatability accuracy to be controlled to ≤0.5mm in both the x and y directions.

[0105] The working process of this invention includes system deployment, laser networking, navigation operation, and dynamic optimization, specifically as follows:

[0106] (1) System deployment

[0107] Based on 3D modeling data of the warehouse environment, BIM (Building Information Modeling) technology is used to plan the placement of guide magnetic strips 15, QR code labels 16, and laser communication units, ensuring that the magnetic strip paths cover all shelf aisles, picking points, and charging stations (such as...). Figure 1 In areas 26, 27, and 28, and the turning area, the QR code markings and the laser communication unit are not obstructed.

[0108] The magnetic strip laying accuracy is calibrated using a laser rangefinder. The magnetic strip at the turning point adopts an arc transition design with a curvature radius ≥1.5m. After the magnetic strip is laid, it passes the tensile test (no displacement when tensile force ≥50N) and the wear resistance test (no damage after 100,000 cycles of reciprocating friction).

[0109] The coordinates of QR code label 16 are calibrated. The three-dimensional coordinate information of the label is entered through the host computer platform to complete the binding of the label ID and coordinates. After calibration, the repeatability accuracy is ≤ ±0.5mm in the x and y directions and ≤ ±1mm in the z direction.

[0110] (2) Laser networking

[0111] The system divides communication zones and overlaps coverage at the zone boundaries; dynamically adjusts the horizontal and vertical divergence angles; establishes bidirectional Mesh links between edge nodes and slave nodes in their respective zones; slave nodes in adjacent zones share signal strength and load status data; backup redundant nodes adopt a "dynamic sleep-wake-up" mechanism; new nodes adopt a fault prediction mechanism; and network performance is tested using "full-scenario dynamic testing + AI prediction optimization".

[0112] (3) Navigation operation

[0113] After the AGV intelligent robot is powered on, it completes system initialization. Its laser communication terminal automatically scans the surrounding laser communication nodes and establishes an encrypted communication link with the nearest slave node through the secure communication encryption module to complete identity authentication and key negotiation.

[0114] The host computer generates scheduling instructions based on the warehousing task, which are encrypted and transmitted to the AGV through the main control node and the slave node. After the AGV main control unit parses the instructions, it tracks the guide magnetic strip 15 through the magnetic strip detection sensor. The differential drive correction module adjusts the driving posture in real time based on the position deviation data (deviation threshold ≤ 0.2mm).

[0115] When the AGV travels to the location marked by the QR code, the high-definition vision sensor automatically collects the image of the tag, combines it with the attitude data of the inertial measurement unit (IMU), calculates the current precise position through the Kalman filter fusion algorithm, and uploads the position information to the host computer platform in encryption.

[0116] The host computer platform dynamically updates the optimal path based on the real-time location of the AGV, the task progress, and the operating status of other AGVs. It then issues path adjustment commands through the laser communication network to achieve collaborative obstacle avoidance and efficient operation of multiple AGVs.

[0117] (4) Dynamic optimization

[0118] The host computer collaborative scheduling platform monitors the bit error rate, latency and node working status of the laser communication link in real time. When a slave node fails, the backup redundant node automatically switches to fill the gap within 50ms to ensure that the communication link is not interrupted.

[0119] The DL-GA-ACO algorithm based on AGVs optimizes path planning parameters and the transmission power and alignment angle of laser communication nodes by accumulating operational data, thereby reducing energy consumption while improving navigation accuracy and communication stability.

[0120] When the warehouse layout is adjusted, the coordinate information of the QR code identifiers and the networking parameters of the laser communication nodes can be reconfigured through the host computer platform without large-scale changes to the hardware deployment, thus achieving flexible adaptation of the system.

[0121] It should be noted that although the embodiments described above are illustrative, they are not intended to limit the invention. Therefore, the invention is not limited to the specific embodiments described above. Any other embodiments obtained by those skilled in the art under the guidance of this invention without departing from its principles are considered to be within the protection scope of this invention.

Claims

1. A secure navigation warehousing method based on a laser communication network, characterized in that, A safe navigation warehousing system is adopted, which includes several shelves and several robots. Aisles are provided between two adjacent shelves arranged side-by-side, and each robot moves within the aisles. The safe navigation warehousing method based on a laser communication network includes the following steps: (1) Construct a laser communication network using a three-level networking initialization mechanism of "master node - slave node - edge node"; deploy slave nodes and edge nodes in the warehouse environment based on the principle of "adaptive clustering partition networking + dynamic node collaboration"; conduct networking performance testing of the completed laser communication network using "full-scenario dynamic testing + AI prediction optimization" until the preset communication performance indicators are met. (2) Each robot establishes an encrypted communication link with the nearest slave node to complete identity authentication and key negotiation; The scheduling instructions for warehousing tasks are encrypted and transmitted to the robot through the master node and slave node, and are dynamically allocated. The robot adjusts its driving posture and direction accordingly based on the scheduling instructions, the information of the guide magnetic strip (15), and the information of the QR code (16) to achieve path optimization. (3) Monitor the bit error rate, delay and working status of each node in real time of the laser communication link, and dynamically fill in interrupted nodes.

2. The secure navigation and warehousing method based on a laser communication network according to claim 1, characterized in that: In step (1), the three-level network initialization mechanism of "master node-slave node-edge node" specifically includes the following steps: (11) Deploy one master node in the warehouse environment, wherein the master node performs data forwarding, communication status monitoring, slave node collaborative scheduling, channel resource allocation and fault alarm; (12) Deploy at least one slave node in each zone within the warehouse environment. Each slave node provides laser signal coverage to the warehouse area of ​​its assigned zone and establishes a Mesh link with adjacent slave nodes to achieve full laser communication coverage of the warehouse area. (13) Deploy edge nodes in the gaps between shelves and at corners in the warehouse environment. The edge nodes perform edge blind spot filling of the laser communication network.

3. The secure navigation and warehousing method based on a laser communication network according to claim 1, characterized in that: In step (1), the principle of "adaptive clustering partitioning networking + dynamic node collaboration" specifically includes: (1-1) Based on the working density of each robot and the occlusion of the shelf, the communication partition is automatically divided by K-means clustering algorithm. At least one slave node is deployed in each partition. The partition boundary adopts an overlapping coverage design with an overlap width of ≥5m. (1-2) Each slave node dynamically adjusts the horizontal and vertical divergence angles based on the occlusion data; the edge node establishes a bidirectional Mesh link with the slave node of its partition, forming a dual redundancy architecture of "partition main coverage + edge blind spot filling"; (1-3) A new inter-node collaborative sensing protocol is added, in which slave nodes of adjacent partitions share signal strength and load status data in real time. When the robot density of a certain partition increases suddenly or temporary occlusion occurs, the slave nodes of adjacent partitions are automatically scheduled to expand the coverage area and the edge nodes to increase the transmission power.

4. The secure navigation and warehousing method based on a laser communication network according to claim 1, characterized in that: In step (2), the dynamic allocation of the warehousing tasks and the path optimization of the robot are realized based on the improved genetic algorithm. The improved genetic algorithm includes an upper-level genetic algorithm and a lower-level ant colony algorithm, specifically: (2-1) The upper-level genetic algorithm includes: A two-dimensional encoding is used: one dimension is the mapping between robot number and task number, and the two dimensions are the path node sequence. A multi-objective weighted fitness function is adopted, which includes the total length of the robot's planned path and its corresponding weight, the path conflict probability and its corresponding weight, the total time to complete the task and its corresponding weight, and the stability of the laser communication signal and its corresponding weight. The selection operator uses a combination of "elite retention + roulette wheel selection", retaining the top 10% of the best individuals, and selecting the remaining individuals based on fitness roulette wheel selection; The crossover operator uses "segmented crossover", which only crosses path node sequence segments, preserving the stability of the AGV-task mapping; The mutation operator introduces "laser communication constraint mutation", which prioritizes path nodes with laser signal intensity ≥ -50dBm during mutation; The genetic algorithm iterates to output the globally optimal robot-warehouse task allocation scheme and the initial path node sequence; (2-2) Local path fine-tuning and real-time conflict avoidance are achieved through the lower-level ant colony algorithm.

5. A secure navigation and warehousing method based on a laser communication network according to claim 4, characterized in that: The upper-level genetic algorithm described in step (2-2) includes: We introduce "laser communication weighted pheromone," where pheromone concentration is positively correlated with laser signal intensity and negatively correlated with path congestion. The heuristic function introduces a "positioning accuracy constraint," prioritizing the selection of path nodes within the precise positioning area provided on-site. The system collects the position and speed information of all robots in real time. When the distance between the intersection points of the paths of two robots and the meeting time are less than a preset threshold, the system temporarily adjusts the local path of the robot with lower priority to avoid conflicts in real time.

6. The secure navigation and warehousing method based on a laser communication network according to claim 4, characterized in that: The multi-objective weighted fitness function is: Where L is the total length of the robot's planned path, with a weight w1=0.4, C is the path conflict probability, with a weight w2=0.3, T is the total time taken to complete the warehousing task, with a weight w3=0.2, and S is the stability of the laser communication signal, with a weight w4=0.

1.

7. A secure navigation and warehousing method based on a laser communication network according to claim 5, characterized in that: The weighted pheromone for laser communication is Where ρ is the pheromone evaporation coefficient. Let α be the pheromone increment of the robot from node i to j during this iteration, α be the laser signal weighting coefficient, Sij be the laser signal intensity from node i to j, and Dij be the path congestion degree from node i to j. and These are the weighted pheromones of laser communication between the robot and node i and j during the next iteration and the current iteration, respectively.

8. A secure navigation and warehousing method based on a laser communication network according to claim 5, characterized in that: The heuristic function is where: is the Euclidean distance between nodes i and j, and β is a positioning accuracy weight, is the positioning accuracy between nodes i and j.

9. A secure navigation warehousing system based on a laser communication network, used to execute the secure navigation warehousing method based on a laser communication network as described in any one of claims 1-8, characterized in that: Includes a host computer, a three-level distributed wireless laser communication unit, a guide magnetic strip (15), and a QR code label (16). The host computer monitors the signal strength, signal occupancy rate, and data transmission rate of the laser communication network in real time, and supports fault alarms and automatic switching of each node; The three-level distributed wireless laser communication unit includes a master node, at least one slave node, and at least one edge node. The master node is connected to a host computer, and the deployment locations of the slave nodes and edge nodes meet the requirements for full laser communication coverage and edge blind spot filling in the warehouse area. The guide magnetic strip (15) is set on the center line of each tunnel. The guide magnetic strip (15) provides a path reference for each robot and assists in stabilizing the tracking. The QR code (16) is laid on the path of the guide magnetic strip (15) to provide positioning information for each robot; Each robot integrates the communication link provided by the three-level distributed wireless laser communication unit, the magnetic strip position signal provided by the guide magnetic strip (15), and the position coordinates built into the QR code identifier (16) to perform warehousing tasks.

Images