Intelligent sorting and recycling system for circulation box
Through the improved LoRaWAN positioning technology and multi-layer perceptron algorithm, combined with the RFID module, the problem of low positioning accuracy of the loop box in a variable environment is solved, and more efficient channel perception and communication real-time performance is achieved, ensuring the loop box Accurate sorting and recycling.
Patent Information
- Application Number
- CN202410914140.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-09
- Publication Date
- 2024-11-08
AI Technical Summary
In the logistics industry, the loss and omission of loop boxes are difficult to achieve accurate positioning in a changing mobile environment due to the weak channel change perception and real-time performance when applying LoRaWAN remote positioning.
Using improved LoRaWAN positioning technology combined with multi-layer perceptron algorithm and radio frequency identification module, communication parameters are adjusted through multi-channel monitoring and adaptive adjustment, channel quality and node motion speed are monitored in real time, modulation parameters are dynamically adjusted, and channel perception capabilities and communication real-time performance are improved. , and use the radio frequency identification module to accurately identify the cycle box information.
It improves channel change perception capability and communication quality, enhances positioning accuracy, reduces delays, improves data transmission rate and operating efficiency, and ensures effective sorting and recycling in a variable environment.
Smart Images

Figure CN118904748A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of circulation box positioning, and in particular to an intelligent sorting and recycling system for circulation boxes. Background Art
[0002] Circulation boxes are low-cost, sturdy and durable, and are widely used in the logistics industry. However, they are often lost or missed during application. Many logistics companies use LoRaWAN remote positioning to solve the problem of loss of circulation boxes. However, since LoRaWAN remote positioning is conventionally used for static positioning, its channel change perception ability, LoRa communication real-time performance, and modulation parameter adjustment based on the node's movement speed are relatively weak. It often fails to achieve the expected results and positioning accuracy in mobile scenarios with changing environments. Summary of the invention
[0003] In order to solve the above problems, the present invention provides an intelligent sorting and recycling system for circulation boxes, which can improve the perception ability of link status, the reaction speed to changes in channel conditions, and the dynamic balance between data rate and data packet acquisition rate. In response to the changeable mobile environment, the sorting and recycling of circulation boxes can also achieve the expected effect and positioning accuracy.
[0004] To achieve the above object, the technical solution adopted by the present invention is:
[0005] An intelligent sorting and recycling system for a circulation box, comprising: a positioning module, an intelligent analysis module and a radio frequency identification module connected in sequence by wireless communication, wherein:
[0006] The positioning module is used to remotely locate the circulation box through the improved LoRaWAN positioning technology;
[0007] The intelligent analysis module is used to evaluate the data of existing recycling boxes using an improved multi-layer perceptron algorithm to determine the future demand for recycling boxes;
[0008] The radio frequency identification module is used to identify the circulation box through a passive RFID tag to determine the specific information of the circulation box.
[0009] Furthermore, the remote positioning of the circulation box by using the improved LoRaWAN positioning technology specifically includes:
[0010] S1: Each node in multiple channels monitors RSSI and SNR, records and analyzes historical data, adaptively adjusts communication parameters, and multiple nodes share channel status information and combine the geographical location information of each node. Machine learning algorithms are used to analyze and predict, and channel change information is fed back to the network server.
[0011] S2: Monitor channel quality, including RSSI and SNR, remove redundant information, optimize MAC layer protocol, set appropriate retransmission times and retransmission intervals, perform priority scheduling and multi-gateway deployment, dynamically allocate time slots and optimize terminal devices, plan gateway and node layout, and introduce cache mechanisms on nodes and gateways.
[0012] S3: Detect the movement speed of the node, adjust the modulation parameters in real time according to the movement speed of the node, and send feedback information to the gateway and server through the node.
[0013] Further, in step S1, each node of the multi-channel monitors RSSI and SNR, records and analyzes historical data, adaptively adjusts communication parameters, and multiple nodes share channel status information and combine the geographical location information of each node, use machine learning algorithms to analyze and predict, and feed back channel change information to the network server, specifically including:
[0014] S11: The node continuously monitors RSSI and SNR through different channels. The node records historical signal monitoring data and analyzes historical periodic changes and abnormal conditions.
[0015] S12: According to the analysis results of S11, the nodes adaptively adjust the communication parameters, and the multiple nodes cooperate to communicate and share the channel state information. Combined with the geographical location information of the nodes, the relationship between the channel change and the geographical location is analyzed;
[0016] S13: Use support vector machine to identify channel change patterns and trends, send current channel status information and adjusted communication parameters to the network server in real time, and the network server performs global optimization and management based on node feedback.
[0017] Further, in step S2, the monitoring of channel quality, including RSSI and SNR, removing redundant information, optimizing MAC layer protocol, setting appropriate retransmission times and retransmission intervals, performing priority scheduling and multi-gateway deployment, dynamically allocating time slots and optimizing terminal devices, planning gateway and node layout, and introducing a cache mechanism at the node and gateway end specifically include:
[0018] S21: Real-time monitoring of channel quality, including RSSI and SNR, dynamically selecting the appropriate data rate based on the monitoring results, compressing and streamlining the transmitted data, removing redundant information, and using data encoding to reduce the size of the data packet;
[0019] S22: Adjust the frame structure, reduce the overhead of the frame header and frame tail, adopt a fast confirmation mechanism, set the number of retransmissions and retransmission intervals, dynamically adjust the retransmission strategy according to the channel quality, and set priorities for different types of data;
[0020] S23: Multiple gateways are deployed in the coverage area, and nodes select the optimal gateway for communication based on signal strength, and time slots are dynamically allocated based on real-time business needs;
[0021] S24: Plan the layout of gateways and nodes, regularly evaluate and optimize the network performance, and set up caches on nodes and gateways.
[0022] Further, in step S3, the detection node's movement speed, adjusting the modulation parameters in real time according to the node's movement speed, and sending feedback information to the gateway and the server through the node specifically include:
[0023] S31: Use the acceleration sensor to detect the speed, and obtain the speed information by integrating the acceleration;
[0024] S32: Adjusting LoRaWAN modulation parameters according to the detected speed information;
[0025] S33: adjust the modulation parameters in real time according to the movement speed of the node to balance the requirements of transmission distance and data rate;
[0026] S34: The node sends feedback information to the server to inform the server of the current movement speed and the effect of the modulation parameter adjustment. The server further optimizes the modulation parameter adjustment strategy according to the feedback information.
[0027] Furthermore, the specific implementation steps of using the improved multi-layer perceptron algorithm to evaluate the data of the existing circulation boxes to determine the demand for the circulation boxes in the future include:
[0028] A1: Collect historical inventory data, including inventory levels, sales data, and relevant features of promotional activities, clean and preprocess the data, extract meaningful features from the raw data, and perform feature selection and feature dimension reduction;
[0029] A2: Determine the MLP architecture in building a multi-layer perceptron model, select a suitable activation function, train the model using the training set data, select the stochastic gradient descent optimization algorithm, and adjust the model weights through iterative training;
[0030] A3: Use validation set data to evaluate the performance of the model. Based on the evaluation results, adjust the model architecture and hyperparameters to optimize the model.
[0031] A4: Input the preprocessed current inventory data and related features into the trained MLP model to perform inventory evaluation based on the predicted demand and current inventory level.
[0032] Further, the radio frequency identification module includes a tag unit and an identification unit;
[0033] The tag unit is used to input information into the passive RFID tag set on each circulation box;
[0034] The identification unit is used for the reader to transmit a radio frequency signal of a specific frequency to the passive RFID tag, and obtain the identification information content through electromagnetic induction reflection tag information. After identification, it is stored and transmitted to the intelligent analysis module.
[0035] Furthermore, it also includes a sorting robot, which is used to obtain the label information transmitted by the identification unit and carry the corresponding circulation box for sorting and recycling.
[0036] Furthermore, the improved LoRaWAN positioning technology is used to remotely position the circulation box, and the positioning operation is performed by setting the time interval through the improved dynamic adjustment time interval algorithm, which specifically includes:
[0037] B1: Set the initial test time interval, speed threshold and adjustment coefficient. When the timer times out, perform the positioning operation to obtain the position and calculate the moving speed.
[0038] B2: Adjust the time interval and reset the timer according to the time interval to wait for the next trigger.
[0039] Furthermore, the positioning information of the positioning module is implanted into a geographic information system and presented in a visual form on a mobile terminal and a control terminal.
[0040] The beneficial effects of the present invention are:
[0041] 1. The present invention improves the channel change perception capability. By real-time monitoring of the node's movement speed, LoRaWAN can better perceive channel changes to adapt to channel changes and improve communication quality and reliability;
[0042] 2. The present invention improves the real-time performance of LoRa communication. According to the movement speed and direction of the node, LoRaWAN can predict its future location and communication needs, thereby allocating resources to it in advance and reducing channel competition and conflict;
[0043] 3. The present invention adjusts the modulation parameters and dynamically adjusts the modulation parameters according to the movement speed of the node, which can increase the data transmission rate and reduce the delay under the premise of ensuring the reliability of communication, thereby improving the real-time performance of communication;
[0044] 4. The intelligent analysis module provided in the present invention can specifically analyze and judge the existing inventory and operation status of the circulation box, predict the future demand for the circulation box, make preparations in advance, and avoid the increase of physical pressure caused by inventory shortage;
[0045] 5. The radio frequency identification module of the present invention can perform information statistics and identification on each circulation box, obtain the sorting and recycling information of the circulation box, and accurately improve the operation efficiency;
[0046] 6. The present invention can effectively perform positioning operations at a lower energy consumption state while ensuring the positioning effect by setting the time interval through an improved dynamic adjustment time interval algorithm, ensuring that the positioning equipment of the circulation box can also operate on demand when there is a long-term lack of power supply;
[0047] 7. The present invention can implant the positioning information into the geographic information system, so as to obtain the position of each circulation box more intuitively and directly, and to schedule and use the circulation boxes more efficiently. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] Figure 1 This is a flowchart of the specific improvements to LoRaWAN positioning technology;
[0049] Figure 2 It is a flowchart of the specific implementation steps of the improved multi-layer perceptron algorithm;
[0050] Figure 3 It is a flowchart of the specific implementation steps of the improved dynamic adjustment time interval algorithm. DETAILED DESCRIPTION
[0051] See also Figure 1 , Figure 2 and Figure 3 As shown, the present invention relates to an intelligent sorting and recycling system for a circulation box, comprising: a positioning module, an intelligent analysis module and a radio frequency identification module connected in sequence by wireless communication, wherein:
[0052] The positioning module is used to remotely locate the circulation box through the improved LoRaWAN positioning technology;
[0053] The intelligent analysis module is used to evaluate the data of existing recycling boxes using an improved multi-layer perceptron algorithm to determine the future demand for recycling boxes;
[0054] The radio frequency identification module is used to identify the circulation box through a passive RFID tag to determine the specific information of the circulation box.
[0055] Furthermore, the remote positioning of the circulation box by using the improved LoRaWAN positioning technology specifically includes:
[0056] S1: Each node in multiple channels monitors RSSI and SNR, records and analyzes historical data, adaptively adjusts communication parameters, and multiple nodes share channel status information and combine the geographical location information of each node. Machine learning algorithms are used to analyze and predict, and channel change information is fed back to the network server.
[0057] S2: Monitor channel quality, including RSS I and SNR, remove redundant information, optimize MAC layer protocol, set appropriate retransmission times and retransmission intervals, perform priority scheduling and multi-gateway deployment, dynamically allocate time slots and optimize terminal devices, plan gateway and node layout, and introduce cache mechanisms at the node and gateway end;
[0058] S3: Detect the movement speed of the node, adjust the modulation parameters in real time according to the movement speed of the node, and send feedback information to the gateway and server through the node.
[0059] Further, in step S1, each node of the multi-channel monitors RSSI and SNR, records and analyzes historical data, adaptively adjusts communication parameters, and multiple nodes share channel status information and combine the geographical location information of each node, use machine learning algorithms to analyze and predict, and feed back channel change information to the network server, specifically including:
[0060] S11: The node continuously monitors RSS I and SNR through different channels. The node records historical signal monitoring data and analyzes historical periodic changes and abnormal conditions.
[0061] S12: According to the analysis results of S11, the nodes adaptively adjust the communication parameters, and the multiple nodes cooperate to communicate and share the channel state information. Combined with the geographical location information of the nodes, the relationship between the channel change and the geographical location is analyzed;
[0062] S13: Use support vector machine to identify channel change patterns and trends, send current channel status information and adjusted communication parameters to the network server in real time, and the network server performs global optimization and management based on node feedback.
[0063] Specifically, nodes should continuously monitor received signal strength (RSSI) and signal-to-noise ratio (SNR) parameters, which can provide real-time information about channel quality; use multiple channels for monitoring to obtain more comprehensive channel status information, and monitor on different channels to better understand channel changes; nodes can record historical signal monitoring data to analyze channel change trends. By analyzing historical data, periodic changes or abnormalities in the channel can be discovered.
[0064] According to the monitored channel changes, nodes can adaptively adjust communication parameters, such as data rate and transmission power, to optimize communication performance; multiple nodes can cooperate to communicate and share channel status information. Through information exchange between nodes, channel changes can be more accurately perceived; combined with the geographical location information of the nodes, the relationship between channel changes and geographical location can be better understood. For example, different channel interference conditions may exist in different geographical locations.
[0065] By using, for example, a support vector machine (SVM), the input features include RSSI sequence, SNR sequence, geographic location, etc., and the output predicted channel change trends and optimal communication parameters, it can help identify the patterns and trends of channel changes, so as to make adjustments in advance; the node sends the current channel status information and adjusted communication parameters to the network server in real time, and the network server performs global optimization and management based on the feedback from multiple nodes.
[0066] This solution improves the ability to perceive channel changes. By monitoring the movement speed of nodes in real time, LoRaWAN can better perceive channel changes to adapt to channel changes and improve communication quality and reliability.
[0067] Further, in step S2, the monitoring of channel quality, including RSSI and SNR, removing redundant information, optimizing MAC layer protocol, setting appropriate retransmission times and retransmission intervals, performing priority scheduling and multi-gateway deployment, dynamically allocating time slots and optimizing terminal devices, planning gateway and node layout, and introducing a cache mechanism at the node and gateway end specifically include:
[0068] S21: Real-time monitoring of channel quality, including RSSI and SNR, dynamically selecting the appropriate data rate based on the monitoring results, compressing and streamlining the transmitted data, removing redundant information, and using data encoding to reduce the size of the data packet;
[0069] S22: Adjust the frame structure, reduce the overhead of the frame header and frame tail, adopt a fast confirmation mechanism, set the number of retransmissions and retransmission intervals, dynamically adjust the retransmission strategy according to the channel quality, and set priorities for different types of data;
[0070] S23: Multiple gateways are deployed in the coverage area, and nodes select the optimal gateway for communication based on signal strength, and time slots are dynamically allocated based on real-time business needs;
[0071] S24: Plan the layout of gateways and nodes, regularly evaluate and optimize the network performance, and set up caches on nodes and gateways.
[0072] Specifically, the channel quality is monitored in real time, including signal strength (RSSI) and signal-to-noise ratio (SNR). The measurement value of RSSI is usually in dBm, and the calculation formula is: RSSI = 10*log10(P_r / P_0), where P_r is the received power and P_0 is the reference power; the calculation formula of SNR is: SNR = Signa l_Power / Noi se_Power, where Signa l_Power is the signal power and Noi se_Power is the noise power; according to the monitoring results, the formula such as DataRate = f(RSSI, SNR) is used to dynamically select the appropriate data rate. Under the premise of ensuring communication reliability, a higher data rate is selected as much as possible to reduce the transmission time; the data to be transmitted is compressed and streamlined, unnecessary redundant information is removed, and a more efficient data encoding method is adopted to reduce the actual size of the data packet.
[0073] Adjust the frame structure to reduce the overhead of the frame header and the frame footer. Assuming that the frame header overhead is Header_Overhead, the frame footer overhead is Footer_Overhead, and the effective data length is Data_Length, the total frame length Frame_Length = Data_Length + Header_Overhead + Footer_Overhead; adopt a fast confirmation mechanism, such as setting the confirmation timeout ACK_Timeout. If the confirmation is not received within the ACK_Timeout, retransmit; set the appropriate number of retransmissions and retransmission intervals, and dynamically adjust the retransmission strategy according to the channel quality, such as increasing the number of retransmissions when the channel conditions are poor; set priorities for different types of data, such as 1 for high priority, 2 for medium priority, and 3 for low priority. High-priority data is sent first, and queued according to priority. The sending order is: send data with priority 1 first, then send data with priority 2, and finally send data with priority 3.
[0074] Deploy multiple gateways in the coverage area. Assuming that the distance from the node to gateway i is D i stance_i and the signal strength is RSS I_i, the formula for node selection of gateway is: Selected_Gateway = argmax_i (RSS I_i / D istance_i), that is, select the gateway with the largest ratio of signal strength to distance for communication; dynamically allocate time slots according to real-time service requirements to avoid time slot waste and conflict. According to real-time service requirements, dynamically calculate the required number of time slots S lot_Number, the formula is: S lot_Number = Data_Vo l ume / (Data_Rate*S lot_Duration), where Data_Vo l ume is the data volume, Data_Rate is the data rate, and S lot_Duration is the time slot duration; select high-performance chips and hardware to improve processing speed and response time. For example, select a processor with a higher main frequency to reduce data processing time.
[0075] Reasonably plan the layout of gateways and nodes, calculate the coverage according to the signal propagation model, avoid signal obstruction and interference, and regularly evaluate the network performance; for example, evaluate network performance by calculating indicators such as packet loss rate and throughput, and make optimization adjustments based on the evaluation results; set the cache size Buffer_S i ze on the node and gateway end to temporarily store data when the network is congested. The cached data amount Buffered_Data cannot exceed Buffer_S i ze, and it will be sent when conditions permit.
[0076] This solution improves the real-time performance of LoRa communication. According to the movement speed and direction of the node, LoRaWAN can predict its future location and communication needs, thereby allocating resources to it in advance and reducing channel competition and conflicts.
[0077] Further, in step S3, the detection node's movement speed, adjusting the modulation parameters in real time according to the node's movement speed, and sending feedback information to the gateway and the server through the node specifically include:
[0078] S31: Use the acceleration sensor to detect the speed, and obtain the speed information by integrating the acceleration;
[0079] S32: Adjusting LoRaWAN modulation parameters according to the detected speed information;
[0080] S33: adjust the modulation parameters in real time according to the movement speed of the node to balance the requirements of transmission distance and data rate;
[0081] S34: The node sends feedback information to the server to inform the server of the current movement speed and the effect of the modulation parameter adjustment. The server further optimizes the modulation parameter adjustment strategy according to the feedback information.
[0082] Specifically, the movement speed v of the node is obtained using an acceleration sensor device. For the acceleration sensor, the speed can be calculated by integrating the acceleration.
[0083] According to the detected speed information, the LoRaWAN modulation parameters such as the spreading factor (SF) are adjusted. A lower SF can provide a longer transmission distance, but a lower data rate; a higher SF can provide a higher data rate, but a shorter transmission distance. The selection of the spreading factor (SF) is determined according to the speed and distance requirements.
[0084] Modulation parameters are adjusted in real time according to the node's movement speed to balance the requirements of transmission distance and data rate. For example, when the node is slow or stationary, a lower SF can be selected to increase the transmission distance; when the node is fast, a higher SF can be selected to increase the data rate. The speed and related parameters are re-detected at certain time intervals to adapt to changes in the node's movement speed.
[0085] The node can send feedback information to the gateway or server to inform the current movement speed and the effect of the modulation parameter adjustment. The gateway or server can further optimize the modulation parameter adjustment strategy based on the feedback information; the node calculates the received signal strength indication (RSSI) and signal-to-noise ratio (SNR) and sends them to the gateway or server. The RSSI is usually in dBm, and the calculation formula is RSSI=10*log10(P_r / P_0), where P_r is the received power, P_0 is the reference power, and the calculation formula of SNR is SNR=Signal_Power / Noise_Power; the gateway or server further optimizes the modulation parameter adjustment strategy based on the fedback RSSI, SNR and the movement speed reported by the node.
[0086] This scheme adjusts the modulation parameters and dynamically adjusts the modulation parameters according to the movement speed of the node. Under the premise of ensuring communication reliability, it can increase the data transmission rate and reduce the delay, thereby improving the real-time performance of communication.
[0087] Furthermore, the specific implementation steps of using the improved multi-layer perceptron algorithm to evaluate the data of the existing circulation boxes to determine the demand for the circulation boxes in the future include:
[0088] A1: Collect historical inventory data, including inventory levels, sales data, and relevant features of promotional activities, clean and preprocess the data, extract meaningful features from the raw data, and perform feature selection and feature dimension reduction;
[0089] A2: Determine the MLP architecture in building a multi-layer perceptron model, select a suitable activation function, train the model using the training set data, select the stochastic gradient descent optimization algorithm, and adjust the model weights through iterative training;
[0090] A3: Use validation set data to evaluate the performance of the model. Based on the evaluation results, adjust the model architecture and hyperparameters to optimize the model.
[0091] A4: Input the preprocessed current inventory data and related features into the trained MLP model to perform inventory evaluation based on the predicted demand and current inventory level.
[0092] Specifically, historical inventory data is collected, including inventory levels, sales data, seasonal factors, promotional activities and other related features, and the data is cleaned and preprocessed, such as handling missing values and outliers, standardizing or normalizing the data, and dividing the data into training sets, validation sets and test sets; and meaningful features are extracted from the original data. For example, to calculate inventory turnover rate, average inventory holding time, etc., feature selection or feature dimensionality reduction can be performed to reduce the input dimension, improve model training efficiency and avoid overfitting.
[0093] Determine the architecture of MLP, including the number of input layer nodes (based on the number of features), the number of hidden layers and the number of nodes in each hidden layer, select a suitable activation function, such as ReLU, Sigmo id, etc.; define the loss function, the common one is mean square error (MSE): MSE = (1 / N) * sum ((y_pred-y_true)^2), where N is the number of samples, y_pred is the predicted value, y_true is the true value, use the training set data to train the model, select the stochastic gradient descent (SGD) optimization algorithm, and the formula for updating weights is: w = wl earning_rate * dw, where w is the weight, l earning_rate is the learning rate, dw is the gradient of the weight, and minimize the loss function.
[0094] Use the validation set data to evaluate the performance of the model, such as calculating the root mean square error, mean absolute error and other indicators. According to the evaluation results, adjust the model architecture (increase or decrease the number of hidden layers, the number of nodes), hyperparameters (learning rate, regularization parameters, etc.) to optimize the model.
[0095] The preprocessed current inventory data and related features are input into the trained MLP model, and the model outputs the demand forecast value for a period of time in the future; inventory evaluation is performed based on the predicted demand and current inventory level. For example, when the predicted demand is greater than the current inventory level, it is recommended to increase the purchase or production quantity.
[0096] The intelligent analysis module set up in this solution can specifically analyze and judge the existing inventory and operation status of the circulation boxes, predict the future demand for circulation boxes, prepare in advance, and avoid the increase in physical pressure caused by inventory shortages.
[0097] Further, the radio frequency identification module includes a tag unit and an identification unit;
[0098] The tag unit is used to input information into the passive RFID tag set on each circulation box;
[0099] The identification unit is used for the reader to transmit a radio frequency signal of a specific frequency to the passive RFID tag, and obtain the identification information content through electromagnetic induction reflection tag information. After identification, it is stored and transmitted to the intelligent analysis module.
[0100] Specifically, when the circulation box is manually transported, the tag unit enters information into the passive RFID tag before the circulation box leaves the warehouse. The information includes transportation product information, transportation vehicle information, destination information, inventory location, group category quantity, and LoRaWAN positioning device number.
[0101] Passive RFID tags have no power source themselves, and the energy required for their operation comes from the radio frequency signal emitted by the reader.
[0102] Before entering the warehouse, the identification unit uses a handheld reader to transmit a radio frequency signal of a specific frequency. The tag antenna receives the signal and generates an induced current on the tag antenna through electromagnetic induction, thereby providing the chip inside the tag with the energy required for operation. After receiving the energy, the chip will process the instructions and data sent by the reader and modulate the information stored in itself into the reflected radio frequency signal. The reader receives the reflected radio frequency signal with the tag information, and after demodulation and decoding, it obtains the information stored in the tag, and then transmits the information to the intelligent analysis module through wireless communication to update the inventory data for analysis and judgment.
[0103] The radio frequency identification module of this solution can perform information statistics and identification on each circulation box, obtain the sorting and recycling information of the circulation box, and accurately improve the operation efficiency.
[0104] Furthermore, it also includes a sorting robot, which is used to obtain the label information transmitted by the identification unit and carry the corresponding circulation box for sorting and recycling.
[0105] Specifically, in order to efficiently complete the sorting and recycling work, the system involved in the present invention can be provided with sorting robots, which are provided with a reader-writer unit, a route recognition unit, and an unloading and transportation unit. The reader-writer unit obtains the information of the recycling box by scanning the passive RFID tag of the recycling box, understands the placement location of the recycling box in the warehouse, plans the round-trip route through the route recognition unit, and completes the sorting and recycling work of the recycling box according to the unloading and transportation unit.
[0106] The reader-writer unit set in the sorting robot can also send the reading and writing information to the intelligent analysis module for analysis and prediction by means of wireless communication transmission.
[0107] Furthermore, the recycling box is remotely located by the improved LoRaWAN positioning technology, and the time interval is set by the improved dynamic adjustment time interval algorithm for positioning operations, specifically including:
[0108] B1: Set the initial time interval, speed threshold, and adjustment coefficient. When the timer times out, perform a positioning operation to obtain the position and calculate the moving speed.
[0109] B2: Adjust the time interval and reset the timer according to the time interval to wait for the next trigger.
[0110] Specifically, set the initial time interval T0, for example, T0 = 10 seconds; set the speed threshold V_threshold, for example, V_threshold = 5 m / s; set the adjustment coefficients k1 (used to shorten the time interval when the speed is greater than the threshold) and k2 (used to extend the time interval when the speed is less than the threshold), for example, k1 = 0.5, k2 = 1.5; set the position accuracy requirement to determine whether the positioning result is valid; when the timer times out, perform a positioning operation to obtain the current position, record the previous position, calculate the moving distance and moving time, and calculate the moving speed v.
[0111] If V > V_threshold, the new time interval T_new = T0 * k1; if V < V_threshold, the new time interval T_new = T0 * k2. To avoid the time interval being too small or too large, upper and lower limits can be set; calculate the deviation between the current positioning result and the previous valid positioning result. If there is a deviation, the current positioning result is invalid, and the previous valid position is continued to be used for speed calculation and time interval adjustment.
[0112] By this method, according to the change situation of the positioning result, such as shortening the time interval when the position moves fast and extending the time interval when the moving speed is slow, it can effectively operate in a lower energy consumption state on the premise of ensuring the positioning effect, and ensure that the positioning device of the recycling box can also operate on demand in the case of long-term lack of power supply.
[0113] Furthermore, the positioning information of the positioning module is implanted into a geographic information system and presented in a visual form on a mobile terminal and a control terminal.
[0114] Specifically, the following are the implementation steps:
[0115] 1. Obtain the original positioning data according to the positioning module, including latitude, longitude and timestamp information, and preprocess the data, such as cleaning outliers and correcting erroneous data; select QG IS as the geographic information platform to be implanted, and convert the positioning data into the CSV format supported by the selected QG IS platform.
[0116] 2. In the QG IS platform, use the corresponding import function to import the processed data, set the correct geographic coordinate system for the data, and ensure that the positioning information is accurately displayed on the map.
[0117] 3. Select a suitable visualization method, such as point markers, track lines, etc., and set the marker style, color, size and other attributes to distinguish different positioning objects or states; if the positioning data contains time information, you can set the positioning track to be displayed dynamically in chronological order.
[0118] 4. Other geographic information data can be overlaid, such as map basemaps, roads, buildings, etc., to provide richer context, mark and annotate important places or information; interactive function settings, such as zooming, panning, querying location point details, etc., to enhance the interactivity between users and maps; and the visualization results can be input to the mobile terminal and the control terminal.
[0119] The present invention can more intuitively and directly obtain the position of each circulation box through the above method, and more efficiently schedule and use the circulation box.
[0120] Specifically, the present invention provides a passive RFID tag on each circulation box; the circulation boxes are grouped according to demand, and each group is provided with a LoRaWAN positioning device, and the grouping needs to meet the requirement that each vehicle can fully accommodate all the circulation boxes in each group during transportation, so as to avoid the grouped circulation boxes not being on the same vehicle.
[0121] Due to the particularity of the positioning performance of the positioning equipment, passive RFID tags and related equipment are only used for the operation of the circulating box when it arrives at the shipping or receiving address, and are responsible for small-scale positioning and information transmission; LoRaWAN positioning equipment is used in the transportation of transportation equipment to transmit the positioning information of the circulating box group during mid-distance transportation without crossing provinces or cities; if long-distance transportation across provinces is required, in addition to passive RFID tags and related equipment and LoRaWAN positioning equipment, GPS positioning equipment must also be installed with the vehicle for positioning monitoring to obtain relevant information on the transportation situation.
[0122] The above implementation modes are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present invention.
Claims
1. An intelligent sorting and recycling system for circulation boxes, characterized in that: include: The positioning module, the intelligent analysis module and the radio frequency identification module are connected in sequence by wireless communication, wherein: The positioning module is used to remotely locate the circulation box through the improved LoRaWAN positioning technology; The intelligent analysis module is used to evaluate the data of existing recycling boxes using an improved multi-layer perceptron algorithm to determine the future demand for recycling boxes; The radio frequency identification module is used to identify the circulation box through a passive RFID tag to determine the specific information of the circulation box.
2. The intelligent sorting and recycling system for circulation boxes according to claim 1 is characterized in that: The improved LoRaWAN positioning technology is used to remotely position the circulation box, specifically including: S1: Each node in multiple channels monitors RSSI and SNR, records and analyzes historical data, adaptively adjusts communication parameters, and multiple nodes share channel status information and combine the geographical location information of each node. Machine learning algorithms are used to analyze and predict, and channel change information is fed back to the network server. S2: Monitor channel quality, including RSSI and SNR, remove redundant information, optimize MAC layer protocol, set appropriate retransmission times and retransmission intervals, perform priority scheduling and multi-gateway deployment, dynamically allocate time slots and optimize terminal devices, plan gateway and node layout, and introduce cache mechanisms on nodes and gateways. S3: Detect the movement speed of the node, adjust the modulation parameters in real time according to the movement speed of the node, and send feedback information to the gateway and server through the node.
3. The intelligent sorting and recycling system for a circulation box according to claim 2 is characterized in that: In step S1, each node of the multi-channel monitors RSSI and SNR, records and analyzes historical data, adaptively adjusts communication parameters, and multiple nodes share channel status information and combine the geographical location information of each node, use machine learning algorithms to analyze and predict, and feed back channel change information to the network server, specifically including: S11: The node continuously monitors RSSI and SNR through different channels. The node records historical signal monitoring data and analyzes historical periodic changes and abnormal conditions. S12: According to the analysis results of S11, the nodes adaptively adjust the communication parameters, and the multiple nodes cooperate to communicate and share the channel state information. Combined with the geographical location information of the nodes, the relationship between the channel change and the geographical location is analyzed; S13: Use support vector machine to identify channel change patterns and trends, send current channel status information and adjusted communication parameters to the network server in real time, and the network server performs global optimization and management based on node feedback.
4. The intelligent sorting and recycling system for circulation boxes according to claim 2 is characterized in that: In step S2, the monitoring of channel quality, including RSSI and SNR, removing redundant information, optimizing MAC layer protocol, setting appropriate retransmission times and retransmission intervals, performing priority scheduling and multi-gateway deployment, dynamically allocating time slots and optimizing terminal devices, planning gateway and node layout, and introducing a cache mechanism at the node and gateway end specifically includes: S21: Real-time monitoring of channel quality, including RSSI and SNR, dynamically selecting the appropriate data rate based on the monitoring results, compressing and streamlining the transmitted data, removing redundant information, and using data encoding to reduce the size of the data packet; S22: Adjust the frame structure, reduce the overhead of the frame header and frame tail, adopt a fast confirmation mechanism, set the number of retransmissions and retransmission intervals, dynamically adjust the retransmission strategy according to the channel quality, and set priorities for different types of data; S23: Multiple gateways are deployed in the coverage area, and nodes select the optimal gateway for communication based on signal strength, and time slots are dynamically allocated based on real-time business needs; S24: Plan the layout of gateways and nodes, regularly evaluate and optimize the network performance, and set up caches on nodes and gateways.
5. The intelligent sorting and recycling system for circulation boxes according to claim 2 is characterized in that: In step S3, the detection node's movement speed, adjusting the modulation parameters in real time according to the node's movement speed, and sending feedback information to the gateway and the server through the node specifically include: S31: Use the acceleration sensor to detect the speed, and obtain the speed information by integrating the acceleration; S32: Adjusting LoRaWAN modulation parameters according to the detected speed information; S33: adjust the modulation parameters in real time according to the movement speed of the node to balance the requirements of transmission distance and data rate; S34: The node sends feedback information to the server to inform the server of the current movement speed and the effect of the modulation parameter adjustment. The server further optimizes the modulation parameter adjustment strategy according to the feedback information.
6. The intelligent sorting and recycling system for circulation boxes according to claim 1 is characterized in that: The specific implementation steps of using the improved multi-layer perceptron algorithm to evaluate the data of the existing circulation box to determine the demand for the circulation box in the future include: A1: Collect historical inventory data, including inventory levels, sales data, and relevant features of promotional activities, clean and preprocess the data, extract meaningful features from the raw data, and perform feature selection and feature dimension reduction; A2: Determine the MLP architecture in building a multi-layer perceptron model, select a suitable activation function, train the model using the training set data, select the stochastic gradient descent optimization algorithm, and adjust the model weights through iterative training; A3: Use validation set data to evaluate the performance of the model. Based on the evaluation results, adjust the model architecture and hyperparameters to optimize the model. A4: Input the preprocessed current inventory data and related features into the trained MLP model to perform inventory evaluation based on the predicted demand and current inventory level.
7. The intelligent sorting and recycling system for circulation boxes according to claim 1 is characterized in that: The radio frequency identification module includes a tag unit and an identification unit; The tag unit is used to input information into the passive RFID tag set on each circulation box; The identification unit is used for the reader to transmit a radio frequency signal of a specific frequency to the passive RFID tag, and obtain the identification information content through electromagnetic induction reflection tag information. After identification, it is stored and transmitted to the intelligent analysis module.
8. The intelligent sorting and recycling system for circulation boxes according to claim 7 is characterized in that: It also includes a sorting robot, which is used to obtain the label information transmitted by the identification unit and carry the corresponding circulation box for sorting and recycling.
9. The intelligent sorting and recycling system for circulation boxes according to claim 1, characterized in that: The improved LoRaWAN positioning technology is used to remotely locate the circulation box, and the improved dynamic adjustment time interval algorithm is used to set the time interval for positioning operation, which specifically includes: B1: Set the initial test time interval, speed threshold and adjustment coefficient. When the timer times out, perform the positioning operation to obtain the position and calculate the moving speed. B2: Adjust the time interval and reset the timer according to the time interval to wait for the next trigger.
10. The intelligent sorting and recycling system for circulation boxes according to claim 1, characterized in that: The positioning information of the positioning module is implanted into the geographic information system and presented in a visual form on the mobile terminal and the control terminal.
