RFID multi-tag information fusion and unmanned vehicle path real-time correction method, system and device based on improved whale optimization, and storage medium

Abstract

The invention discloses an RFID multi-label information fusion and unmanned vehicle path real-time correction method, system and device based on improved whale optimization, and a storage medium, and belongs to the technical field of the crossing field of the Internet of Things and an intelligent traffic system, and the method comprises the steps: collecting received signal strength indication information and label position information, constructing a label signal propagation model and a positioning probability distribution model; performing fusion processing of multi-label positioning information, adopting a dynamic weighting strategy to adjust signal contribution, obtaining an unmanned vehicle position estimation value, taking the unmanned vehicle position estimation value as state input of path optimization, constructing a path cost function, introducing a whale optimization search mechanism based on dynamic update parameters, performing iterative optimization on a path candidate solution, and obtaining a path optimization result; outputting a path correction result; and a real-time path instruction is issued to the vehicle control module, and the unmanned vehicle is guided to complete path correction and navigation control in a dynamic environment. According to the method, the problems of poor path adaptability, low fusion precision and local optimum in the existing method are solved.

Description

Technical Field

[0001] This invention relates to the technical field of the intersection of the Internet of Things and intelligent transportation systems, specifically to a method, system, device and storage medium for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion. Background Technology

[0002] In existing technologies, research on RFID multi-tag information fusion and unmanned vehicle path correction mainly focuses on traditional optimization algorithms and path planning in static environments. These methods have poor adaptability in dynamic environments and struggle to respond in real time to environmental changes, such as dynamic obstacles and multi-tag interference in complex scenarios. For example, particle swarm optimization-based methods have low computational efficiency when processing large-scale RFID tag data and are prone to getting trapped in local optima; while information fusion techniques based on Kalman filtering lack accuracy when dealing with nonlinear dynamic environments and cannot meet the requirements of high-precision navigation.

[0003] Furthermore, existing autonomous vehicle path correction methods typically rely on pre-built static maps, lacking the ability to adjust in real time to dynamic environments. For example, some studies have proposed path planning methods based on deep learning or reinforcement learning, but these methods require large amounts of training data and computational resources, making them difficult to apply in real time in resource-constrained autonomous vehicle systems. Although hybrid methods combining statistical filtering and machine learning have emerged in recent years, they suffer from significant feature fragmentation and lack dynamic adjustment mechanisms. Existing technologies still fail to address pain points such as fixed weights for multimodal feature fusion, lack of protection for critical areas, and low hardware energy efficiency. Summary of the Invention

[0004] To address the aforementioned technical challenges, a method for RFID multi-tag information fusion and real-time path correction of unmanned vehicles based on improved whale optimization is proposed. This method includes collecting received signal strength indication information and corresponding tag location information of preset RFID tags in the environment, and constructing a tag signal propagation model and a positioning probability distribution model.

[0005] Based on the tag signal propagation model and the positioning probability distribution model, the fusion processing of multi-tag positioning information is performed, and a dynamic weighting strategy is used to adjust the signal contribution of each tag to obtain the estimated position of the unmanned vehicle at the current moment.

[0006] The estimated position of the autonomous vehicle is used as the state input for path optimization. A path cost function is constructed and a whale optimization search mechanism based on dynamically updated parameters is introduced to iteratively optimize the candidate path solutions and output the path correction results.

[0007] Based on the path correction results, real-time path commands are sent to the vehicle control module to guide the unmanned vehicle to complete path correction and navigation control in dynamic environments.

[0008] As a preferred embodiment of the improved whale-optimized RFID multi-tag information fusion and real-time path correction method for unmanned vehicles described in this invention, the method involves: collecting preset RFID tag received signal strength indication information and corresponding tag location information in the environment; and constructing a tag signal propagation model and a positioning probability distribution model.

[0009] The received signal strength indication information and the corresponding tag location information are jointly modeled. By fitting the relationship between the received signal strength and the spatial distance between the unmanned vehicle and each tag, a signal propagation model is established.

[0010] Based on the signal propagation model and combined with the received signal strength samples observed in actual observation, a probability distribution model reflecting the conditional probability between the vehicle's current position and the received signal strength is constructed to evaluate the probability density of the unmanned vehicle receiving each tag signal at different candidate positions.

[0011] The signal propagation model and probability distribution model provide a basis for signal strength prediction and a joint probability calculation framework for subsequent multi-label information fusion.

[0012] As a preferred embodiment of the improved whale-optimized RFID multi-tag information fusion and real-time path correction method for unmanned vehicles described in this invention, the multi-tag positioning information fusion processing includes:

[0013] Based on the constructed tag signal propagation model and positioning probability distribution model, the contribution of each RFID tag to the position estimation of the unmanned vehicle at the current moment is evaluated, and dynamic fusion weights are assigned based on the current contribution.

[0014] By weighted integration of the location information of several tags and the received signal strength, an estimated value of the unmanned vehicle's location information at the current moment is generated;

[0015] The estimated value serves as the input state parameter for subsequent path optimization.

[0016] The beneficial effects of this preferred technical solution are as follows: By introducing a contribution evaluation mechanism based on the tag signal propagation model and the positioning probability distribution model, the credibility of each RFID tag in the current location estimation can be dynamically judged, and fusion weights can be allocated accordingly, thereby realizing differentiated processing of multi-tag information and improving the anti-interference capability and accuracy stability of the fusion positioning results. Compared with static weighted or average fusion methods, this solution can significantly suppress the interference of long-distance weak signals or obstructed tags in complex channel environments, effectively improving the positioning robustness of unmanned vehicles in dynamic scenarios and the initial state accuracy of path optimization.

[0017] As a preferred embodiment of the improved whale optimization-based RFID multi-tag information fusion and real-time path correction method for unmanned vehicles described in this invention, the whale optimization search mechanism based on dynamically updated parameters iteratively optimizes candidate path solutions, including:

[0018] Based on the environmental disturbance information during the operation of the unmanned vehicle, the disturbance index is calculated in real time, and the search step size parameter and convergence control parameter in the whale optimization algorithm are dynamically adjusted based on the preset disturbance mapping function.

[0019] In each round of optimization iteration, based on the updated search parameters, the current candidate path solution is selected and encirclement behavior, bubble net behavior or search behavior is executed to generate path candidate update results;

[0020] And when the optimization gets stuck in a local optimum, a new search position calculation mechanism based on the perturbation direction vector and the escape distance is triggered;

[0021] The new search position replaces the solution vector of the current individual in the path population and serves as the path candidate input in the next iteration, participating in the path fitness calculation and update operation of the optimization algorithm.

[0022] The beneficial effects of this preferred technical solution are as follows: By introducing a dynamic parameter adjustment mechanism based on environmental disturbances, the whale optimization algorithm can adjust the search step size and convergence speed in real time according to dynamic factors such as obstacle changes and path blocking in the unmanned vehicle's operating environment, thereby improving the adaptability of the path search process to environmental changes; further, by combining the escape direction vector and the disturbance distance to construct a new position substitution mechanism, the path search can be effectively prevented from getting stuck in local optima, improving global optimization ability and path diversity; overall, this solution significantly enhances the convergence efficiency and global optimization stability of path planning in dynamic scenarios, and improves the real-time performance and robustness of unmanned vehicle path correction.

[0023] As a preferred embodiment of the improved whale-optimized RFID multi-tag information fusion and real-time path correction method for unmanned vehicles described in this invention, the path cost function includes:

[0024] Define a method for calculating the total path length of a path node sequence, based on the sum of the Euclidean distances between adjacent nodes in the path;

[0025] Define the distance between each node in the path and the nearest obstacle, construct an obstacle distance cost term, and combine the path length and obstacle cost term with weighting coefficients to generate a comprehensive path evaluation function.

[0026] As a preferred embodiment of the RFID multi-tag information fusion and real-time path correction method for unmanned vehicles based on improved whale optimization described in this invention, the identification of optimization trapping in a local optimum includes:

[0027] The search space is divided into several sub-regions, and path search calculations are performed independently within each sub-region.

[0028] The optimization algorithm runs in parallel in each sub-region. By calculating the path fitness, the optimal path candidate solutions of each sub-region are fused and compared, and the solution with the best fitness is selected as the global optimal path solution for the current round.

[0029] As a preferred embodiment of the improved whale-optimized RFID multi-tag information fusion and real-time path correction method for unmanned vehicles described in this invention, the step of issuing real-time path commands to the vehicle control module based on the path correction results to guide the unmanned vehicle to complete path correction and navigation control in dynamic environments includes:

[0030] A control instruction set is generated based on the optimized path node sequence, and the path instruction set is transmitted to the vehicle control in segments according to the preset control cycle. The vehicle control executes the control instructions sequentially to complete the path tracking control.

[0031] The control instruction set includes target pose parameters, velocity constraint parameters, and path segment index information.

[0032] Another objective of this invention is to provide an RFID multi-tag information fusion and unmanned vehicle path real-time correction system based on improved whale optimization. Under conditions of multi-tag interference and dynamic environmental changes, this invention improves the positioning accuracy, real-time performance, and global optimization capability of unmanned vehicles, solving the problems of poor path adaptability, low fusion accuracy, and local optimum dilemma in existing methods.

[0033] As a preferred embodiment of the RFID multi-tag information fusion and unmanned vehicle path real-time correction system based on improved whale optimization described in this invention, it is characterized by including a data acquisition and modeling module, an information fusion module, a path optimization module, and a path control module.

[0034] The data acquisition and modeling module collects the received signal strength indication information and corresponding tag location information of preset RFID tags in the environment, and constructs a tag signal propagation model and a positioning probability distribution model.

[0035] The information fusion module is based on the tag signal propagation model and the positioning probability distribution model. It performs multi-tag positioning information fusion processing and uses a dynamic weighting strategy to adjust the signal contribution of each tag to obtain the estimated position of the unmanned vehicle at the current moment.

[0036] The path optimization module takes the estimated position of the unmanned vehicle as the state input for path optimization, constructs a path cost function, and introduces a whale optimization search mechanism based on dynamically updated parameters to iteratively optimize the candidate path solutions and output the path correction results.

[0037] The path control module sends real-time path commands to the vehicle control module based on the path correction results, guiding the unmanned vehicle to complete path correction and navigation control in dynamic environments.

[0038] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the described method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion.

[0039] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion.

[0040] The beneficial effects of this invention are as follows: By improving the dynamic adaptation mechanism and local optimal escape strategy of whale optimization, this invention significantly enhances the path planning accuracy and global optimization capability of unmanned vehicles in dynamic environments, making it particularly suitable for scenarios with complex obstacle distributions. A multi-tag information fusion framework based on probability distribution, combined with dynamic weight allocation, significantly improves the positioning accuracy and anti-interference capability of RFID tags, adapting to the high-precision requirements of complex signal environments. The algorithm efficiency is optimized through a hierarchical search strategy and parallel computing technology, while dynamic obstacle avoidance constraints are introduced to ensure real-time path correction and safe obstacle avoidance performance of the unmanned vehicle in complex environments. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 The above is a flowchart of an embodiment of the present invention, which provides a method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion.

[0043] Figure 2 This is a schematic diagram of RFID tag-based unmanned vehicle positioning, provided as an embodiment of the present invention, for an RFID multi-tag information fusion and real-time path correction method for unmanned vehicles based on an improved whale-optimized RFID method.

[0044] Figure 3 The diagram shows the improved phase-shift control results of an improved whale-optimized RFID multi-tag information fusion and unmanned vehicle path real-time correction method provided in one embodiment of the present invention. Detailed Implementation

[0045] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0046] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion, including:

[0047] S1. Collect the received signal strength indication information and corresponding tag location information of the preset RFID tags in the environment, and construct the tag signal propagation model and the positioning probability distribution model.

[0048] The received signal strength indication information and the corresponding tag location information are jointly modeled. By fitting the relationship between the received signal strength and the spatial distance between the unmanned vehicle and each tag, a signal propagation model is established.

[0049] Based on the signal propagation model and combined with the received signal strength samples observed in actual observation, a probability distribution model reflecting the conditional probability between the vehicle's current position and the received signal strength is constructed to evaluate the probability density of the unmanned vehicle receiving each tag signal at different candidate positions.

[0050] The signal propagation model and probability distribution model provide a basis for signal strength prediction and a joint probability calculation framework for subsequent multi-label information fusion.

[0051] The data acquisition phase is the first step in the RFID multi-tag information fusion and real-time path correction method for unmanned vehicles based on the improved whale optimization algorithm. Its main purpose is to provide real and reliable data input for subsequent model building, information fusion, and path optimization. Data acquisition mainly includes two parts: RFID tag signal acquisition and environmental perception data acquisition.

[0052] The autonomous vehicle uses an equipped RFID reader to scan RFID tags in the surrounding environment in real time, obtaining the Received Signal Strength Indication (RSSI) value of each tag and its corresponding coordinate information (x, y). i ,y iRFID tags are pre-deployed in the environment, each with a unique ID. Signal strength reflects the distance between the autonomous vehicle and the tag. To ensure data accuracy, the collected RSSI values ​​undergo Kalman filtering to remove noise interference and improve data reliability. By collecting RFID tag signals, the autonomous vehicle can obtain accurate environmental positioning information, laying a data foundation for subsequent model building and path optimization.

[0053] Reference Figure 2 An optimization model is constructed based on the collected signals. The model construction phase involves establishing mathematical models for positioning and path optimization based on the collected RFID signals and environmental perception data. This mainly includes two parts: signal propagation model construction and probability distribution model construction. These models provide theoretical support for subsequent information fusion and path optimization.

[0054] The RFID signal propagation model describes the relationship between signal strength (RSSI) and distance, and its mathematical expression is as follows:

[0055] RSSI i =α-10βlog 10 (d i )+∈ (1)

[0056] Among them: RSSI i This represents the received signal strength indication value (dBm) of the i-th RFID tag; d i α represents the straight-line distance from the autonomous vehicle to the i-th RFID tag; α represents the signal strength at the reference distance; β represents the path loss exponent, reflecting the rate at which the signal decays with distance; ∈ represents the noise term, which measures the impact of the environment on the signal.

[0057] Obtain a set of known distances d from the data acquisition phase. i and corresponding RSSI i RFID tag data {(d1,RSSI1),(d2,RSSI2),...,(d n RSSI n The error function E is defined as the sum of squared residuals for all data points:

[0058]

[0059] Where Xi is the logarithmic distance between the i-th tag and the sensor, and RSSI i Let represent the signal strength received at the i-th point, α represent the constant term of the signal strength model, β represent the path loss factor, and E represent the total cost function.

[0060] Minimize the error function:

[0061]

[0062] Where n is the total number of observations for the current region label, Xi is the logarithmic distance between the i-th label and the sensor, and RSSI i This represents the signal strength received at the i-th point.

[0063] Solving the system of equations yields α and β:

[0064]

[0065] Then, a probability distribution model is constructed to describe the likelihood of the autonomous vehicle receiving RFID tag signals at different locations (x, y). Its mathematical expression is:

[0066]

[0067] Where: P i (x,y) represents the probability density of receiving the i-th RFID tag signal at location (x,y); RSSI i This represents the actual measured signal strength; σ represents the signal strength predicted based on the signal propagation model. i It represents the noise standard deviation, reflecting the uncertainty of the signal.

[0068] Using the calibrated signal propagation model, calculate the predicted signal strength received at location (x, y) for the i-th RFID tag. Based on actual measured RSSI i and predicted value Calculate the probability density P i (x,y). σ i The range of signal strength fluctuations can be obtained through statistical analysis of experimental data.

[0069] S2. Based on the tag signal propagation model and the positioning probability distribution model, perform multi-tag positioning information fusion processing, and use a dynamic weighting strategy to adjust the signal contribution of each tag to obtain the estimated position of the unmanned vehicle at the current moment.

[0070] Based on the constructed tag signal propagation model and positioning probability distribution model, the contribution of each RFID tag to the position estimation of the unmanned vehicle at the current moment is evaluated, and dynamic fusion weights are assigned based on the current contribution.

[0071] By weighted integration of the location information of several tags and the received signal strength, an estimated value of the unmanned vehicle's location information at the current moment is generated;

[0072] The estimated value serves as the input state parameter for subsequent path optimization.

[0073] A preferred embodiment of the present invention for performing multi-tag location information fusion processing is as follows: For multiple RFID tags, their probability densities are calculated separately, and the multi-tag information is fused through a joint probability distribution.

[0074]

[0075] Where P(x,y) represents the joint probability density of the target at position (x,y), P i (x,y) represents the local probability estimate of the target at location (x,y) from the i-th observation point (e.g., an RFID tag or base station), where n is the total number of observations from tags in the current area. The collected information is then fused. Information fusion integrates data from multiple RFID tags to improve the accuracy and robustness of autonomous vehicle positioning.

[0076] First, dynamic weight allocation is adopted based on the signal strength RSSI of the RFID tag. i Distance d from label i Assign different weights to each label. i This reflects its credibility. The specific formula is as follows:

[0077]

[0078] Among them, RSSI i d represents the signal strength of the RFID tag. i w represents the distance between labels. i The corresponding weight for each label.

[0079] RSSI i The smaller the value w, the weaker the signal strength, indicating that the tag may be far from the autonomous vehicle or that the signal is being interfered with. Therefore, the weight w is... i It should be relatively small. d i The larger the value w, the farther the tag is from the autonomous vehicle, and its location information may be inaccurate. Therefore, the weight w is relatively high. i It should be relatively small. By taking the reciprocal, RSSI is reduced. i and d i Taking all the factors into account, we ensure that tags with weak signal strength or far distance have lower weights.

[0080] Then, based on the weighted least squares method and fusing the positioning information from multiple RFID tags, the optimal position estimate (x) of the unmanned vehicle is solved. r ,y r Its optimization objective is:

[0081]

[0082] Where, d i w represents the Euclidean distance from the i-th label to the inferred location. iThe confidence score or weight of the i-th label, α and β are the coefficients obtained from the model fitting, and RSSI i This refers to the signal strength of the RFID tag.

[0083] Where d i Location of the driverless car (x) r ,y r ) to the position of the i-th label (x i ,y i Euclidean distance:

[0084]

[0085] Define the objective function E1, which represents the weighted sum of squared residuals for all labels:

[0086]

[0087] Calculate the partial derivatives, set them equal to zero, and you will obtain the optimal solution:

[0088]

[0089] Calculation yields:

[0090]

[0091] The purpose of finding the partial derivative is to minimize the cost function and find the optimal positioning position of the unmanned vehicle in the two-dimensional plane, so as to minimize the sum of squares of all weighted positioning errors, taking into account the influence of distance and signal strength factors.

[0092] In one optional embodiment of the present invention, the fusion processing of multi-tag location information does not employ a dynamic fusion strategy based on signal propagation model and probability distribution model, but instead uses a static weighted average method to achieve the fusion of multiple RFID tag data.

[0093] In one optional embodiment of the present invention, the multi-tag localization information fusion processing does not employ optimization strategies such as probability distribution models and weighted least squares methods, but instead uses the maximum signal strength tag localization method as the basis for estimating the location of the unmanned vehicle.

[0094] Specifically, within each positioning cycle, this method selects only the tag with the highest signal strength (RSSI) among all currently received RFID tags and uses the known spatial location of that tag as an estimate of the autonomous vehicle's current location. That is, there is no need to calculate the weights of multiple tags or perform signal propagation models; no need for probabilistic modeling and weighted optimization; the system determines the unique "most reliable" tag as the positioning reference point through a simple maximum value retrieval strategy.

[0095] A preferred embodiment of the invention introduces a signal propagation model and a positioning probability distribution model during the fusion stage. By calculating the signal probability density value of each tag at its current location, fusion weights are dynamically allocated, and a weighted least squares method is used to estimate the global optimal location. This scheme not only significantly improves the adaptability and accuracy of the information fusion process, but also has stronger error resistance and stability in complex interference environments, effectively supporting the state input requirements of the subsequent path optimization module.

[0096] S3. Using the estimated position of the unmanned vehicle as the state input for path optimization, a path cost function is constructed and a whale optimization search mechanism based on dynamically updated parameters is introduced to iteratively optimize the candidate path solutions and output the path correction results.

[0097] Based on the environmental disturbance information during the operation of the unmanned vehicle, the disturbance index is calculated in real time, and the search step size parameter and convergence control parameter in the whale optimization algorithm are dynamically adjusted based on the preset disturbance mapping function.

[0098] In each round of optimization iteration, based on the updated search parameters, the current candidate path solution is selected and encirclement behavior, bubble net behavior or search behavior is executed to generate path candidate update results;

[0099] And when the optimization gets stuck in a local optimum, a new search position calculation mechanism based on the perturbation direction vector and the escape distance is triggered;

[0100] The new search position replaces the solution vector of the current individual in the path population and serves as the path candidate input in the next iteration, participating in the path fitness calculation and update operation of the optimization algorithm.

[0101] Define a method for calculating the total path length of a path node sequence, based on the sum of the Euclidean distances between adjacent nodes in the path;

[0102] Define the distance between each node in the path and the nearest obstacle, construct an obstacle distance cost term, and combine the path length and obstacle cost term with weighting coefficients to generate a comprehensive path evaluation function.

[0103] In a preferred embodiment of the present invention, the whale optimization search mechanism with dynamically updated parameters is as follows:

[0104] WOA is an optimization algorithm based on the natural feeding behavior of whales. In path optimization, WOA gradually optimizes the path by simulating the whale's search, encirclement, and bubble-net feeding behaviors.

[0105] First, the parameters are initialized, including the initial search step size, convergence factor, perturbation coefficient, and random vector. The search space is then initialized, defining the start and end points of the autonomous vehicle. To adapt to environmental changes (such as dynamic obstacles) and increase the algorithm's robustness, the search step size and convergence speed are dynamically adjusted.

[0106]

[0107] in: This represents the parameters that the algorithm dynamically adjusts after being affected by the environment. γ represents the initial parameters set in the dynamic optimization algorithm, where γ is the adjustment coefficient and ΔE represents the environmental change. A larger ΔE indicates significant environmental change, thus increasing the search step size and expanding the search range. Conversely, a smaller ΔE indicates a stable environment, thus decreasing the search step size and improving search accuracy.

[0108] In the improved whale path planning method proposed in this invention, the search behavior, encirclement behavior, and bubble net predation behavior are not executed independently in sequence, but rather based on the intensity of environmental disturbances fed back by the RFID tag. And dynamic collaborative driving of path deviation. Specifically, the algorithm obtains the current environmental state in real time through the disturbance perception module, and calculates the weight coefficients of the three behaviors based on the disturbance intensity and path error, and dynamically allocates the behavior proportions. Each behavior generates a path correction instruction, which is then weighted and summarized in a unified path fusion expression to form the next direction of motion, thereby realizing the adaptive adjustment and multi-objective trade-off of the autonomous vehicle path.

[0109] To improve behavioral flexibility, the system introduces a perturbation-driven weight allocation mechanism:

[0110] w encircle +w bubble +w search =1 (17)

[0111] Among them, w encircle For the weights of the enclosing behavior, w bubble w is the weight for bubble behavior. search These are the weights assigned to search behavior. Their respective calculation methods are as follows:

[0112]

[0113] w encircle =1-w search -w bubble (20)

[0114] Where, ε p This represents the deviation value of the current path, ΔE is the amount of environmental change, and ∈ is a small quantity to prevent division by zero.

[0115] The final outcome of the action was:

[0116] P t+1 =w encircle ·P encircle +w bubble ·P bubble +w search ·P searc h (twenty one)

[0117] Among them, P encircle P represents the path that encloses the output of the behavior. bubble P represents the path of the bubble behavior output. bubble This represents the path output by the bubble behavior. Path fusion is then performed to obtain the path input for the next iteration.

[0118] Encirclement behavior:

[0119] D e = |C e ·P * (t)-P(t)∣ (22)

[0120] P encircle =P(t)-A e ·D e (twenty three)

[0121] Among them, P * (t) represents the current optimal path, P(t) represents the individual position at time t in the current round, and A e C e D is the parameter vector enclosing the behavior. e Let P be the distance vector between an individual and the optimal path. encircle This represents the path that encloses the output of the enclosing behavior. This step is used to represent the behavior of individual paths approaching the current optimal path.

[0122] Bubble web predation behavior:

[0123] P bubble =P * (t)+D b ·e bl ·cos(2πl) (24)

[0124] Where b is the spiral shape constant, l is the random perturbation variable in the range [-1,1], and D b P represents the Euclidean distance between the current individual and the optimal path. bubble This represents the path output by the bubble behavior. This step simulates the local search behavior of spiraling towards the target.

[0125] Search behavior:

[0126] D s = |Cs ·P rand (t)-P(t)∣ (25)

[0127] P search =P rand (t)-A s ·D s (26)

[0128] Among them, P rand (t) represents a randomly selected path, D s Let A represent the distance between an individual and a random path during the search process. s C s P represents the vector of behavioral parameters in the search behavior. search This indicates the path output by the search action. This step simulates a large-scale jump search.

[0129] Then, a path optimization objective function is constructed to evaluate the quality of the path, including path length and obstacle collision risk:

[0130]

[0131] Where λ is the tradeoff coefficient. C distance The calculation method for (P) is as follows:

[0132]

[0133] Where P i Let be the i-th node of the path, and n be the total number of nodes in the path.

[0134] Solve for the obstacle collision risk cost C obstacle (P), calculates the proximity of the path to the obstacle:

[0135]

[0136] Among them, risk(P) i ) represents node P i The closer the distance to the nearest obstacle, the greater the risk. i ε represents the distance between the node and the obstacle, and ε represents the decimal to prevent division by zero.

[0137] Furthermore, to avoid the algorithm getting trapped in local optima, a local optimum escape strategy was constructed:

[0138] X(t+1)=X rand +η·D escape (31)

[0139]

[0140] Where X(t+1) represents the whale's current new position; X rand D represents a randomly selected location; escape The escape direction vector is represented by η; the perturbation coefficient controls the escape step size; X(t) represents the current position trapped in a local optimum; X best δ represents the globally optimal individual; rand It represents a random, tiny perturbation.

[0141] Although the variables involved in the three behaviors mentioned above are similar in form, their meanings are clearly differentiated in specific behaviors. They are bound to the corresponding behavior modules during use to avoid confusion. Simultaneously, the three types of path points will be weighted using a unified path fusion formula to output the next path point, achieving a closed-loop expression of the path update logic. This mechanism ensures that the algorithm not only possesses global search capabilities but also fine-grained obstacle avoidance and convergence capabilities, enabling highly robust path planning and real-time correction in complex RFID environments. Furthermore, a hierarchical search strategy and parallel computing technology are employed to accelerate the path optimization process.

[0142]

[0143] The search space is divided into multiple sub-regions, and each sub-region is searched independently. The search results from each sub-region are merged to obtain the global optimal solution. Multithreading or distributed computing techniques are used to evaluate multiple candidate paths simultaneously.

[0144] In one optional embodiment of the present invention, the whale optimization search mechanism with dynamically updated parameters is to use a standard whale optimization algorithm with fixed parameters to perform path search.

[0145] In the initialization phase of path optimization, fixed optimization control parameters such as search step size, convergence control parameters, and perturbation coefficients are preset and remain unchanged throughout the iteration process, lacking a dynamic adjustment mechanism. During the path search process, the algorithm performs three types of fixed behaviors in each iteration based on the state relationship between the current path and the optimal path solution:

[0146] If the current path individual is close to the current global optimal path, perform an encirclement action to move the path solution closer to the optimal direction;

[0147] If the set spiral update cycle is met, the bubble net predation behavior is executed, and the path solution converges and updates spirally around the optimal path;

[0148] Under other conditions, the search behavior is performed, and the path solution is updated with positional perturbation relative to another randomly selected solution vector.

[0149] The three types of behavior are switched according to fixed rules and are not dynamically adjusted according to environmental conditions or search progress; the control parameters used remain constant throughout the entire algorithm operation.

[0150] The objective function for path optimization consists of the total path length and the obstacle distance cost term. These terms are calculated based on the distance between path nodes and the shortest distance to obstacles, and then combined using weighted coefficients to form a comprehensive evaluation index.

[0151] In one optional embodiment of the present invention, the whale optimization search mechanism with dynamically updated parameters still uses the whale optimization algorithm (WOA) as the basic framework of the search strategy during the path optimization process. However, it does not introduce a perturbation-driven dynamic allocation mechanism for behavior weights, nor does it use behavior fusion expressions to weightedly combine the path outputs of the three types of behaviors. Instead, it adopts a fixed-ratio behavior selection + discrete sampling method based on the perturbation range to achieve simplified control of the path search process.

[0152] Specifically, in this embodiment, each round of path search iteration sets a fixed probability of behavior execution, for example: encirclement behavior accounts for 40%; bubble web predation behavior accounts for 30%; and search behavior accounts for 30%.

[0153] During path optimization, the algorithm does not evaluate current environmental disturbance information, nor does it update behavior weights in real time. Instead, it samples the behavior set in rounds according to a set ratio and selects the corresponding path update formula to execute.

[0154] The preferred embodiment of the present invention introduces a dynamic control parameter adjustment strategy based on environmental disturbances, which can adjust the search step size and convergence speed in real time according to the degree of disturbance, thereby improving the adaptability of path search to dynamic environments. At the same time, a local optimum escape mechanism is constructed, which generates alternative paths through disturbance vectors when the path gets stuck in a local extremum, effectively expanding the search space. In addition, the preferred embodiment further adopts a hierarchical search and parallel computing strategy to improve the efficiency of path optimization and ensure the real-time performance and accuracy of path planning in complex environments, which is significantly better than the optional solution.

[0155] S4. Based on the path correction results, send real-time path instructions to the vehicle control module to guide the unmanned vehicle to complete path correction and navigation control in dynamic environments.

[0156] The search space is divided into several sub-regions, and path search calculations are performed independently within each sub-region.

[0157] The optimization algorithm runs in parallel in each sub-region. By calculating the path fitness, the optimal path candidate solutions of each sub-region are fused and compared, and the solution with the best fitness is selected as the global optimal path solution for the current round.

[0158] A control instruction set is generated based on the optimized path node sequence, and the path instruction set is transmitted to the vehicle control in segments according to a preset control cycle. The vehicle control executes the control instructions sequentially to complete the path tracking control. The control instruction set includes target pose parameters, speed constraint parameters, and path segment index information.

[0159] The system parses the path node sequence output by the path optimization module, extracting the target position coordinates, desired driving direction, and corresponding speed limit information for each node, and assigns path segment indices according to the path's logical structure. This information is then uniformly encapsulated into a structured control instruction set. Each control instruction corresponds to a path segment and includes target pose parameters, speed constraint parameters, and path segment index information.

[0160] Subsequently, the system segments and schedules the generated path instruction set according to a preset control cycle or path segment transition rules. The scheduling mechanism can flexibly set the control frequency based on task complexity and environmental dynamics, supporting fixed-cycle scheduling or trigger-based scheduling strategies based on obstacle proximity. During each scheduling, the system sends the control instruction for the current path segment to the vehicle control module and waits for the segment to complete execution or receive update feedback before issuing the next instruction.

[0161] Upon receiving the path control command, the vehicle control module analyzes the target pose and speed parameters, which are then executed by the underlying controller. During execution, the vehicle continuously compares the deviation between its current position and the target path segment, dynamically adjusting its direction and speed to achieve continuous tracking control of the optimized path. When the control module detects that the vehicle deviates from the path by more than a set threshold, or that the path ahead becomes unreachable due to environmental changes (such as sudden obstacles), it will proactively report the abnormal path status and request path replanning, thus re-entering the path optimization phase in a closed-loop manner.

[0162] Through the above control strategy, the present invention achieves seamless connection between path correction results and vehicle execution behavior, ensuring the feasibility and robustness of path planning in actual dynamic environments, and improving the overall system's navigation stability and path tracking accuracy.

[0163] It should be further explained that:

[0164] The proposed method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion is achieved through dynamic adaptation mechanism improvement, multi-tag information fusion model optimization, local optimal escape strategy, path correction and obstacle avoidance integration, and computational efficiency optimization. This method effectively solves the technical problems of low path correction efficiency, insufficient positioning accuracy, and high computational complexity of unmanned vehicles in dynamic environments, and improves the path planning accuracy and dynamic environment adaptability of unmanned vehicles, especially in scenarios with complex obstacle distribution and signal interference.

[0165] Table 1 Comparison of the present invention with previous technologies

[0166]

[0167]

[0168] Example 2, refer to Figure 3 This is the second embodiment of the present invention, providing a method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion. To verify the beneficial effects of the invention, scientific demonstration was conducted through experiments. In 10 independent positioning experiments, the test environment was a semi-open indoor warehouse scenario measuring 15m × 8m × 3.5m (length × width × height). The system exhibited millimeter-level high-precision positioning performance, achieving an average positioning error of only 0.011 meters (11 millimeters) and a standard deviation of 0.0073 meters (7.3 millimeters). This result demonstrates that even under conditions of low tag deployment density, the proposed positioning algorithm can still stably provide sub-centimeter-level spatial resolution with a small error fluctuation range (standard deviation only 66.4% of the average error), fully verifying the high reliability and repeatability of the system in sparse RFID environments. This performance can effectively support the precise pose control requirements of applications such as unmanned vehicles in complex scenarios.

[0169] Example 3 is the third embodiment of the present invention, which differs from the previous two embodiments in that:

[0170] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0171] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0172] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0173] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0174] Example 4 is the fourth embodiment of the present invention. This embodiment provides an RFID multi-tag information fusion and unmanned vehicle path real-time correction system based on improved whale optimization, including a data acquisition and modeling module, an information fusion module, a path optimization module, and a path control module.

[0175] The data acquisition and modeling module collects the received signal strength indication information and corresponding tag location information of preset RFID tags in the environment, and constructs a tag signal propagation model and a positioning probability distribution model.

[0176] The information fusion module is based on the tag signal propagation model and the positioning probability distribution model. It performs multi-tag positioning information fusion processing and uses a dynamic weighting strategy to adjust the signal contribution of each tag to obtain the estimated position of the unmanned vehicle at the current moment.

[0177] The path optimization module takes the estimated position of the unmanned vehicle as the state input for path optimization, constructs a path cost function, and introduces a whale optimization search mechanism based on dynamically updated parameters to iteratively optimize the candidate path solutions and output the path correction results.

[0178] The path control module sends real-time path commands to the vehicle control module based on the path correction results, guiding the unmanned vehicle to complete path correction and navigation control in dynamic environments.

[0179] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion, characterized in that: include, Collect the received signal strength indication information and corresponding tag location information of preset RFID tags in the environment, and construct a tag signal propagation model and a positioning probability distribution model; Based on the tag signal propagation model and the positioning probability distribution model, the fusion processing of multi-tag positioning information is performed, and a dynamic weighting strategy is used to adjust the signal contribution of each tag to obtain the estimated position of the unmanned vehicle at the current moment. The estimated position of the autonomous vehicle is used as the state input for path optimization. A path cost function is constructed and a whale optimization search mechanism based on dynamically updated parameters is introduced to iteratively optimize the candidate path solutions and output the path correction results. Based on the path correction results, real-time path commands are sent to the vehicle control module to guide the unmanned vehicle to complete path correction and navigation control in dynamic environments.

2. The method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion as described in claim 1, characterized in that: The data acquisition environment includes preset RFID tag received signal strength indication information and corresponding tag location information. The construction of a tag signal propagation model and a positioning probability distribution model includes... The received signal strength indication information and the corresponding tag location information are jointly modeled. By fitting the relationship between the received signal strength and the spatial distance between the unmanned vehicle and each tag, a signal propagation model is established. Based on the signal propagation model and combined with the received signal strength samples observed in actual observation, a probability distribution model reflecting the conditional probability between the vehicle's current position and the received signal strength is constructed to evaluate the probability density of the unmanned vehicle receiving each tag signal at different candidate positions. The signal propagation model and probability distribution model provide a basis for signal strength prediction and a joint probability calculation framework for subsequent multi-label information fusion.

3. The method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion as described in claim 2, characterized in that: The process of fusing multi-tag location information includes, Based on the constructed tag signal propagation model and positioning probability distribution model, the contribution of each RFID tag to the position estimation of the unmanned vehicle at the current moment is evaluated, and dynamic fusion weights are assigned based on the current contribution. By weighted integration of the location information of several tags and the received signal strength, an estimated value of the unmanned vehicle's location information at the current moment is generated; The estimated value serves as the input state parameter for subsequent path optimization.

4. The method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion as described in claim 3, characterized in that: The whale optimization search mechanism based on dynamically updated parameters iteratively optimizes candidate path solutions, including... Based on the environmental disturbance information during the operation of the unmanned vehicle, the disturbance index is calculated in real time, and the search step size parameter and convergence control parameter in the whale optimization algorithm are dynamically adjusted based on the preset disturbance mapping function. In each round of optimization iteration, based on the updated search parameters, the current candidate path solution is selected and encirclement behavior, bubble net behavior or search behavior is executed to generate path candidate update results; And when the optimization gets stuck in a local optimum, a new search position calculation mechanism based on the perturbation direction vector and the escape distance is triggered; The new search position replaces the solution vector of the current individual in the path population and serves as the path candidate input in the next iteration, participating in the path fitness calculation and update operation of the optimization algorithm.

5. The method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion as described in claim 4, characterized in that: The construction path cost function includes, Define a method for calculating the total path length of a path node sequence, based on the sum of the Euclidean distances between adjacent nodes in the path; Define the distance between each node in the path and the nearest obstacle, construct an obstacle distance cost term, and combine the path length and obstacle cost term with weighting coefficients to generate a comprehensive path evaluation function.

6. The method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion as described in claim 5, characterized in that: The identification of optimization getting trapped in a local optimum includes, The search space is divided into several sub-regions, and path search calculations are performed independently within each sub-region. The optimization algorithm runs in parallel in each sub-region. By calculating the path fitness, the optimal path candidate solutions of each sub-region are fused and compared, and the solution with the best fitness is selected as the global optimal path solution for the current round.

7. The method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion as described in claim 6, characterized in that: The step of issuing real-time path commands to the vehicle control module based on the path correction results to guide the autonomous vehicle to complete path correction and navigation control in dynamic environments includes: A control instruction set is generated based on the optimized path node sequence, and the path instruction set is transmitted to the vehicle control in segments according to the preset control cycle. The vehicle control executes the control instructions sequentially to complete the path tracking control. The control instruction set includes target pose parameters, velocity constraint parameters, and path segment index information.

8. A system for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion, applying the method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion as described in any one of claims 1 to 7, characterized in that, include: The module includes data acquisition and modeling, information fusion, path optimization, and path control. The data acquisition and modeling module collects the received signal strength indication information and corresponding tag location information of preset RFID tags in the environment, and constructs a tag signal propagation model and a positioning probability distribution model. The information fusion module is based on the tag signal propagation model and the positioning probability distribution model. It performs multi-tag positioning information fusion processing and uses a dynamic weighting strategy to adjust the signal contribution of each tag to obtain the estimated position of the unmanned vehicle at the current moment. The path optimization module takes the estimated position of the unmanned vehicle as the state input for path optimization, constructs a path cost function, and introduces a whale optimization search mechanism based on dynamically updated parameters to iteratively optimize the candidate path solutions and output the path correction results. The path control module sends real-time path commands to the vehicle control module based on the path correction results, guiding the unmanned vehicle to complete path correction and navigation control in dynamic environments.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of any one of claims 1 to 7 of the method for real-time path correction of unmanned vehicles based on improved whale-optimized RFID multi-tag information fusion.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of any one of the claims 1 to 7: an improved whale-optimized RFID multi-tag information fusion and unmanned vehicle path real-time correction method.

Images