Bird trajectory prediction system based on bird detection radar and big data model analysis

Abstract

The invention relates to the technical field of bird trajectory prediction, in particular to a bird trajectory prediction system based on bird detection radar and big data model analysis, which comprises a data acquisition module, a prediction model establishment module and a prediction model updating module, and is characterized in that the data acquisition module acquires the speed, acceleration, trajectory and time point data of a bird; the prediction model establishing module sets a trajectory prediction model by using the data, predicts the bird trajectory of the next time point according to the data of the preorder time point, calculates a prediction error, optimizes the error through a training strategy in the model so as to update prediction parameters, and records the time required for prediction; the prediction model updating module calculates an influence coefficient according to the error and the prediction time so as to adjust and update a training strategy of the model; according to the process, the problem that a traditional system is difficult to dynamically adapt to bird movement changes is solved, the prediction accuracy and efficiency are improved through a self-adaptive training strategy, and a more stable and reliable bird trajectory prediction effect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of bird trajectory prediction, and in particular to a bird trajectory prediction system based on bird detection radar and big data model analysis. Background Art

[0002] Traditional bird trajectory prediction systems often face several technical problems in practical applications. The first is the limitation of data processing. These systems usually rely on fixed parameter adjustments and static models, which are difficult to effectively deal with the dynamic changes that occur during the bird's movement, resulting in large deviations in the prediction results. Due to the lack of real-time update and adaptive adjustment mechanisms, traditional systems are unable to correct errors in the model in a timely manner and achieve accurate predictions of complex natural environments.

[0003] Traditional systems often neglect to fully quantify and optimize errors during data processing, which means that it is difficult for the model to maintain high-precision stable output when processing diverse and complex data sets. The system may also not fully consider the efficiency of computing time during training, resulting in prediction delays and affecting the effectiveness of real-time applications. For these systems, a single static learning strategy cannot adapt to the changing needs of different data scenarios. The lack of strategy makes the model inefficient and inflexible when faced with a large amount of changing data.

[0004] In addition, because traditional systems fail to use advanced algorithms such as gradient descent to optimize errors, model parameter adjustment lacks flexibility and precision, which poses a challenge to the accurate prediction of complex trajectories. This not only reduces the reliability of the prediction system, but may also cause waste in resource utilization. Therefore, when facing bird trajectory prediction, how to use a more dynamic and adaptive strategy to update the model has become a problem that traditional systems need to solve urgently. Summary of the invention

[0005] The purpose of the present invention is to provide a bird trajectory prediction system based on bird detection radar and big data model analysis to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a bird trajectory prediction system based on bird detection radar and big data model analysis, including a data acquisition module, a prediction model establishment module and a prediction model update module, wherein: The data acquisition module acquires bird detection radar data, which includes bird speed, bird acceleration, bird trajectory and time point; The prediction model establishment module sets a trajectory prediction model, and the trajectory prediction model is trained according to the bird trajectories at different time points in the bird detection radar data; the trajectory prediction model is used to predict the bird trajectory at the next time point according to the bird speed, bird acceleration and bird trajectory at the previous time point in the bird detection radar data, and the error is calculated, wherein the error is calculated based on the bird trajectory predicted by the trajectory prediction model at the next time point and the bird trajectory in the bird detection radar data at that time point; and the training strategy in the trajectory prediction model is used to optimize the error to update the prediction parameters in the trajectory prediction model, and the time required for the bird trajectory at the next time point predicted by the trajectory prediction model is recorded as the model prediction time; The prediction model updating module calculates the influence coefficient according to the error sum of each calculation and the recorded model prediction time, and updates the training strategy in the trajectory prediction model according to the influence coefficient; and uses the updated trajectory prediction model to predict the bird trajectory at future time points in the bird detection radar data.

[0007] As a further improvement of the technical solution, the prediction model building module includes an initial prediction unit, and the process of the initial prediction unit recording the model prediction time specifically includes: The bird-detecting radar data is divided into multiple time segments, and a sliding window is created. Each sliding window contains a series of continuous time points. The bird speed, bird acceleration and bird trajectory at the previous time points in the bird-detecting radar data are used as the input of the trajectory prediction model, that is, the previous time point data of the last time point data in the sliding window is used as the input of the trajectory prediction model; the trajectory prediction model is used to predict the bird trajectory at the next time point, the prediction result of the trajectory prediction model is output, and the time required for the trajectory prediction model to predict the bird trajectory at the next time point is recorded as the model prediction time.

[0008] As a further improvement of the present technical solution, the prediction model building module includes an error optimization unit, and the error optimization unit performs error optimization specifically including: Calculate the error between the bird trajectory predicted by the trajectory prediction model at the next time point and the bird trajectory of the bird detection radar data at that time point. Use the gradient descent training strategy to optimize the error. By counting the mean square error of the error and calculating the partial derivative of the mean square error, use the chain rule to calculate the gradient of the mean square error to the parameters in the trajectory prediction model. The calculation formula of the mean square error is as follows: ,in It indicates that the trajectory prediction model predicts the bird's trajectory at the next time point; Indicates the bird trajectory of the bird detection radar data at this time point, specifically the bird trajectory at the last time point in the sliding window; Indicates the number of samples of bird radar data; Relying on the optimizer, the calculated gradient is used to update the prediction parameters of the trajectory prediction model. Through batch processing of multiple training cycles, the prediction parameters of the trajectory prediction model are continuously updated to minimize the error. The prediction parameter adjustment formula is as follows: For the prediction parameters , ,in, is the learning rate, is the mean square error with respect to the parameter gradient.

[0009] As a further improvement of the present technical solution, the prediction model updating module includes an influence coefficient calculation unit, and the process of calculating the influence coefficient by the influence coefficient calculation unit specifically includes: The influence coefficient is calculated based on the error of each calculation and the recorded model prediction time. The calculation formula is: ,in represents the influence coefficient; Represents the error of each calculation, ; The model prediction time for each record, that is, the output the time used; is a factor used to normalize the influence coefficient.

[0010] As a further improvement of the present technical solution, the prediction model updating module includes a model updating unit, and the process of the model updating unit performing model updating specifically includes: The training strategy in the trajectory prediction model is updated according to the influence coefficient, and the training strategy of the gradient descent method is updated. Make adjustments, and the adjustment formula is as follows: , which means that if the influence coefficient is higher, the learning rate is reduced, otherwise it is increased.

[0011] Compared with the prior art, the present invention has the following beneficial effects: The bird trajectory prediction system based on bird radar and big data model analysis predicts the bird trajectory at the next time point using the trajectory prediction model according to the bird speed, acceleration and trajectory at the previous time point in the bird radar data, and calculates the error, where the error is obtained by comparing the predicted trajectory with the bird trajectory in the actual bird radar data; this error calculation is a key feedback mechanism that enables the model to identify and quantify the deviation from the real data; therefore, the error calculation directly affects the accuracy and precision of the model. The smaller the error, the closer the performance of the prediction model is to reality. In this process, the core task is to continuously optimize the error using the training strategy in the trajectory prediction model. Specifically, the model gradually reduces the error by updating the prediction parameters, which is usually achieved through techniques such as gradient descent. This optimization strategy not only improves the accuracy of the model, but also makes the model more sensitive and reliable when processing actual data; in the optimization process, the parameters are continuously adjusted and improved, and the model gradually learns and adapts to the changes in the data, improving its performance in predicting trajectories.

[0012] In addition, the influence coefficient is calculated based on the error of each calculation and the recorded model prediction time. This coefficient integrates the dual information of error size and prediction time consumption, which not only reflects the prediction accuracy of the model, but also points out its computational efficiency. The influence coefficient is an important reference for adjusting the model training strategy, helping to achieve adaptive optimization of the training process. Specifically, a higher influence coefficient indicates that the prediction error is high or the time consumption is too long. At this time, the learning rate needs to be reduced to avoid the model from updating parameters too aggressively to ensure stability and convergence. A lower influence coefficient indicates that the prediction is more accurate and efficient, and the model learning speed can be accelerated by increasing the learning rate, thereby saving training time. In this way, the influence coefficient realizes the dynamic adjustment of the training strategy, enabling the model to adjust its behavior according to real-time data, improving overall performance and resource utilization. This adaptive training strategy not only maintains the efficiency and stability of the model in a constantly changing environment, but also greatly reduces the limitations caused by the conventional static learning rate setting, and improves the practicality and reliability of the model in complex situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the overall module of the present invention; Figure 2 A schematic diagram of a module for establishing a prediction model of the present invention; Figure 3 Schematic diagram of the prediction model updating module of the present invention.

[0014] In the figure: 100, data acquisition module; 200, prediction model building module; 201, initial prediction unit; 202, error optimization unit; 300, prediction model updating module; 301, influence coefficient calculation unit; 302, model updating unit. DETAILED DESCRIPTION

[0015] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] Next, see Figure 1-Figure 3 The present invention provides a technical solution: a bird trajectory prediction system based on bird detection radar and big data model analysis, including a data acquisition module 100, a prediction model building module 200 and a prediction model updating module 300.

[0017] The data acquisition module 100 acquires bird-detecting radar data, which includes bird speed, bird acceleration, bird trajectory and time point, wherein the bird speed represents the speed of the bird at each time point, which can usually be a vector, including size and direction; the bird acceleration represents the rate of change of the bird speed; the bird trajectory represents the position information of the bird at each time point; the time point represents the specific time corresponding to the data. In the bird-detecting radar data, the role and effect of the time point is to ensure the timing and continuity of the data, which is crucial for the construction and optimization of the prediction model; the time point not only provides the instantaneous state of the bird's movement, but also is the key node connecting the entire data stream; in the process of establishing the prediction model, the time point is used to divide the data set, and through the sliding window technology, the model can construct a prediction of the future bird trajectory in a continuous time segment; this method enables the model to extract patterns from known data to predict unseen data; this not only improves the prediction ability of the model, but also helps optimize the model parameters to cope with different changes in motion behavior; in addition, the time point is also used to record and calculate the time required for prediction, which is used to evaluate the efficiency of the model and is an important component in the calculation of the influence coefficient.

[0018] The initial prediction unit 201 in the prediction model building module 200 sets a trajectory prediction model. The trajectory prediction model is trained according to the bird trajectories at different time points in the bird detection radar data, and the time required for the bird trajectory at the next time point predicted by the trajectory prediction model is recorded as the model prediction time, which specifically includes: The bird radar data is divided into multiple time segments. A sliding window is created. Each sliding window contains a series of continuous time points. The bird speed, bird acceleration and bird trajectory at the previous time points in the bird radar data are used as the input of the trajectory prediction model. That is, the previous time point data of the last time point data in the sliding window is used as the input of the trajectory prediction model. For example, assuming there is a time point , for a window size of 3, Indicates the previous time point of the last time point data in the sliding window. Indicates the last time point in the sliding window; The trajectory prediction model is used to predict the bird's trajectory at the next time point, the prediction result of the trajectory prediction model is output, and the time required for the trajectory prediction model to predict the bird's trajectory at the next time point is recorded as the model prediction time.

[0019] The error optimization unit 202 in the prediction model building module 200 predicts the bird trajectory at the next time point based on the bird speed, bird acceleration and bird trajectory at the previous time point in the bird detection radar data using the trajectory prediction model, and calculates the error, wherein the error is calculated based on the bird trajectory at the next time point predicted by the trajectory prediction model and the bird trajectory of the bird detection radar data at that time point; and the training strategy in the trajectory prediction model is used to optimize the error to update the prediction parameters in the trajectory prediction model, specifically including: Calculate the error between the bird trajectory predicted by the trajectory prediction model at the next time point and the bird trajectory of the bird detection radar data at that time point. Use the gradient descent training strategy to optimize the error. By counting the mean square error of the error and calculating the partial derivative of the mean square error, use the chain rule to calculate the gradient of the mean square error to the parameters in the trajectory prediction model. The calculation formula of the mean square error is as follows: ,in It indicates that the trajectory prediction model predicts the bird's trajectory at the next time point; Indicates the bird trajectory of the bird detection radar data at this time point, specifically the bird trajectory at the last time point in the sliding window; Indicates the number of samples of bird radar data; Relying on the optimizer, the calculated gradient is used to update the prediction parameters of the trajectory prediction model. Through batch processing of multiple training cycles, the prediction parameters of the trajectory prediction model are continuously updated to minimize the error. The prediction parameter adjustment formula is as follows: For the prediction parameters , ,in, is the learning rate, is the mean square error with respect to the parameter gradient.

[0020] The influence coefficient calculation unit 301 in the prediction model update module 300 calculates the influence coefficient according to the error sum calculated each time and the recorded model prediction time, specifically including: The influence coefficient is calculated based on the error of each calculation and the recorded model prediction time. The calculation formula is: ,in represents the influence coefficient; Represents the error of each calculation, ; The model prediction time for each record, that is, the output the time used; is a factor used to normalize the impact coefficient to ensure that it is within a reasonable range.

[0021] The model updating unit 302 in the prediction model updating module 300 updates the training strategy in the trajectory prediction model according to the influence coefficient, by updating the training strategy in the gradient descent method. Make adjustments, and the adjustment formula is as follows: , which means that if the influence coefficient is higher (such as large error or long prediction time), the learning rate is reduced, otherwise it is increased; Optimizing the learning process by collecting errors and model prediction time helps the model remain stable and efficient in complex situations. In specific scenarios, targeted optimization can be performed based on errors and time, such as accelerating convergence at key time nodes. The core role of calculating the influence coefficient and updating the training strategy of the trajectory prediction model based on it is to achieve automatic adjustment during dynamic training, which is a significant improvement over the traditional static training strategy. The influence coefficient provides a way to balance the accuracy and training efficiency of the model by combining the error to measure the current performance of the model at a specific data point, and the time reflects the consumption of the model's computing resources. Specifically, when the influence coefficient is high, it means that the model has significant errors in prediction and may take a long time. At this time, slowing down the learning rate helps to fine-tune the model parameters, thereby improving accuracy and stability, and avoiding overfitting or instability due to large changes. When the influence coefficient is low, the learning rate can be appropriately increased to speed up the optimization speed, especially when the model has roughly converged, which helps to shorten the training time. Overall, this adaptive training strategy enhances the adaptability of the model in various training scenarios, allowing the model to not only maintain good performance in simple and complex data input environments, but also to efficiently predict in uncertain scenarios.

[0022] Adjusting the training strategy through the influence coefficient has brought significant effects on the prediction accuracy and computational efficiency of the model, and its most direct manifestation is in the stability and convergence of the model; first, by using the method of dynamically adjusting the learning rate, the model can better adapt to different training stages and data characteristics. For example, in the early stage, a higher learning rate helps to quickly approach the optimal solution, while in the later stage of training, reducing the learning rate can prevent the violent oscillation of the model parameters, thereby more accurately fine-tuning the model and improving the accuracy of the final model. Secondly, by combining error and time indicators, the model training process becomes more intelligent, and can actively identify and adapt to data or tasks of different complexity, which is especially important in real-time or variable application scenarios; in addition, this mechanism also improves resource utilization efficiency. By optimizing the learning rate, unnecessary training time is shortened, and excessive computing pressure is reduced, which ultimately leads to faster processing speed, which not only reduces the training cost, but also enables the model to maintain an efficient learning process under the constraints of software and hardware resources; overall, the use of influence coefficients not only improves the prediction performance of the model, but also makes it more adaptable to the needs and limitations of actual application scenarios.

[0023] Finally, the prediction model updating module 300 uses the updated trajectory prediction model to predict the trajectory of the flying bird at a future time point in the bird detection radar data.

[0024] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A bird trajectory prediction system based on bird detection radar and big data model analysis, characterized in that: The method comprises a data acquisition module (100), a prediction model establishment module (200) and a prediction model update module (300), wherein: The data acquisition module (100) acquires bird detection radar data, the bird detection radar data including bird speed, bird acceleration, bird trajectory and time point; The prediction model establishment module (200) sets a trajectory prediction model, and the trajectory prediction model is trained according to the bird trajectories at different time points in the bird detection radar data; according to the bird speed, bird acceleration and bird trajectory at the previous time point in the bird detection radar data, the trajectory prediction model is used to predict the bird trajectory at the next time point, and the error is calculated, wherein the error is calculated based on the bird trajectory predicted by the trajectory prediction model at the next time point and the bird trajectory of the bird detection radar data at the time point; and the training strategy in the trajectory prediction model is used to optimize the error to update the prediction parameters in the trajectory prediction model, and the time required for the trajectory prediction model to predict the bird trajectory at the next time point is recorded as the model prediction time; The prediction model updating module (300) calculates the influence coefficient according to the error sum calculated each time and the recorded model prediction time, and updates the training strategy in the trajectory prediction model according to the influence coefficient; and uses the updated trajectory prediction model to predict the trajectory of the flying bird at a future time point in the bird detection radar data.

2. The bird trajectory prediction system based on bird detection radar and big data model analysis according to claim 1 is characterized in that: The prediction model building module (200) comprises an initial prediction unit (201), and the process of recording the model prediction time by the initial prediction unit (201) specifically comprises: The bird-detecting radar data is divided into multiple time segments, and a sliding window is created. Each sliding window contains a series of continuous time points. The bird speed, bird acceleration and bird trajectory at the previous time points in the bird-detecting radar data are used as the input of the trajectory prediction model, that is, the previous time point data of the last time point data in the sliding window is used as the input of the trajectory prediction model; the trajectory prediction model is used to predict the bird trajectory at the next time point, the prediction result of the trajectory prediction model is output, and the time required for the trajectory prediction model to predict the bird trajectory at the next time point is recorded as the model prediction time.

3. The bird trajectory prediction system based on bird detection radar and big data model analysis according to claim 2 is characterized in that: The prediction model building module (200) comprises an error optimization unit (202), and the error optimization unit (202) performs error optimization specifically including: Calculate the error between the bird trajectory predicted by the trajectory prediction model at the next time point and the bird trajectory of the bird detection radar data at that time point. Use the gradient descent training strategy to optimize the error. By counting the mean square error of the error and calculating the partial derivative of the mean square error, use the chain rule to calculate the gradient of the mean square error to the parameters in the trajectory prediction model. The calculation formula of the mean square error is as follows: ,in It indicates that the trajectory prediction model predicts the bird's trajectory at the next time point; Indicates the bird trajectory of the bird detection radar data at this time point, specifically the bird trajectory at the last time point in the sliding window; Indicates the number of samples of bird radar data; Relying on the optimizer, the calculated gradient is used to update the prediction parameters of the trajectory prediction model. Through batch processing of multiple training cycles, the prediction parameters of the trajectory prediction model are continuously updated to minimize the error. The prediction parameter adjustment formula is as follows: For the prediction parameters , ,in, is the learning rate, is the mean square error with respect to the parameter gradient.

4. The bird trajectory prediction system based on bird detection radar and big data model analysis according to claim 3 is characterized in that: The prediction model updating module (300) comprises an influence coefficient calculation unit (301), and the process of calculating the influence coefficient by the influence coefficient calculation unit (301) specifically comprises: The influence coefficient is calculated based on the error of each calculation and the recorded model prediction time. The calculation formula is: ,in represents the influence coefficient; Represents the error of each calculation, ; The model prediction time for each record, that is, the output the time used; is a factor used to normalize the influence coefficient.

5. The bird trajectory prediction system based on bird detection radar and big data model analysis according to claim 4 is characterized in that: The prediction model updating module (300) comprises a model updating unit (302), and the process of the model updating unit (302) performing model updating specifically comprises: The training strategy in the trajectory prediction model is updated according to the influence coefficient, and the training strategy of the gradient descent method is updated. Make adjustments, and the adjustment formula is as follows: , which means that if the influence coefficient is higher, the learning rate is reduced, otherwise it is increased.

Images