Unmanned aerial vehicle flight path optimization system and method based on artificial intelligence and Internet of Things

Abstract

The invention discloses an unmanned aerial vehicle flight path optimization system and method based on artificial intelligence and Internet of Things, and relates to the technical field of unmanned aerial vehicle control, and the method comprises the steps: collecting communication delay, environment data and an unmanned aerial vehicle state in real time; analyzing historical communication delay data by using a Markov chain, and predicting a basic delay value in a short time in the future; correcting the basic delay value through a machine learning model; outputting a final delay prediction result; calculating a delay risk index, and judging a delay risk level; based on the delay risk level, when the delay risk level is low risk, sending a control instruction in advance according to the final delay prediction result; and when the delay risk level is medium or high risk, calling a pre-trained reinforcement learning strategy library, and generating a self-adaptive control instruction. According to the method, the multi-dimensional risk index is constructed, multiple strategies are provided to adjust the flight path according to the real-time data and the risk index, and the flight stability of the unmanned aerial vehicle is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) control, and specifically to a UAV flight trajectory optimization system and method based on artificial intelligence and the Internet of Things. Background Art

[0002] With the rapid development of UAV technology, UAVs are increasingly widely used in various industries, especially in fields such as logistics distribution, agricultural monitoring, and environmental monitoring. The optimization of UAV flight trajectories is one of the important technologies to ensure the efficient and safe flight of UAVs. Currently, most UAV flight control systems rely on traditional preset flight paths and fixed control instructions, lacking sufficient intelligence and adaptability. Especially in complex environments, UAVs often face problems such as communication delays, environmental interference, and unstable flight states, which affect flight stability and efficiency.

[0003] Traditional communication technologies have limitations. The Wi-Fi transmission range is limited, and although the 4G / 5G cellular network has a far transmission range, there are communication delays. Especially, the control link delay can pose a safety hazard. The environmental data and the UAV's own state data are complex and variable, and existing methods are difficult to comprehensively consider the impact of these factors on communication delays, resulting in inaccurate delay prediction. Regarding the risks brought by communication delays, there is a lack of effective evaluation and response mechanisms, and it is impossible to provide precise control strategies according to different risk levels, making it difficult to ensure the safe flight and mission execution of UAVs in complex situations. Summary of the Invention

[0004] The purpose of the present invention is to provide a UAV flight trajectory optimization system and method based on artificial intelligence and the Internet of Things to solve the problems proposed in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A UAV flight trajectory optimization method based on artificial intelligence and the Internet of Things, the method comprising: Real-time collecting communication delays, environmental data, and UAV states; Using a Markov chain to analyze historical communication delay data to predict the basic delay value in the short term in the future; analyzing environmental data and UAV states through a machine learning model to correct the predicted basic delay value; combining the basic delay value and the correction result to output the final delay prediction result; Calculating a delay risk index and judging the delay risk level; Based on the delay risk level, when the delay risk level is low risk, sending control instructions in advance according to the final delay prediction result; when the delay risk level is medium or high risk, calling a pre-trained reinforcement learning strategy library to generate adaptive control instructions.

[0006] According to the above solution, the use of Markov chain to analyze historical communication delay data and predict the basic delay value in the short term in the future includes: Collect historical communication delay data, arrange them in chronological order, and construct a continuous historical communication delay time series data set; according to the distribution characteristics of the historical communication delay data, divide the continuous delay values into several discrete state intervals with clear boundaries, and each interval represents a specific delay range level; Statistically analyze the transition frequencies between different delay states in the historical data, calculate the transition probabilities between each pair of states, and construct a Markov state transition probability matrix; Map the real-time collected communication delay to the corresponding discrete state in the Markov state transition probability matrix as the initial state of the Markov chain; based on the initial state and the state transition probability matrix, iteratively calculate the delay state distribution in the short term in the future, and deduce the delay state probability distribution in the short term in the future through multiple state transition simulations; Convert the predicted discrete state probability distribution into a specific delay value range, and use a time series smoothing algorithm to optimize the prediction result, effectively reducing the prediction fluctuation introduced by state discretization, and output the basic delay value.

[0007] According to the above solution, the analysis of environmental data and UAV state by a machine learning model to correct the predicted basic delay value includes: The environmental data includes real-time wind speed, temperature gradient and terrain; the UAV state includes flight altitude, horizontal speed and remaining battery power; Convert the environmental data and UAV state into environmental features and UAV state features; Based on the environmental features and UAV state features, construct a multi-dimensional input vector; input the multi-dimensional input vector into a pre-trained gradient boosting decision tree model, and output a delay correction amount through feature importance weighted analysis. The delay correction amount includes a positive correction amount and a negative correction amount; Perform dynamic weighted fusion on the basic delay prediction value and the delay correction amount, where the weight coefficient is dynamically adjusted according to the confidence of the gradient boosting decision tree model and the feature stability; the confidence of the gradient boosting decision tree model is obtained through the prediction error of the gradient boosting decision tree model; the feature stability is evaluated through the variance of feature importance within a sliding window; Record the deviation between the actual communication delay and the prediction result. When it continuously exceeds the set threshold, trigger the incremental learning update mechanism of the gradient boosting decision tree model to adaptively optimize the correction parameters.

[0008] According to the above solution, the conversion of the environmental data and UAV state into environmental features and UAV state features includes: Calculate the mean and standard deviation of the wind speed data within a sliding time window to generate wind speed stability features; calculate the linear regression of the temperature data and the real-time altitude data to extract the vertical temperature gradient features; quantify the complexity features of the terrain and landform based on the terrain and landform. Calculate the height change acceleration feature of the flight altitude; through coordinate transformation, decompose the horizontal speed into longitudinal and lateral speed components; fit a quadratic curve to the battery power time series data to extract the discharge acceleration feature.

[0009] According to the above solution, combining the basic delay value and the correction result, and outputting the final delay prediction result, including: Apply a correction amount constraint to the basic delay value to generate a preliminary fusion result, and distinguish the positive compensation and negative compensation intervals according to the directional characteristics of the correction amount; Perform credibility grading on the fusion result through a dynamic confidence evaluation mechanism, and the credibility grading includes calculating the mean square error between the predicted value and the actual delays in the last 3 times, and generating a confidence score in combination with the feature stability index; Apply a sliding time window smoothing algorithm to eliminate the short-term fluctuations of the prediction result, and perform logical verification on the abnormal prediction values in combination with the current network topology state to generate a delay prediction range with a confidence interval; When the confidence of the prediction result is higher than the set threshold, directly output the final predicted value; when the confidence is lower than the threshold, trigger the redundant communication link verification mechanism; The redundant communication link verification mechanism includes that when the delay prediction confidence of the primary communication link is lower than the set threshold, at least one standby communication link is automatically activated; probe data packets are synchronously sent through the primary and standby links and real-time communication delay data is collected; weighted fusion calculation is performed on the multi-link measurement results, and the weights in the weighted fusion calculation are dynamically allocated according to the link signal strength, historical stability and environmental adaptability; compare the fusion result with the original predicted value, and when the difference exceeds the fault tolerance threshold, then adopt the verification value and mark the primary link as abnormal, and update the communication strategy at the same time.

[0010] According to the above solution, calculating the delay risk index and judging the delay risk level, including: Based on the final delay prediction result, calculate the deviation degree of the predicted delay from the safety threshold, count the fluctuation range of the delay prediction results in the recent time window, extract the environmental features and UAV state features, and construct a multi-dimensional risk factor set, where the multi-dimensional risk factor set includes the deviation degree, the fluctuation range, the environmental features and the UAV features; and dynamically adjust the weight allocation ratio according to the current task type of the UAV, and output the standardized risk index value. Set multiple levels of delay risk threshold intervals, and map the calculated risk index values to the corresponding delay risk levels; the delay risk levels include low risk, medium risk, and high risk; The low risk means that the predicted delay does not exceed the safety threshold and the fluctuation is gentle, the medium risk means that the predicted delay exceeds the safety threshold but the backup link is available, and the high risk means that the predicted delay exceeds the safety threshold and there is no reliable backup link.

[0011] According to the above solution, when the delay risk level is low risk, send a control instruction in advance according to the final delay prediction result, including: Determine the basic compensation time according to the final delay prediction value, adjust the compensation amount in combination with the current motion state of the drone. The higher the flight speed, the longer the compensation time to ensure the timeliness of the instruction during high-speed flight; add a timestamp and a sequence number mark to the control instruction. After receiving the control instruction, the drone executes the instruction according to the timestamp and the sequence number mark to avoid control errors caused by out-of-order or delay; at the same time, monitor the instruction execution status, record the matching degree between the instruction sending time and the predicted communication delay, monitor the actual instruction execution time of the drone, and compare the deviation between the actual instruction execution time and the expected time, and synchronously update the deviation data to the predicted error accumulation record; continuously optimize the prediction accuracy through incremental learning by accumulating prediction error data.

[0012] According to the above solution, when the delay risk level is medium or high risk, call the pre-trained reinforcement learning policy library to generate an adaptive control instruction, including: According to the current risk level, and in combination with the real-time collected environmental data and the drone state, match the candidate policies corresponding to the risk level from the pre-trained reinforcement learning policy library; evaluate the expected benefits and execution costs of the candidate policies, and select the candidate policy with the highest comprehensive score as the preliminary policy; the expected benefits include communication reliability, flight safety, and mission effectiveness, and the execution costs include energy consumption and operation risks; Integrate and encode the environmental data and the drone state, and convert them into the standard input vector of the reinforcement learning model; based on the standard input vector, use the reinforcement learning model to generate a preliminary control instruction sequence, and combine the dynamic constraints to perform a feasibility check on the preliminary control instruction; Apply dynamic constraints and obstacle avoidance rules to the instructions that pass the feasibility check; the dynamic constraints include maximum acceleration and angular velocity limits; the obstacle avoidance rules include maintaining a minimum safe distance from obstacles; Output an adaptive control instruction with a safety margin; the setting of the safety margin takes into account various uncertainty factors and potential risks, provides additional safety guarantees for the flight of the drone, and ensures that the drone can still execute tasks safely and stably in a complex and changing environment.

[0013] According to the above solution, the pre-trained reinforcement learning policy library includes: Establish a UAV dynamics model and a communication environment model to simulate various risk scenarios; use a deep reinforcement learning algorithm to generate policies through virtual trial-and-error training; Classify and store policies according to the risk level. Medium-risk policies focus on communication recovery, and high-risk policies focus on emergency avoidance; the medium-risk policies include communication frequency band switching policies, flight speed adjustment policies, and data transmission compression policies; the high-risk policies include emergency obstacle avoidance maneuver policies, forced return policies, and emergency landing policies; During the process of applying the policies, record the execution effects of the policies and dynamically update the policy priorities.

[0014] A UAV flight trajectory optimization system based on artificial intelligence and the Internet of Things. The system includes: a data acquisition module, a delay prediction module, a risk assessment module, and a control decision module; The data acquisition module includes a communication data unit, an environmental data unit, and a UAV status unit; the communication data unit is used to collect communication delay, packet loss rate, jitter situation, signal strength index, signal-to-noise ratio, frequency band information, and retransmission information; the environmental data unit is used to collect real-time wind speed, temperature gradient, and terrain; the UAV status unit is used to collect the UAV flight altitude, horizontal speed, and remaining battery power; The delay prediction module includes a prediction unit, a correction unit, and a result fusion unit; the prediction unit generates a basic prediction value based on historical communication delay data; the correction unit dynamically corrects the basic prediction value through environmental data and UAV status characteristics; the result fusion unit weights and integrates the basic prediction value and the correction amount, and outputs the final delay prediction result; The risk assessment module is used to quantify the communication delay risk level and trigger a hierarchical response policy; The control decision module includes a low-risk control unit, a reinforcement learning policy library, and an instruction verification unit; the communication delay compensation unit, when the risk assessment unit assesses that the communication delay risk is low risk, sends a control instruction in advance according to the final delay prediction result; the reinforcement learning policy library, when the delay risk level is medium or high risk, calls the pre-trained reinforcement learning policy library to generate an adaptive control instruction; the instruction verification unit verifies the feasibility of the instruction through dynamic constraints and obstacle avoidance rules.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. The present invention analyzes historical communication delay data to predict the basic delay value, and combines a machine learning model to correct it according to environmental and UAV status data, and can accurately predict communication delay; 2. The present invention constructs a multi-dimensional risk index, and based on real-time data and the risk index, provides a variety of strategies to adjust the flight trajectory, improving the flight stability of the drone. 3. The present invention performs dynamic constraints and obstacle avoidance rule checks on control instructions, outputs instructions with safety margins, reduces flight accidents caused by unreasonable instructions, and enhances the flight reliability of the drone. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flowchart of the steps of the method for optimizing the flight trajectory of a drone based on artificial intelligence and the Internet of Things according to the present invention; Figure 2 is a schematic structural diagram of the system for optimizing the flight trajectory of a drone based on artificial intelligence and the Internet of Things according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] Embodiment: As Figure 1 - Figure 2 shown, the present invention provides a technical solution, a method for optimizing the flight trajectory of a drone based on artificial intelligence and the Internet of Things, and the method includes the steps: S1. Real-time collect communication delay, environmental data, and drone status; Specifically, the environmental data includes real-time wind speed, temperature gradient, and terrain and landform; the drone status includes flight altitude, horizontal speed, and remaining battery power; for example, the collected data: the current link is the 4G main link, and the last three communication delays of the 4G main link are 120 ms, 150 ms, and 110 ms respectively; the real-time wind speed is 8.2 m / s, the temperature gradient is -6.5 °C / 100 m, the terrain and landform quantization value is 0.78; the flight altitude is 152 m, the horizontal speed is 12 m / s (decomposed into a longitudinal speed of 10.2 m / s and a lateral speed of 6.5 m / s), and the battery power is 65%. This is only for illustrative purposes and is not limited.

[0019] S2. Use the Markov chain to analyze historical communication delay data and predict the basic delay value in the short term in the future; Specifically, collect historical communication delay data, arrange it in chronological order, and construct a continuous historical communication delay time series data set; according to the distribution characteristics of the historical communication delay data, divide the continuous delay values into several discrete state intervals with clear boundaries, and each interval represents a specific delay range level. For example, discretize the historical delay data into 3 state intervals, where State 1: [100ms, 130ms), State 2: [130ms, 160ms), State 3: ≥160ms; count the transition frequencies between different delay states in the historical data, calculate the transition probabilities between each pair of states, and construct a Markov state transition probability matrix; map the real-time collected communication delay to the corresponding discrete state in the Markov state transition probability matrix as the initial state of the Markov chain; based on the initial state and the state transition probability matrix, iteratively calculate the delay state distribution in the future short time, and through multiple state transitions simulation, deduce the delay state probability distribution in the future short time; convert the predicted discrete state probability distribution into a specific delay value range, and use a time series smoothing algorithm to optimize the prediction result, effectively reducing the prediction fluctuation introduced by state discretization, and output the basic delay value. For example: the current communication delay is 125ms, mapped to State Interval 1, the predicted next state probability is 60% for State 1, and the predicted communication delay range is [100ms, 130ms). After smoothing, the output basic delay value is 118ms.

[0020] S3. Analyze the environmental data and the UAV state through a machine learning model to correct the predicted basic delay value; Specifically, based on the environmental characteristics and UAV state characteristics, construct a multi-dimensional input vector. For example: [wind speed stability = 0.15, temperature gradient = -6.5, terrain complexity = 0.78, altitude acceleration = 0.12, longitudinal speed = 10.2, discharge acceleration = -1.2]. This is only for illustrative purposes and not for limitation; input the multi-dimensional input vector into a pre-trained gradient boosting decision tree model, and through feature importance weighted analysis, output the delay correction amount, where the delay correction amount includes a positive correction amount and a negative correction amount; perform dynamic weighted fusion on the basic delay prediction value and the delay correction amount, where the weight coefficient is dynamically adjusted according to the gradient boosting decision tree model confidence and feature stability; the gradient boosting decision tree model confidence is obtained through the prediction error of the gradient boosting decision tree model; the feature stability is evaluated through the variance of feature importance within a sliding window; record the deviation between the actual communication delay and the prediction result, and when it continuously exceeds the set threshold, trigger the incremental learning update mechanism of the gradient boosting decision tree model to adaptively optimize the correction parameters; Further, calculate the mean and standard deviation of the wind speed data within a sliding time window to generate a wind speed stability feature; calculate the linear regression of the temperature data and the real-time altitude data to extract a vertical temperature gradient feature; quantify the complexity feature of the terrain and landform based on the terrain and landform; calculate the altitude change acceleration feature of the flight altitude; decompose the horizontal speed into longitudinal and lateral speed components features through coordinate transformation; fit a quadratic curve to the battery power time series data to extract a discharge acceleration feature.

[0021] S4. Combine the basic delay value and the correction result to output a final delay prediction result; Further, impose a correction amount constraint on the basic delay value to generate a preliminary fusion result, and distinguish the positive compensation and negative compensation intervals according to the directional feature of the correction amount; perform credibility grading on the fusion result through a dynamic confidence evaluation mechanism, and the credibility grading includes calculating the mean square error between the predicted value and the last 3 actual delays, and generating a confidence score in combination with a feature stability index; apply a sliding time window smoothing algorithm to eliminate the short-term fluctuations of the prediction result, and perform logical verification on the abnormal prediction value in combination with the current network topology state to generate a delay prediction range with a confidence interval; when the confidence of the prediction result is higher than the set threshold, directly output the final predicted value; when the confidence is lower than the threshold, trigger a redundant communication link verification mechanism; the redundant communication link verification mechanism includes that when the delay prediction confidence of the primary communication link is lower than the set threshold, automatically activate at least one standby communication link; send probe data packets synchronously through the primary and standby links and collect real-time communication delay data; perform weighted fusion calculation on the multi-link measurement results, and the weights in the weighted fusion calculation are dynamically allocated according to the link signal strength, historical stability, and environmental adaptability; compare the fusion result with the original predicted value, and when the difference exceeds the fault tolerance threshold, then adopt the verification value and mark the primary link as abnormal, and update the communication strategy at the same time.

[0022] S5. Calculate the delay risk index and judge the delay risk level; Specifically, based on the final delay prediction result, calculate the deviation degree between the predicted delay and the safety threshold, statistically analyze the fluctuation range of the delay prediction results in the most recent time window, extract environmental features and UAV state features, and construct a multi-dimensional risk factor set, which includes the deviation degree, the fluctuation range, environmental features, and UAV features; and dynamically adjust the weight allocation ratio according to the current task type of the UAV, and output the risk index value after standardized processing; set multiple levels of delay risk threshold intervals, and map the calculated risk index value to the corresponding delay risk level; for example: through the calculation of the delay risk index, it is obtained that the deviation from the safety threshold is -10 ms, the fluctuation range is the standard deviation of the delay within the sliding window of 18 ms, the environmental risk is terrain complexity and wind speed, the output delay risk index is 0.35, and the delay risk level is low risk; the delay risk levels include low risk, medium risk, and high risk; the low risk means that the predicted delay does not exceed the safety threshold and the fluctuation is gentle, the medium risk means that the predicted delay exceeds the safety threshold but the backup link is available, and the high risk means that the predicted delay exceeds the safety threshold and there is no reliable backup link.

[0023] S6. When the delay risk level is low risk, send a control instruction in advance according to the final delay prediction result; Specifically, when the delay risk level is low risk, sending a control instruction in advance according to the final delay prediction result includes: determining the basic compensation time according to the final delay prediction value, adjusting the compensation amount in combination with the current motion state of the UAV, the higher the flight speed, the longer the compensation time, to ensure the timeliness of the instruction during high-speed flight; adding a timestamp and a sequence number mark to the control instruction, and after receiving the control instruction, the UAV executes the instruction according to the timestamp and the sequence number mark to avoid control errors caused by out-of-order or delay; for example: send a pitch angle adjustment instruction 140 ms in advance, and the timestamp is +140 ms; at the same time, monitor the instruction execution status, record the matching degree between the instruction sending time and the predicted communication delay, monitor the actual instruction execution time of the UAV, and compare the deviation between the actual instruction execution time and the expected time, and synchronously update the deviation data to the predicted error cumulative record; for example: the actual execution deviation is +5 ms, and it is updated to the predicted error cumulative record; by continuously accumulating and analyzing the predicted error data, continuously optimize the accuracy of the communication delay prediction.

[0024] S7. When the delay risk level is medium or high risk, call the pre-trained reinforcement learning policy library to generate an adaptive control instruction; Specifically, when the delay risk level is medium or high risk, a pre-trained reinforcement learning policy library is called to generate an adaptive control instruction, including: For example, when encountering a strong wind of 15 m / s, the communication delay becomes 210 ms, and the delay risk level is high risk; according to the current risk level, combined with the real-time collected environmental data and the UAV state, a candidate policy corresponding to the risk level is matched from the pre-trained reinforcement learning policy library; the expected benefits and execution costs of the candidate policy are evaluated, and the candidate policy with the highest comprehensive score is selected as the preliminary policy; For example, the matched candidate policies are Policy A: forced return, and Policy B: emergency altitude reduction; the candidate policies are evaluated, and the evaluation score of Policy B is higher, so Policy B is selected; the expected benefits include communication reliability, flight safety and mission effectiveness, and the execution costs include energy consumption and operation risk; the environmental data and the UAV state are integrated and encoded into a standard input vector of the reinforcement learning model; based on the standard input vector, using the reinforcement learning model, a preliminary control instruction sequence is generated, and combined with the dynamic constraints, the feasibility of the preliminary control instruction is verified; the dynamic constraints and obstacle avoidance rules are applied to the instructions that pass the feasibility verification; For example, after selecting Policy B, a descent instruction is generated, and the dynamic constraint is a maximum descent acceleration of 2 m / s², and the obstacle avoidance rule is to ensure a distance greater than 50 m from the mountain; the dynamic constraints include maximum acceleration and angular velocity limits; the obstacle avoidance rules include maintaining a minimum safe distance from obstacles; an adaptive control instruction with a safety margin is output; the setting of the safety margin takes into account various uncertainty factors and potential risks, provides additional safety protection for the flight of the UAV, and ensures that the UAV can still perform tasks safely and stably in a complex and changeable environment.

[0025] Furthermore, a UAV dynamics model and a communication environment model are established to simulate various risk scenarios; a deep reinforcement learning algorithm is used to generate policies through virtual trial-and-error training; the policies are classified and stored according to the risk level, with medium-risk policies focusing on communication recovery and high-risk policies focusing on emergency avoidance; the medium-risk policies include communication frequency band switching policies, flight speed adjustment policies and data transmission compression policies, and the high-risk policies include emergency obstacle avoidance maneuver policies, forced return policies and emergency landing policies; during the application of the policies, the execution effects of the policies are recorded, and the policy priorities are dynamically updated; For example, after executing Policy B, the execution effect of the policy is recorded as the delay is reduced to 170 ms, the mission progress is delayed by 8%, and the delay risk level is still high risk, and the desired effect is not achieved, so the policy priority is updated, and the weight of Policy A in the same type of scenario is increased by 15%.

[0026] The present invention provides another technical solution, a UAV flight trajectory optimization system based on artificial intelligence and the Internet of Things, which includes: a data acquisition module, a delay prediction module, a risk assessment module and a control decision module; The data acquisition module includes a communication data unit, an environmental data unit, and a UAV status unit; the communication data unit is used to collect communication delay, packet loss rate, jitter condition, signal strength index, signal-to-noise ratio, frequency band information, and retransmission information; the environmental data unit is used to collect real-time wind speed, temperature gradient, and terrain and landform; the UAV status unit is used to collect UAV flight altitude, horizontal speed, and remaining battery power; The delay prediction module includes a prediction unit, a correction unit, and a result fusion unit; the prediction unit generates a basic prediction value based on historical communication delay data; the correction unit dynamically corrects the basic prediction value through environmental data and UAV status characteristics; the result fusion unit weights and integrates the basic prediction value and the correction amount, and outputs the final delay prediction result; The risk assessment module is used to quantify the communication delay risk level and trigger a hierarchical response strategy; The control decision module includes a low-risk control unit, a reinforcement learning policy library, and an instruction verification unit; the communication delay compensation unit, when the risk assessment unit assesses that the communication delay risk is low risk, sends a control instruction in advance according to the final delay prediction result; the reinforcement learning policy library, when the delay risk level is medium or high risk, calls the pre-trained reinforcement learning policy library to generate an adaptive control instruction; the instruction verification unit verifies the feasibility of the instruction through dynamic constraints and obstacle avoidance rules.

[0027] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed invention.

Claims

1. An unmanned aerial vehicle flight trajectory optimization method based on artificial intelligence and the Internet of Things, characterized in that: The method includes: Collecting communication delay, environmental data, and UAV status in real time; Analyzing historical communication delay data using a Markov chain to predict the basic delay value in the short term in the future; analyzing environmental data and UAV status through a machine learning model to correct the predicted basic delay value; combining the basic delay value and the correction result to output the final delay prediction result; Calculating the delay risk index and judging the delay risk level; Based on the delay risk level, when the delay risk level is low risk, send a control instruction in advance according to the final delay prediction result; when the delay risk level is medium or high risk, call a pre-trained reinforcement learning policy library to generate an adaptive control instruction.

2. The method for optimizing the flight trajectory of an unmanned aerial vehicle based on artificial intelligence and the Internet of Things according to claim 1, wherein: The step of analyzing historical communication delay data using a Markov chain to predict the basic delay value in the short term in the future includes: Collecting historical communication delay data, arranging it in chronological order, and constructing a continuous historical communication delay time series dataset; according to the distribution characteristics of historical communication delay data, dividing the continuous delay values into several discrete state intervals with clear boundaries, and each interval represents a specific delay range level; Statistically analyzing the transition frequencies between different delay states in historical data, calculating the transition probabilities between each pair of states, and constructing a Markov state transition probability matrix; Mapping the real-time collected communication delay to the corresponding discrete state in the Markov state transition probability matrix as the initial state of the Markov chain; based on the initial state and the state transition probability matrix, iteratively calculating the delay state distribution in the short term in the future, and deducing the delay state probability distribution in the short term in the future through multiple state transitions; Converting the predicted discrete state probability distribution into a specific delay value range, and using a time series smoothing algorithm to optimize the prediction result, effectively reducing the prediction fluctuation introduced by state discretization, and outputting the basic delay value.

3. The method for optimizing the flight trajectory of an unmanned aerial vehicle based on artificial intelligence and the Internet of Things according to claim 1, wherein: The step of analyzing environmental data and UAV status through a machine learning model to correct the predicted basic delay value includes: The environmental data includes real-time wind speed, temperature gradient, and terrain and landform; the UAV status includes flight altitude, horizontal speed, and remaining battery power; Converting the environmental data and UAV status into environmental features and UAV status features; Based on the environmental features and UAV status features, constructing a multi-dimensional input vector; inputting the multi-dimensional input vector into a pre-trained gradient boosting decision tree model, and outputting a delay correction amount through feature importance weighted analysis, where the delay correction amount includes a positive correction amount and a negative correction amount; Performing dynamic weighted fusion on the basic delay prediction value and the delay correction amount, where the weight coefficient is dynamically adjusted according to the confidence of the gradient boosting decision tree model and the feature stability; the confidence of the gradient boosting decision tree model is obtained through the prediction error of the gradient boosting decision tree model; the feature stability is evaluated through the variance of feature importance within a sliding window; Recording the deviation between the actual communication delay and the prediction result, and triggering the incremental learning update mechanism of the gradient boosting decision tree model to adaptively optimize the correction parameters when the deviation continuously exceeds the set threshold.

4. The method for optimizing the flight trajectory of a drone based on artificial intelligence and the Internet of Things according to claim 3, characterized in that: Converting the environmental data and UAV status into environmental features and UAV status features includes: Calculating the mean and standard deviation of the wind speed data within a sliding time window to generate a wind speed stability feature; calculating the linear regression of the temperature data and the real-time altitude data to extract a vertical temperature gradient feature; quantifying the complexity feature of the terrain based on the terrain and landform; Calculating the altitude change acceleration feature of the flight altitude; decomposing the horizontal speed into longitudinal and lateral speed components features through coordinate transformation; fitting a quadratic curve to the battery power time series data to extract a discharge acceleration feature.

5. The method for optimizing the flight trajectory of an unmanned aerial vehicle based on artificial intelligence and the Internet of Things according to claim 1, characterized in that: Combining the basic delay value and the correction result and outputting the final delay prediction result includes: Applying a correction amount constraint to the basic delay value to generate a preliminary fusion result, and distinguishing between the positive compensation and negative compensation intervals according to the directional feature of the correction amount; Performing credibility grading on the fusion result through a dynamic confidence evaluation mechanism; Applying a sliding time window smoothing algorithm to eliminate the short-term fluctuations of the prediction result, and combining the current network topology state to perform logical verification on abnormal prediction values to generate a delay prediction range with a confidence interval; When the confidence of the prediction result is higher than the set threshold, directly output the final prediction value; when the confidence is lower than the threshold, trigger a redundant communication link verification mechanism.

6. The method for optimizing the flight trajectory of a drone based on artificial intelligence and the Internet of Things according to claim 1, wherein: Calculating the delay risk index and judging the delay risk level includes: Based on the final delay prediction result, calculating the deviation degree of the predicted delay from the safety threshold, statistically analyzing the fluctuation range of the delay prediction results in the recent time window, extracting environmental features and UAV status features, and constructing a multi-dimensional risk factor set, where the multi-dimensional risk factor set includes the deviation degree, the fluctuation range, environmental features, and UAV features; and dynamically adjusting the weight allocation ratio according to the current task type of the UAV, and outputting the risk index value after standardized processing; Setting multiple levels of delay risk threshold intervals and mapping the calculated risk index value to the corresponding delay risk level; the delay risk level includes low risk, medium risk, and high risk.

7. The method for optimizing the flight trajectory of an unmanned aerial vehicle based on artificial intelligence and the Internet of Things according to claim 1, wherein: When the delay risk level is low risk, sending a control instruction in advance according to the final delay prediction result includes: Determining the basic compensation time according to the final delay prediction value, and adjusting the compensation amount in combination with the current motion state of the UAV. The higher the flight speed, the longer the compensation time, to ensure the timeliness of the instruction during high-speed flight; adding a time stamp and a sequence number mark to the control instruction. After receiving the control instruction, the UAV executes the instruction according to the time stamp and the sequence number mark to avoid control errors caused by out-of-order or delay; at the same time, monitoring the instruction execution status, recording the matching degree between the instruction sending time and the predicted communication delay, monitoring the actual instruction execution time of the UAV, and comparing the deviation between the actual instruction execution time and the expected time, and synchronously updating the deviation data to the predicted error accumulation record.

8. The method for optimizing the flight trajectory of an unmanned aerial vehicle based on artificial intelligence and the Internet of Things according to claim 1, wherein: When the delay risk level is medium or high risk, calling a pre-trained reinforcement learning policy library to generate an adaptive control instruction includes: Based on the current risk level, combined with environmental data and the UAV state, match candidate strategies corresponding to the risk level from the pre-trained reinforcement learning strategy library; evaluate the expected benefits and execution costs of the candidate strategies, and select the candidate strategy with the highest comprehensive score as the preliminary strategy; Integrate and encode the environmental data and the UAV state, and convert them into the standard input vector of the reinforcement learning model; based on the standard input vector, use the reinforcement learning model to generate a preliminary control instruction sequence, and combine the dynamic constraints to verify the feasibility of the preliminary control instructions; Apply dynamic constraints and obstacle avoidance rules to the instructions that pass the feasibility verification; the dynamic constraints include maximum acceleration and angular velocity limits; the obstacle avoidance rules include maintaining a minimum safe distance from obstacles; Output an adaptive control instruction with a safety tolerance.

9. The method for optimizing the flight trajectory of a drone based on artificial intelligence and the Internet of Things according to claim 8, characterized in that: The pre-trained reinforcement learning strategy library includes: Establish a UAV dynamics model and a communication environment model to simulate various risk scenarios; use a deep reinforcement learning algorithm to generate strategies through virtual trial-and-error training; Classify and store strategies according to risk levels. Medium-risk strategies focus on communication recovery, and high-risk strategies focus on emergency avoidance; the medium-risk strategies include communication frequency band switching strategies, flight speed adjustment strategies, and data transmission compression strategies; the high-risk strategies include emergency obstacle avoidance maneuver strategies, forced return strategies, and emergency landing strategies; During the process of applying strategies, record the execution effects of the strategies and dynamically update the strategy priorities.

10. An unmanned aerial vehicle flight trajectory optimization system based on artificial intelligence and the Internet of Things, characterized in that: The system includes: a data acquisition module, a delay prediction module, a risk assessment module, and a control decision module; The data acquisition module includes a communication data unit, an environmental data unit, and a UAV state unit; the communication data unit is used to collect communication delay, packet loss rate, jitter situation, signal strength index, signal-to-noise ratio, frequency band information, and retransmission information; the environmental data unit is used to collect real-time wind speed, temperature gradient, and terrain and landform; the UAV state unit is used to collect the UAV flight altitude, horizontal speed, and remaining battery power; The delay prediction module includes a prediction unit, a correction unit, and a result fusion unit; the prediction unit generates a basic prediction value based on historical communication delay data; the correction unit dynamically corrects the basic prediction value through environmental data and UAV state characteristics; the result fusion unit weights and integrates the basic prediction value and the correction amount, and outputs the final delay prediction result; The risk assessment module is used to quantify the communication delay risk level and trigger a hierarchical response strategy; The control decision module includes a low-risk control unit, a reinforcement learning strategy library, and an instruction verification unit; the communication delay compensation unit, when the risk assessment unit assesses that the communication delay risk is low risk, sends a control instruction in advance according to the final delay prediction result; the reinforcement learning strategy library, when the delay risk level is medium or high risk, calls the pre-trained reinforcement learning strategy library to generate an adaptive control instruction; the instruction verification unit verifies the feasibility of the instruction through dynamic constraints and obstacle avoidance rules.

Images