An intelligent aircraft turning control system

Abstract

The present invention belongs to the technical field of aviation flight control, and specifically discloses an intelligent aircraft turning regulation system, including: a data acquisition and preprocessing module, which collects flight data and preprocesses the collected flight data; a DQN intelligent optimization module, which outputs the best turning action a based on the aircraft state s to achieve an optimal turning strategy; the DQN intelligent optimization module is introduced into airborne equipment, and the DQN intelligent optimization module adjusts the control parameters of the aircraft to control the aircraft to complete the turning action. The present invention uses the Deep Q-Network (DQN) deep reinforcement learning algorithm to establish a real-time optimization model, which can adaptively calculate the optimal turning angle, bank angle, and turning rate based on multiple information such as the current state parameters of the aircraft, environmental factors, and route planning objectives, realizing safer and more efficient autonomous turning operations of the aircraft in complex environments.

Description

Technical Field

[0001] The present invention belongs to the technical field of aircraft flight control, and particularly relates to an intelligent aircraft turning control system. Background Art

[0002] The turning operation of an aircraft involves the coordinated work of multiple components and systems. For example, when using ailerons to make the aircraft bank and turn, coordinated operation with the rudder is required. Because when the aircraft banks, a yaw moment will be generated. If the rudder is not used for coordination, the aircraft will not only turn but also skid. In modern aircraft, the autopilot system can effectively assist the pilot in completing the turning operation, and can coordinate control surfaces such as ailerons, rudders, and elevators according to parameters such as the set course and turning angle for precise turning operations. However, under complex weather conditions or special flight states, it is difficult to achieve precise and safe turning, and individual differences in the pilot's control skills will also affect the turning efficiency and safety. Summary of the Invention

[0003] The purpose of the present invention is to overcome the defects existing in the prior art and provide an intelligent aircraft turning control system.

[0004] The present invention provides an intelligent aircraft turning control system, including:

[0005] A data acquisition and preprocessing module, which collects flight data from actual flight records, simulator simulations, or wind tunnel experiments, and preprocesses the collected flight data; the flight data includes flight performance parameters and corresponding turning operation details under different speeds, altitudes, weight distributions, and meteorological conditions, and the preprocessing includes removing outliers, filling in missing values, and converting the flight data format to make it suitable for a deep learning model;

[0006] A DQN intelligent optimization module, which outputs the best turning action a based on the aircraft state s to achieve an optimal turning strategy;

[0007] The DQN intelligent optimization module is imported into the airborne equipment, and the DQN intelligent optimization module adjusts the control parameters of the aircraft to control the aircraft to complete the turning action.

[0008] A further solution is that the DQN intelligent optimization module includes:

[0009] A Deep Q-Network model, which adjusts the control parameters of the aircraft to control the aircraft to complete the turning action;

[0010] An environment definition unit for defining the flight environment state, quantifying and discretizing the flight environment state into a finite state space S; S = {s_1, s_2,..., s_n}, where s_n represents the state at time t; and defining an action space A; A = {a_1, a_2,..., a_m}, where a_m represents the action taken at time t;

[0011] The flight environment includes the aircraft state and executable actions;

[0012] A reward function unit for constructing a reward function R(s, a, s') to measure the effect of reaching a new state s' after performing an action a in state s;

[0013] A Q-value function unit for obtaining the value evaluation Q(s, a) of action a when in state s;

[0014] An experience replay buffer unit for storing a large number of experiences (<s, a, r, s'>), where s is the current state, a is the action taken, r is the immediate reward received, and s' is the next state after performing the action.

[0015] In a further solution, the Q-value function unit uses a deep neural network as an approximator of the Q-value function. The input of the deep neural network is the state s, and the output is the value evaluation Q(s, a) of action a in state s;

[0016] The network architecture of the deep neural network adopts a multi-layer perceptron, and the last layer corresponds to the value evaluation Q(s, a) of each action.

[0017] In a further solution, the training process of the Deep Q-Network model includes:

[0018] Exploring the environment through an ε-greedy strategy, performing random actions or selecting the action with the highest current Q value;

[0019] After performing the action, obtaining the new state s' and the corresponding reward r, and storing the experience in the experience replay buffer unit;

[0020] Sampling a batch of data from the experience replay buffer unit for learning, updating the DQN network parameters to narrow the gap between the value evaluation Q(s, a) and the actual return.

[0021] In a further solution, the control system further includes: a dynamic bank-to-turn control module for continuously and smoothly adjusting the tilt of the aircraft body, reducing the turning radius and reducing the overload;

[0022] The dynamic bank-to-turn control module includes:

[0023] A motion model construction unit that establishes a motion model of an aircraft based on the principles of aircraft dynamics and aerodynamics. The motion model includes six-degree-of-freedom motion states and physical quantities related to the six-degree-of-freedom motion states. The motion model equations are as follows:

[0024] m*d^2x / dt^2 = F_x + m*g*sin(θ)

[0025] m*d^2y / dt^2 = F_y

[0026] I_z*d^2ψ / dt^2 = L_z

[0027] Where m is the mass, I_z is the moment of inertia, (x, y) are the position coordinates, ψ is the roll angle, (F_x, F_y) are the components of the resultant force in the xy plane, L_z is the moment about the z-axis, and θ is the pitch angle;

[0028] An index definition unit for defining performance indices. The quadratic form of the performance index J is:

[0029] J = ∫[Q(x, u) + R(u)]dt

[0030] Where x is the state vector, u is the control input vector, Q(x, u) is the state cost function, and R(u) is the control input cost function;

[0031] A turning dynamic adjustment unit that dynamically adjusts the attitude and thrust distribution of the aircraft during a turn based on the calculated control rate to achieve a banked turn action based on a preset performance index J;

[0032] A closed-loop control unit that obtains the output result of the Deep Q-Network model, generates an action instruction based on the output structure, and the airborne equipment receives and executes the instruction to form a feedback loop.

[0033] A further solution is that the control rate is a mathematical model for calculating the control input based on the error, and the mathematical model is a PID controller; the error is the result of comparing the current aircraft state parameters with the target set value.

[0034] A further solution is that the regulation system further includes a model real-time optimization module. Based on the closed-loop control principle, a closed-loop control system including multi-source sensor data fusion, error detection, real-time calculation, and instruction output is constructed. The closed-loop control is used to monitor the deviation between the actual turning process of the aircraft and the predetermined trajectory and immediately take compensation measures;

[0035] The model real-time optimization module includes:

[0036] A data fusion unit that integrates multi-source information through sensor fusion technology to generate a flight state estimate;

[0037] A feedback unit that obtains an error signal in real time and generates and sends a control instruction immediately after obtaining the error signal to control the input of the Deep Q-Network model, so as to continuously adjust the output of the Deep Q-Network model until it approaches or reaches the expected state;

[0038] A performance optimization unit that records key performance indicators in real time, evaluates the effect of the PID controller using the key indicators, and according to the evaluation results, adopts an online self-tuning algorithm or an offline parameter tuning method to continuously optimize the PID controller parameters to improve the control performance;

[0039] A security mechanism unit that sets a safety threshold. When the state variable exceeds the safety threshold, an alarm is triggered or the system enters a safe mode.

[0040] Furthermore, the control system further includes a fault detection module connected to the security mechanism unit. After an alarm is triggered or the system enters a safe mode, the security mechanism unit sends a fault code to the fault detection module. The fault detection module identifies the fault code and switches to a backup system or takes emergency measures to ensure flight safety.

[0041] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0042] The present invention uses the Deep Q-Network (DQN) deep reinforcement learning algorithm to establish a real-time optimization model. This model can adaptively calculate the optimal turning angle, bank angle, and turning rate based on multiple information such as the current state parameters of the aircraft, environmental factors, and route planning objectives, realizing safer and more efficient autonomous turning operations of the aircraft in complex environments.

[0043] The control system of the present invention adopts a three-dimensional space dynamic tilting technology, which can continuously and smoothly tilt and adjust the aircraft body. On the premise of ensuring flight stability and passenger comfort, it reduces the turning radius and overload, reduces fuel consumption during the turning process of the aircraft, thereby reducing the operating cost. It also constructs a closed-loop control system including multi-source sensor data fusion, error detection, real-time calculation, and instruction output, which can real-time monitor the deviation between the actual turning process of the aircraft and the predetermined trajectory and immediately take compensation measures to ensure that the aircraft can maintain the predetermined turning performance indicators at any flight stage, enhancing the intelligence level and market competitiveness of the aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The following drawings are only schematic illustrations and explanations of the present invention and are not used to limit the scope of the present invention, where:

[0045] Figure 1 : The principle block diagram of the control system of the present invention;

[0046] In the figure: 1, airborne equipment; 2, data acquisition and preprocessing module; 3, DQN intelligent optimization module; 4, dynamic bank-to-turn control module; 5, model real-time optimization module; 6, fault detection module; 7, environment definition unit; 8, Deep Q-Network model; 9, reward function unit; 10, Q-value function unit; 11, experience replay buffer unit; 12, deep neural network; 13, motion model construction unit; 14, closed-loop control unit; 15, index definition unit; 16, turn dynamic adjustment unit; 17, data fusion unit; 18, feedback unit; 19, performance optimization unit; 20, security mechanism unit. Specific implementation mode

[0047] In order to make the purpose, technical solution, design method and advantages of the present invention clearer, the present invention will be further described in detail below through specific embodiments with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0048] As Figure 1 shown, the present invention provides an intelligent aircraft turning control system, which improves the turning efficiency, accuracy and fuel economy of the aircraft through innovative design, and specifically includes:

[0049] A data acquisition and preprocessing module, which collects flight data in actual flight records, simulator simulations or wind tunnel experiments, and preprocesses the collected flight data; the flight data includes flight performance parameters and corresponding turning operation details under different speeds, altitudes, weight distributions and meteorological conditions, and the preprocessing includes removing outliers, filling in missing values and converting the flight data format to make it suitable for deep learning models;

[0050] A DQN intelligent optimization module, which outputs the best turning action a based on the aircraft state s to achieve the optimal turning strategy;

[0051] The DQN intelligent optimization module is imported into the airborne equipment, and the DQN intelligent optimization module adjusts the control parameters of the aircraft to control the aircraft to complete the turning action. Among them, the DQN intelligent optimization module includes:

[0052] A Deep Q-Network model, which adjusts the control parameters of the aircraft to control the aircraft to complete the turning action;

[0053] An environment definition unit, which is used to define the flight environment state, and quantize and discretize the flight environment state into a finite state space S; S = {s_1, s_2,..., s_n}, where s_n represents the state at time t; and define the action space A; A = {a_1, a_2,..., a_m}, where a_m represents the action taken at time t;

[0054] The flight environment includes the aircraft state and executable actions. The aircraft state includes, for example, speed, altitude, attitude, etc., and the executable operations include, for example, the rudder surface deflection angle, thrust adjustment amount, etc.

[0055] A reward function unit that constructs a reward function R(s, a, s'), which is used to measure the effect of reaching a new state s' after executing an action a in state s;

[0056] A Q-value function unit, which is used to obtain the value evaluation Q(s, a) of action a when in state s;

[0057] An experience replay buffer unit, which is used to store a large number of experiences (<s, a, r, s'>), where s is the current state, a is the action taken, r is the immediate reward received, and s' is the next state after executing the action. Some core pseudocode is as follows:

[0058] # Pseudocode

[0059] replay_buffer = deque(maxlen = replay_buffer_capacity)

[0060] In the above, the Q-value function unit uses a deep neural network as an approximator of the Q-value function. The input of the deep neural network is the state s, and the output is the value evaluation Q(s, a) of action a in state s;

[0061] The network architecture of the deep neural network adopts a multi-layer perceptron, and the last layer corresponds to the value evaluation Q(s, a) of each action; among them, the main difference between the multi-layer perceptron and the single-layer neural network is that the multi-layer perceptron has an additional hidden layer. The single-layer neural network only has an input layer and an output layer, while the multi-layer neural network has one or more hidden layers located between the input layer and the output layer. Therefore, the multi-layer perceptron is also more complex. The multi-layer perceptron (Multi-Layer Perceptron) is abbreviated as MLP, which is a feedforward neural network, a fully connected and forward-propagating neural network, composed of an input layer, one or more hidden layers, and an output layer.

[0062] Some core pseudocode is as follows:

[0063] # Pseudocode

[0064] def Q_network(input_shape, output_dim, learning_rate):

[0065] model = Sequential()

[0066] # Build a deep neural network structure (such as fully connected layers, convolutional layers, etc.)

[0067] #...

[0068] model.compile(optimizer=Adam(lr=learning_rate), loss='mse')

[0069] return model

[0070] In the above, the training process of the Deep Q-Network model includes:

[0071] Explore the environment through the ε-greedy strategy, perform random actions or select the action with the highest current Q value;

[0072] After performing the action, obtain the new state s' and the corresponding reward r, and store the experience in the experience replay buffer unit;

[0073] Sample a batch of data from the experience replay buffer unit for learning, update the DQN network parameters to narrow the gap between the value evaluation Q(s,a) and the actual return.

[0074] The following is part of the core pseudocode:

[0075] · Select an action using the ε-greedy strategy:

[0076] epsilon = get_epsilon(timestep) # Gradually decrease the value of ε as the training progresses. If np.random.rand() < epsilon:

[0077] action = np.random.choice(A) # Random exploration

[0078] else:

[0079] action = np.argmax(Q_network.predict(state)) # Use the Q-network to select the optimal action

[0080] · Execute the action and obtain the result:

[0081] new_state, reward, done, _ = environment.step(action)

[0082] replay_buffer.append((state, action, reward, new_state, done))

[0083] · Update Q-network parameters:

[0084] if len(replay_buffer) > batch_size:

[0085] mini_batch = random.sample(replay_buffer, batch_size)

[0086] states_mb, actions_mb, rewards_mb, next_states_mb, dones_mb = zip(*mini_batch)

[0087] target_Q_values = Q_target(next_states_mb).detach().numpy()

[0088] max_Q_next = np.max(target_Q_values, axis = 1)

[0089] targets = rewards_mb + gamma * max_Q_next * (1 - dones_mb)

[0090] with torch.no_grad():

[0091] predicted_Q_values = Q_network(states_mb).gather(1, actions_mb.unsqueeze(-1)).squeeze(-1)

[0092] loss = F.mse_loss(predicted_Q_values, torch.tensor(targets))

[0093] optimizer.zero_grad()

[0094] loss.backward()

[0095] optimizer.step()

[0096] To enable the control system to continuously and smoothly adjust the tilt of the aircraft body, reduce the turning radius and overload while ensuring flight stability and passenger comfort, the control system further includes: a dynamic tilt turning control module for continuously and smoothly adjusting the tilt of the aircraft body, reducing the turning radius and overload;

[0097] The dynamic tilt turning control module includes:

[0098] A motion model construction unit that establishes a motion model of the aircraft based on the principles of aircraft dynamics and aerodynamics. The motion model includes six-degree-of-freedom motion states and physical quantities related to the six-degree-of-freedom motion states. The motion model equations are:

[0099] m*d^2x / dt^2 = F_x + m*g*sin(θ)

[0100] m*d^2y / dt^2 = F_y

[0101] I_z*d^2ψ / dt^2 = L_z

[0102] where m is the mass, I_z is the moment of inertia, (x,y) are the position coordinates, ψ is the roll angle, (F_x,F_y) are the components of the resultant force in the xy plane, L_z is the moment about the z axis, and θ is the pitch angle;

[0103] An index definition unit for defining performance indices. The quadratic form of the performance index J is:

[0104] J = ∫[Q(x,u) + R(u)]dt

[0105] where x is the state vector, u is the control input vector, Q(x,u) is the state cost function, and R(u) is the control input cost function;

[0106] In this embodiment, an optimization objective function can also be set, and this function should reflect the goal of minimizing or maximizing the desired performance indices. Establish constraint conditions, which may involve aircraft operation limitations such as speed range, overload limit, sideslip angle limit, control surface and engine operation limits, etc. Construct the form of the Hamilton-Jacobi-Bellman (HJB) equation or the Lagrange multiplier method:

[0107] H(x,u,λ) = Q(x,u) + λ'f(x,u)

[0108] Select a linear quadratic regulator (LQR):

[0109] u ^u* = -K*x

[0110] K is obtained by solving the continuous-time Riccati equation:

[0111] 0 = A'P + PA - PBR^{-1}B'P + Q

[0112] The gain matrix K = R^{-1}B'P is obtained

[0113] The turning dynamic adjustment unit, based on the calculated control rate, calculates and updates the aircraft's attitude (roll, pitch) and thrust distribution in real time according to the feedback data after the execution action during the turning process, so as to achieve a fast, efficient and safe banked turn action based on the preset performance index J; The following is part of the core pseudocode:

[0114] def compute_control_inputs(state,desired_heading):

[0115] # Calculate the required roll rate and thrust allocation

[0116] roll_rate_command = calculate_optimal_roll_rate(state,desired_heading)

[0117] thrust_distribution = allocate_thrust_for_turn(state)

[0118] return roll_rate_command,thrust_distribution

[0119] The closed-loop control unit obtains the output result of the Deep Q-Network model, generates an action instruction based on the output structure, and the on-board equipment (servo, engine, etc.) receives and executes the instruction to form a feedback loop. The following is part of the core pseudocode:

[0120] error = setpoint - current_state

[0121] control_signal = K*error + optimal_control_input(state)

[0122] apply_control(control_signal)

[0123] In this embodiment, the control rate is a mathematical model calculated based on the error for controlling the input. The mathematical model is a PID controller (Proportional-Integral-Derivative controller), which is used to select an appropriate controller structure, set controller parameters, and analyze the stability, response speed, and steady-state error characteristics of the system using dynamic system modeling techniques. The following is part of the core pseudocode:

[0124]

[0125] The error is the result of comparing the current aircraft state parameters with the target set value. For state variables in multiple dimensions, such as attitude angles, position coordinates, speed, etc., the corresponding error values need to be calculated separately. The following is part of the core pseudocode:

[0126] def calculate_error(target_value,current_value):

[0127] return target_value-current_value

[0128] desired_heading=90# Assume the desired heading is 90 degrees

[0129] actual_heading=get_heading_from_imu()# Obtain the actual heading

[0130] heading_error=calculate_error(desired_heading,actual_heading)

[0131] In this embodiment, the regulation system further includes a model real-time optimization module. Based on the closed-loop control principle, a closed-loop control system including multi-source sensor data fusion, error detection, real-time calculation, and command output is constructed. The closed-loop control is used to monitor the deviation between the actual turning process of the aircraft and the predetermined trajectory and immediately take compensation measures. The model real-time optimization module includes:

[0132] A data fusion unit that uses multiple sensors (such as gyroscopes, accelerometers, magnetometers, GPS navigation systems, etc.) to continuously monitor the attitude, position, speed, heading, and environmental parameters of the aircraft. Through sensor fusion technology, multi-source information is integrated to generate an accurate and complete flight state estimate. The following is part of the core pseudocode:

[0133] # Example pseudocode

[0134] import sensor_library

[0135] imu = sensor_library.IMU() # Initialize the IMU sensor object

[0136] gps = sensor_library.GPS() # Initialize the GPS sensor object

[0137] while True:

[0138] gyro_data, accel_data = imu.read() # Read and store the current attitude angular velocity and acceleration data

[0139] The feedback unit obtains the error signal in real time and generates and sends a control instruction immediately after obtaining the error signal to control the input of the Deep Q-Network model, so as to continuously adjust the output of the Deep Q-Network model until it approaches or reaches the expected state. The following is part of the core code:

[0140] def set_servo_angle(angle):

[0141] servo_device.set_angle(angle)

[0142] # Convert the PID output to the servo angle and execute

[0143] corrected_servo_angle = map_control_signal_to_angle(control_signal)

[0144] set_servo_angle(corrected_servo_angle)

[0145] The performance optimization unit records the key performance indicators in real time, evaluates the effect of the PID controller using the key indicators, and according to the evaluation results, adopts an online self-tuning algorithm or an offline parameter tuning method to continuously optimize the PID controller parameters to improve the control performance. The following is part of the core code:

[0146] record_error_history(heading_error)

[0147] optimize_pid_parameters(error_history)

[0148] The security mechanism unit sets a safety threshold. When the state variable exceeds the safety threshold, an alarm is triggered or the safety mode is entered. The following is part of the core code:

[0149] if abs(control_signal)>MAX_CONTROL_LIMIT:

[0150] trigger_safety_protocol()

[0151] The control system further includes a fault detection module, which is connected to the security mechanism unit. After triggering an alarm or entering the safety mode, the security mechanism unit sends a fault code to the fault detection module. The fault detection module identifies the fault code and switches to the standby system or takes emergency measures to ensure flight safety.

[0152] The embodiments of the present invention have been described above. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to technologies in the market, or to enable other ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. An intelligent aircraft turning control system, characterized in that, It includes: A data acquisition, preprocessing module and a DQN intelligent optimization module; The data acquisition collects flight data in actual flight records, simulator simulations or wind tunnel experiments; the flight data includes flight performance parameters and corresponding turning operation details under different speeds, altitudes, weight distributions, and meteorological conditions; The preprocessing module is connected to the data acquisition and preprocesses the collected flight data; The DQN intelligent optimization module is connected to the preprocessing module, uses a deep learning model to obtain the aircraft state s based on the preprocessed flight data, and outputs the best turning action a based on the aircraft state s to obtain an optimal turning strategy; The DQN intelligent optimization module includes: A Deep Q-Network model for adjusting the control parameters of the aircraft and controlling the aircraft to complete turning actions; An environment definition unit for defining the flight environment state, quantifying and discretizing the flight environment state into a finite state space S; S = {s_1, s_2, ..., s_n}, where s_n represents the state at time t; and defining the action space A; A = {a_1, a_2, ..., a_m}, where a_m represents the action taken at time t; The flight environment includes the aircraft state and executable actions; A reward function unit for constructing a reward function R(s, a, s'), and measuring the effect of reaching a new state s' after executing the action a in the state s based on the reward function R(s, a, s'); A Q-value function unit for obtaining the value evaluation Q(s, a) of the action a in the state s; An experience replay buffer unit for storing a large amount of experience data <s, a, r, s'>, where s is the current state, a is the action taken, r is the immediate reward received, and s' is the next state after executing the action; The Q-value function unit uses a deep neural network as an approximator of the Q-value function. The input of the deep neural network is the state s, and the output is the value evaluation Q(s, a) of the action a in the state s; The network architecture of the deep neural network adopts a multi-layer perceptron, and the last layer of the deep neural network corresponds to the value evaluation Q(s, a) of each action.

2. The intelligent aircraft turning control system according to claim 1, characterized in that, The preprocessing includes removing outliers, filling in missing values, and converting the flight data format to make it suitable for the deep learning model.

3. An intelligent aircraft turning control system according to claim 2, characterized in that Before the Deep Q-Network model is used to adjust the control parameters of the aircraft, it needs to be trained. The training process includes: Exploring the environment through the ε-greedy strategy, performing random actions or selecting the action with the highest current Q value; After executing the action, obtaining the new state s' and the corresponding reward r, and storing the experience in the experience replay buffer unit; Sampling a batch of data from the experience replay buffer unit for learning, updating the DQN network parameters to narrow the gap between the value evaluation Q(s,a) and the actual turning flight execution actions of the aircraft.

4. An intelligent aircraft turning control system according to claim 3, characterized in that, The control system further includes: a dynamic banked turn control module for continuously and smoothly adjusting the tilt of the aircraft body to reduce the turning radius and overload; The dynamic banked turn control module includes: A motion model construction unit that establishes a motion model of the aircraft based on the principles of aircraft dynamics and aerodynamics. The motion model includes six-degree-of-freedom motion states and physical quantities related to the six-degree-of-freedom motion states. The motion model equations are: m * d^2x / dt^2 = F_x + m * g * sin(θ) m * d^2y / dt^2 = F_y I_z * d^2ψ / dt^2 = L_z where m is the mass, I_z is the moment of inertia, (x, y) are the position coordinates, ψ is the roll angle, (F_x, F_y) are the components of the resultant force in the xy plane, L_z is the moment about the z-axis, and θ is the pitch angle; An index definition unit for defining performance indices. The quadratic form of the performance index J is: J = ∫[Q(x, u) + R(u)] dt where x is the state vector, u is the control input vector, Q(x, u) is the state cost function, and R(u) is the control input cost function; A turn dynamic adjustment unit that dynamically adjusts the attitude and thrust distribution of the aircraft during the turn based on the calculated control rate to achieve a banked turn action based on the preset performance index J; A closed-loop control unit that obtains the output result of the Deep Q-Network model, generates an action instruction based on the output structure, and the on-board equipment receives and executes the instruction to form a feedback loop.

5. An intelligent aircraft turning control system according to claim 4, characterized in that, The control rate is a mathematical model for calculating the control input based on the error. The mathematical model is a PID controller; the error is the result of comparing the current aircraft state parameters with the target set value.

6. An intelligent aircraft turning control system according to claim 5, characterized in that, The control system further includes a model real-time optimization module that constructs a closed-loop control system including multi-source sensor data fusion, error detection, real-time calculation, and instruction output based on the closed-loop control principle. The closed-loop control is used to monitor the deviation between the actual turning process of the aircraft and the predetermined trajectory and immediately take compensation measures; The model real-time optimization module includes: A data fusion unit that integrates multi-source information through sensor fusion technology to generate a flight state estimate; A feedback unit that obtains the error signal in real time and immediately generates and sends a control instruction after obtaining the error signal to control the input of the Deep Q-Network model, thereby continuously adjusting the output of the Deep Q-Network model until it approaches or reaches the expected state; A performance optimization unit that records the key performance indices in real time and uses the key performance indices to evaluate the effect of the PID controller. According to the evaluation result, an online self-tuning algorithm or an offline parameter tuning method is used to continuously optimize the PID controller parameters to improve the control performance; A security mechanism unit that sets a safety threshold. When the state variable exceeds the safety threshold, an alarm is triggered or the safety mode is entered.

7. An intelligent aircraft turning control system according to claim 6, characterized in that, The control system further includes a fault detection module, which is connected to the security mechanism unit. After triggering an alarm or entering the safe mode, the security mechanism unit sends a fault code to the fault detection module. The fault detection module identifies the fault code and switches to the standby system or takes emergency measures to ensure flight safety.

Images