Actuator multi-mode failure-oriented distributed driving hovercar self-adaptive fault-tolerant control method

Abstract

The invention relates to the technical field of aerocar mode switching, and discloses a distributed driving aerocar self-adaptive fault-tolerant control method for actuator multimode failure, which comprises the following steps: constructing a unified six-degree-of-freedom dual-mode state space model; residual signals are generated based on extended Kalman filtering and a sliding-mode observer, and fault types and positions are positioned in real time through a lightweight classifier; the method comprises the following steps: extracting residual time-frequency features, identifying hard faults by using a lightweight convolutional neural network, quantifying soft fault degrees through an incremental support vector machine, fusing multi-source information based on a Bayesian network to output fault types, levels and confidence coefficients, and introducing an incremental learning mechanism to realize self-evolution of a diagnosis model; a virtual control instruction is generated by adopting hierarchical sliding mode control, thrust and torque distribution of remaining actuators is optimized based on a dynamic quadratic programming algorithm, control parameters are adjusted online in combination with a Lyapunov adaptive law, aerodynamic interference and model uncertainty are inhibited, attitude stability and trajectory tracking in air-ground mode switching are guaranteed, and the method has the advantages of being high in reliability and high in reliability. And the fault-tolerant performance and the operation safety of the hovercar in the air-ground mode switching process are obviously enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of flying car mode switching, and in particular to an adaptive fault-tolerant control method for a distributed drive flying car oriented to actuator multi-mode failure. Background Art

[0002] With the full coverage of 5G communication networks, the maturity of satellite navigation and high-precision positioning technologies, and continuous breakthroughs in materials science, power batteries and electric drive systems, various types of drones and light aircraft have been put into large-scale use in scenarios such as precision agricultural spraying, smart logistics distribution, urban air taxis, emergency medical rescue, and material transportation in mining areas and islands.

[0003] Flying cars combine the road-driving capabilities of traditional cars with the three-dimensional maneuverability of aircraft, paving the way for urban traffic and optimizing intermodal transport. Currently, several manned flying platforms, both domestically and internationally, have completed prototype verification for vertical takeoff and landing, ferry flights, and short-distance transport. However, most still operate as separate systems, one for the vehicle and the other for the aircraft, and have yet to achieve truly seamless ground-to-air integration.

[0004] On the technical level, the engines, motors, transmission mechanisms, and control systems of new energy vehicles and aircraft each possess certain failure detection and fault-tolerant control capabilities; however, faced with the high coupling and multi-operating mode switching requirements of integrated flying vehicles, existing methods struggle to meet system-level safety and reliability requirements. Existing fault-tolerant control methods mostly rely on splitting the entire vehicle into two redundant subsystems, "road mode" and "flight mode," ignoring the dynamic interaction and fault propagation mechanisms between subsystems under the two operating conditions. To address this challenge, there is an urgent need to build a unified reliability framework covering the multi-mode switching process and develop a full-link solution including distributed redundant design, online health management, adaptive fault-tolerant algorithms, and decision-making layer security assurance to support the safe and efficient operation of future integrated flying vehicles in complex urban airspace and road environments. Summary of the Invention

[0005] To address the above technical problems, the present invention provides an adaptive fault-tolerant control method for a distributed-drive flying vehicle oriented to multi-mode actuator failures. This method enables rapid and accurate fault location and hierarchical diagnosis during flying vehicle mode switching, dynamic and balanced allocation of remaining actuators, and adaptive control parameter adjustment, thus addressing the current lack of a fault-tolerant mechanism for flying vehicle mode switching.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] An adaptive fault-tolerant control method for a distributed drive flying car oriented to actuator multi-mode failures, comprising:

[0008] Establish a dual-mode state-space model that takes the rotor thrust vector and the in-wheel motor torque as input and the vehicle body's six-degree-of-freedom posture as output;

[0009] Based on the actual working state data of each actuator and the data estimated by the dual-mode state space model for the working state of each actuator, a residual is formed, and the fault mode of each actuator is detected in real time based on the residual and in combination with a lightweight classifier;

[0010] A hybrid intelligent algorithm based on short-time Fourier transform and lightweight convolutional neural network is used to extract the time-frequency characteristics of the residual and identify hard faults. Incremental support vector machines are combined to quantify the extent of soft faults. A Bayesian network is used to fuse multi-source information to output the fault type, fault severity level, and confidence level, enabling diagnosis of multi-mode failures.

[0011] Hierarchical sliding mode control and a quadratic programming dynamic allocation algorithm with health weights are used to optimize the thrust and torque output of the remaining actuators in real time. Combined with the Lyapunov-based adaptive law, the control parameters are adjusted online, ensuring that the attitude angle and position of the flying car converge when an actuator fails during air-to-ground mode switching.

[0012] In one embodiment, establishing a dual-mode state space model that takes the rotor thrust vector and the wheel hub motor torque as input and the vehicle body six-degree-of-freedom posture as output specifically includes:

[0013] The nonlinear dynamic sub-models of the rotor-airframe system and the wheel-body system were established using Lagrange equations.

[0014] A sigmoid function is used as the switching mechanism to smoothly transition between the two dynamic sub-models according to the flight altitude or vehicle speed, resulting in a unified dual-mode dynamic model that can be directly used for multi-actuator control distribution.

[0015] In one embodiment, the residual is formed based on the actual working state data of each actuator and the data estimated by the dual-mode state space model on the working state of each actuator. The fault mode of each actuator is detected in real time based on the residual and combined with a lightweight classifier, specifically including:

[0016] A multi-rate sampling-based fault detection system is deployed at each actuator node. The fault detection system includes a rotor speed monitoring module, a wheel speed sensor array, and a motor current detection unit to collect real-time data on the actual working status of each actuator.

[0017] Inputting the actual working state data into an extended Kalman filter, performing a joint fusion estimation of the six-degree-of-freedom motion state of the flying car body and the state of each actuator based on the dual-mode state space model, and continuously calculating the difference between the actual working state of each actuator and the working state predicted by the dual-mode state space model to form a residual;

[0018] Threshold determination and lightweight classification are performed on the residual to identify the actuator fault type and locate the actuator fault position in real time.

[0019] In one embodiment, a hybrid intelligent algorithm based on short-time Fourier transform and lightweight convolutional neural network is used to extract the time-frequency features of the residual and identify hard faults, specifically including:

[0020] First, the residual is decoupled from the time-frequency characteristics: a sliding window short-time Fourier transform is used to extract the time-frequency domain energy distribution, and the high-frequency component is separated from the low-frequency component through principal component analysis to form a high-frequency feature vector representing sudden faults and a low-frequency feature vector representing gradual faults;

[0021] Deploy convolutional neural networks for high-frequency feature vectors: Using the compressed MobileNetV3 Tiny architecture with a deep separable convolutional structure, it identifies transient characteristic patterns of hard faults through parallel computing and outputs discrete fault classification labels and their confidence levels.

[0022] In one embodiment, the high-frequency component is a component greater than 800 Hz, and the low-frequency component is a component less than 100 Hz.

[0023] In one embodiment, the method of combining an incremental support vector machine to quantify the degree of soft faults and fusing multi-source information through a Bayesian network to output the fault type, fault severity level, and confidence level to achieve multi-mode failure diagnosis specifically includes:

[0024] An incremental support vector machine model is constructed for low-frequency feature vectors. A rolling time window is used to perform regression analysis on time-domain slowly varying features, and the soft fault parameters of each actuator are quantified in real time.

[0025] Establish a Bayesian probability decision network: This network integrates the current sensor data of each actuator, the classification results of the hard fault, the quantitative parameters of the soft fault, and the historical fault statistics library to calculate the joint confidence distribution of the fault type and fault severity level. The fault types include hard faults, soft faults, and combined hard and soft faults.

[0026] An incremental learning mechanism is designed to achieve the evolution of the classification system: when it is detected that the joint confidence is continuously lower than the set threshold and the residual exceeds the limit, the classifier fault category label set is automatically expanded, and the convolutional neural network classification nodes, support vector machine model kernel function mapping relationship and Bayesian probability decision network conditional probability table are synchronously updated.

[0027] In one embodiment, the hierarchical sliding mode control and the quadratic programming dynamic allocation algorithm with health weights are used to optimize the thrust and torque output of the remaining actuators in real time, combined with the Lyapunov-based adaptive law to adjust the control parameters online, so that the attitude angle and position of the flying car converge when the actuator fails during air-to-ground mode switching. Specifically, the method includes:

[0028] For the flying car's six-degree-of-freedom dynamics model, adaptive sliding mode control combined with Lyapunov stability theory is used to respond to position and velocity deviations and online generate the virtual control variables required for each degree of freedom. This ensures the tracking accuracy of roll, pitch, yaw, and lift forces while suppressing aerodynamic interference. The virtual control variables include virtual control forces and virtual control torques.

[0029] Based on the health status of each actuator and the characteristics of the air-ground hybrid drive, a dynamic quadratic programming problem with safety limits and mode switching constraints is constructed to map the virtual control variables of each degree of freedom to rotor thrust, motor torque, and control surface deflection commands.

[0030] The actuator weights are dynamically set according to the flying car's height from the ground, the severity of the fault, and the current mode. The Lyapunov method is combined to adjust the quadratic programming and sliding mode control parameters online to achieve convergence of attitude angle and position.

[0031] Compared with the prior art, the beneficial technical effects of the present invention are:

[0032] 1. The present invention achieves smooth switching between flight mode and ground mode in the same framework through a unified six-degree-of-freedom dual-mode state space model constructed based on the Lagrange equation and the S-shaped function, overcoming the problems of inconsistent models, large switching jitter and difficulty in online coordination when switching between two independent subsystems in the prior art.

[0033] 2. The present invention deploys a multi-rate sampling monitoring system at each actuator node, and combines extended Kalman filtering with a sliding mode observer to generate a residual signal. The lightweight classifier locates the fault type and location in real time, achieving rapid and accurate diagnosis of multi-mode failures of rotor and hub motors, significantly reducing fault detection delay and false alarm rate.

[0034] 3. The present invention uses short-time Fourier transform to extract residual time-frequency features, integrates lightweight convolutional neural networks to identify hard faults, and incremental support vector machines to quantify soft faults. It also fuses multi-source information based on a Bayesian network to output fault levels and confidence levels. At the same time, it introduces an incremental learning mechanism to enable the self-evolution of the diagnostic model, greatly improving the accuracy, scalability, and long-term adaptability of the joint diagnosis of hard and soft faults.

[0035] 4. The present invention adopts hierarchical sliding mode control to generate virtual control instructions, a dynamic quadratic programming algorithm to optimize the thrust / torque distribution of the remaining actuators online, and combines the Lyapunov adaptive law to adjust the control parameters in real time. It can quickly restore attitude stability and trajectory tracking after actuator failure, significantly enhancing the fault tolerance and operational safety of the flying car during the air-to-ground mode switching process. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Flowchart of a method in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] A preferred embodiment of the present invention will be described in detail below with reference to the accompanying drawings.

[0038] like Figure 1 As shown, the present invention provides an adaptive fault-tolerant control method for a distributed drive flying car facing multi-mode failure of actuators, comprising the following steps:

[0039] S1, establish a dual-mode state space model with rotor thrust vector and wheel hub motor torque as input and vehicle body six degrees of freedom posture as output;

[0040] S2, generating a residual based on the actual working state data of each actuator and the data estimated by the dual-mode state space model for the working state of each actuator, and performing real-time detection of the fault mode of each actuator based on the residual in combination with a lightweight classifier;

[0041] S3 uses a hybrid intelligent algorithm based on short-time Fourier transform and lightweight convolutional neural network to extract the time-frequency characteristics of residuals and identify hard faults. It also combines incremental support vector machines to quantify the extent of soft faults and fuses multi-source information through a Bayesian network to output fault type, fault severity level, and confidence level, enabling diagnosis of multi-mode failures.

[0042] S4 uses hierarchical sliding mode control and a quadratic programming dynamic allocation algorithm with health weights to optimize the thrust and torque output of the remaining actuators in real time. Combined with the Lyapunov-based adaptive law, it adjusts the control parameters online, allowing the attitude angle and position of the flying car to converge when an actuator fails during air-to-ground mode switching, achieving a safe transition.

[0043] In one embodiment, establishing a dual-mode state space model that takes the rotor thrust vector and the wheel hub motor torque as input and the vehicle body six-degree-of-freedom posture as output specifically includes:

[0044] The nonlinear dynamic sub-models of the rotor-airframe system and the wheel-body system were established using Lagrange equations.

[0045] A sigmoid function is used as the switching mechanism to smoothly transition between the two dynamic sub-models according to the flight altitude or vehicle speed, resulting in a unified dual-mode dynamic model that can be directly used for multi-actuator control distribution.

[0046] In one embodiment, the residual is formed based on the actual working state data of each actuator and the data estimated by the dual-mode state space model on the working state of each actuator. The fault mode of each actuator is detected in real time based on the residual and combined with a lightweight classifier, specifically including:

[0047] A multi-rate sampling-based fault detection system is deployed at each actuator node. The fault detection system includes a rotor speed monitoring module, a wheel speed sensor array, and a motor current detection unit to collect real-time data on the actual working status of each actuator.

[0048] Inputting the actual working state data into an extended Kalman filter, performing a joint fusion estimation of the six-degree-of-freedom motion state of the flying car body and the state of each actuator based on the dual-mode state space model, and continuously calculating the difference between the actual working state of each actuator and the working state predicted by the dual-mode state space model to form a residual;

[0049] The residual error is subjected to threshold determination and lightweight classification to identify the actuator fault type and locate the actuator fault location in real time. The present invention can also report the fault type and location to the main control unit.

[0050] In one embodiment, a hybrid intelligent algorithm based on short-time Fourier transform and lightweight convolutional neural network is used to extract the time-frequency features of the residual and identify hard faults, specifically including:

[0051] First, the residual is decoupled from the time-frequency characteristics: a sliding window short-time Fourier transform is used to extract the time-frequency domain energy distribution, and the high-frequency component is separated from the low-frequency component through principal component analysis to form a high-frequency feature vector representing sudden faults and a low-frequency feature vector representing gradual faults;

[0052] Deploy convolutional neural networks for high-frequency feature vectors: Using the compressed MobileNetV3 Tiny architecture with a deep separable convolutional structure, it identifies transient characteristic patterns of hard faults through parallel computing and outputs discrete fault classification labels and their confidence levels.

[0053] Hard faults include freezing, communication interruption, and other faults.

[0054] In one embodiment, the high-frequency component is a component greater than 800 Hz, and the low-frequency component is a component less than 100 Hz.

[0055] In one embodiment, the method of combining an incremental support vector machine to quantify the degree of soft faults and fusing multi-source information through a Bayesian network to output the fault type, fault severity level, and confidence level to achieve multi-mode failure diagnosis specifically includes:

[0056] An incremental support vector machine model is constructed for low-frequency feature vectors. A rolling time window is used to perform regression analysis on the time-domain slowly varying features, and the soft fault quantization parameters of each actuator are quantified in real time. The kernel function weights are updated online to adapt to different degradation trajectories. The soft fault quantization parameters can include the efficiency attenuation ratio (0-100%) and the aging rate coefficient (% / h).

[0057] Establish a Bayesian probability decision network: This network integrates the current sensor data of each actuator, the classification results of the hard fault, the quantitative parameters of the soft fault, and the historical fault statistics library to calculate the joint confidence distribution of the fault type and fault severity level. The fault types include hard faults, soft faults, and combined hard and soft faults.

[0058] An incremental learning mechanism is designed to achieve the evolution of the classification system: when it is detected that the joint confidence is continuously lower than the set threshold and the residual exceeds the limit, the classifier fault category label set is automatically expanded, and the convolutional neural network classification nodes, support vector machine model kernel function mapping relationship and Bayesian probability decision network conditional probability table are synchronously updated.

[0059] In one embodiment, the hierarchical sliding mode control and the quadratic programming dynamic allocation algorithm with health weights are used to optimize the thrust and torque output of the remaining actuators in real time, combined with the Lyapunov-based adaptive law to adjust the control parameters online, so that the attitude angle and position of the flying car converge when the actuator fails during air-to-ground mode switching. Specifically, the method includes:

[0060] For the flying car's six-degree-of-freedom dynamics model, adaptive sliding mode control combined with Lyapunov stability theory is used to respond to position and velocity deviations and online generate the virtual control variables required for each degree of freedom. This ensures the tracking accuracy of roll, pitch, yaw, and lift forces while suppressing aerodynamic interference. The virtual control variables include virtual control forces and virtual control torques.

[0061] Based on the health status of each actuator and the characteristics of air-ground hybrid drive, a dynamic quadratic programming problem with safety limits and mode switching constraints is constructed. The virtual control variables of each degree of freedom are mapped and distributed to rotor thrust, motor torque, and control surface deflection commands, giving priority to meeting the core requirements of mode switching and avoiding actuator overload.

[0062] The actuator weights are dynamically set according to the flying car's height from the ground, the severity of the fault, and the current mode. The Lyapunov method is combined to online adjust the quadratic programming and sliding mode control parameters to adaptively suppress the model uncertainty caused by the fault, so that the attitude angle and position converge.

[0063] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times. The order of execution of these steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0064] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0065] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. It is intended that all variations within the meaning and range of equivalents of the claims be embraced herein, and any reference signs in the claims should not be construed as limiting the claims to which they relate.

[0066] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. An adaptive fault-tolerant control method for a distributed drive flying car facing multi-mode failure of actuators, characterized by: include: Establish a dual-mode state-space model that takes the rotor thrust vector and the in-wheel motor torque as input and the vehicle body's six-degree-of-freedom posture as output; Based on the actual working state data of each actuator and the data estimated by the dual-mode state space model for the working state of each actuator, a residual is formed, and the fault mode of each actuator is detected in real time based on the residual and in combination with a lightweight classifier; A hybrid intelligent algorithm based on short-time Fourier transform and lightweight convolutional neural network is used to extract the time-frequency characteristics of the residual and identify hard faults. Incremental support vector machines are combined to quantify the extent of soft faults. A Bayesian network is used to fuse multi-source information to output the fault type, fault severity level, and confidence level, enabling diagnosis of multi-mode failures. Hierarchical sliding mode control and a quadratic programming dynamic allocation algorithm with health weights are used to optimize the thrust and torque output of the remaining actuators in real time. Combined with the Lyapunov-based adaptive law, the control parameters are adjusted online, ensuring that the attitude angle and position of the flying car converge when an actuator fails during air-to-ground mode switching.

2. The adaptive fault-tolerant control method for distributed drive flying vehicles facing multi-mode actuator failure according to claim 1, characterized in that: The dual-mode state space model is established, which takes the rotor thrust vector and the wheel hub motor torque as input and the vehicle body six-degree-of-freedom posture as output, specifically including: The nonlinear dynamic sub-models of the rotor-airframe system and the wheel-body system were established using Lagrange equations. A sigmoid function is used as the switching mechanism to smoothly transition between the two dynamic sub-models according to the flight altitude or vehicle speed, resulting in a unified dual-mode dynamic model that can be directly used for multi-actuator control distribution.

3. The adaptive fault-tolerant control method for distributed drive flying vehicles facing multi-mode actuator failure according to claim 1, characterized in that: The actual working state data of each actuator and the data estimated by the dual-mode state space model on the working state of each actuator are used to form a residual, and the fault mode of each actuator is detected in real time based on the residual and in combination with a lightweight classifier, specifically including: A multi-rate sampling-based fault detection system is deployed at each actuator node. The fault detection system includes a rotor speed monitoring module, a wheel speed sensor array, and a motor current detection unit to collect real-time data on the actual working status of each actuator. Inputting the actual working state data into an extended Kalman filter, performing a joint fusion estimation of the six-degree-of-freedom motion state of the flying car body and the state of each actuator based on the dual-mode state space model, and continuously calculating the difference between the actual working state of each actuator and the working state predicted by the dual-mode state space model to form a residual; Threshold determination and lightweight classification are performed on the residual to identify the actuator fault type and locate the actuator fault position in real time.

4. The adaptive fault-tolerant control method for distributed drive flying vehicles facing multi-mode actuator failure according to claim 1, characterized in that: The hybrid intelligent algorithm based on short-time Fourier transform and lightweight convolutional neural network is used to extract the time-frequency features of the residual and identify hard faults, specifically including: First, the residual is decoupled from the time-frequency characteristics: a sliding window short-time Fourier transform is used to extract the time-frequency domain energy distribution, and the high-frequency component is separated from the low-frequency component through principal component analysis to form a high-frequency feature vector representing sudden faults and a low-frequency feature vector representing gradual faults; Deploy convolutional neural networks for high-frequency feature vectors: Adopting a compressed MobileNetV3Tiny architecture with a deep separable convolutional structure, it identifies transient characteristic patterns of hard faults through parallel computing and outputs discretized fault classification labels and their confidence levels.

5. The adaptive fault-tolerant control method for distributed drive flying vehicles facing multi-mode actuator failure according to claim 4, characterized in that: The high-frequency component is a component greater than 800 Hz, and the low-frequency component is a component less than 100 Hz.

6. The adaptive fault-tolerant control method for distributed drive flying vehicles facing multi-mode actuator failure according to claim 4, characterized in that: The method combines the incremental support vector machine to quantify the degree of soft faults, fuses multi-source information through the Bayesian network to output the fault type, fault severity level and confidence level, and realizes the diagnosis of multi-mode failures, specifically including: An incremental support vector machine model is constructed for low-frequency feature vectors. A rolling time window is used to perform regression analysis on time-domain slowly varying features, and the soft fault parameters of each actuator are quantified in real time. Establish a Bayesian probability decision network: This network integrates the current sensor data of each actuator, the classification results of the hard fault, the quantitative parameters of the soft fault, and the historical fault statistics library to calculate the joint confidence distribution of the fault type and fault severity level. The fault types include hard faults, soft faults, and combined hard and soft faults. An incremental learning mechanism is designed to achieve the evolution of the classification system: when it is detected that the joint confidence is continuously lower than the set threshold and the residual exceeds the limit, the classifier fault category label set is automatically expanded, and the convolutional neural network classification nodes, support vector machine model kernel function mapping relationship and Bayesian probability decision network conditional probability table are synchronously updated.

7. The adaptive fault-tolerant control method for distributed drive flying vehicles facing multi-mode actuator failure according to claim 1, characterized in that: The proposed method uses hierarchical sliding mode control and a quadratic programming dynamic allocation algorithm with health weights to optimize the thrust and torque output of the remaining actuators in real time. Combined with Lyapunov's adaptive law, the control parameters are adjusted online to ensure that the attitude angle and position of the flying car converge when an actuator fails during air-to-ground mode switching. Specifically, the method includes: For the flying car's six-degree-of-freedom dynamics model, adaptive sliding mode control combined with Lyapunov stability theory is used to respond to position and velocity deviations and online generate the virtual control variables required for each degree of freedom. This ensures the tracking accuracy of roll, pitch, yaw, and lift forces while suppressing aerodynamic interference. The virtual control variables include virtual control forces and virtual control torques. Based on the health status of each actuator and the characteristics of the air-ground hybrid drive, a dynamic quadratic programming problem with safety limits and mode switching constraints is constructed to map the virtual control variables of each degree of freedom to rotor thrust, motor torque, and control surface deflection commands. The actuator weights are dynamically set according to the flying car's height from the ground, the severity of the fault, and the current mode. The Lyapunov method is combined to adjust the quadratic programming and sliding mode control parameters online to achieve convergence of attitude angle and position.

Images