Intelligent automobile control system and method based on artificial intelligence

Abstract

The invention discloses an intelligent automobile control system and method based on artificial intelligence, and relates to the technical field of new energy automobile intelligent control. Acquiring an electroencephalogram signal of a target driver and performing intention feature extraction to obtain a driving intention instruction; acquiring multi-modal data and extracting environment sensing features to obtain environment sensing parameters; inputting the driving intention instruction and the environment perception parameters into a multi-modal large model for training to obtain a semantic feature instruction, and inputting the semantic feature instruction into an intelligent agent decision model to obtain a new energy vehicle control instruction; and the new energy vehicle is controlled according to the control instruction evaluation result. The brain-computer interface accurately analyzes the intention of the driver, integrates multi-modal environment perception data to realize high-precision scene understanding, and generates a control instruction in combination with an agent decision; the dynamic optimization and safety redundancy mechanism ensures the instruction reliability, the closed-loop feedback continuously optimizes the model, and the control intuition, the decision safety and the system adaptability are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control technology for new energy vehicles, and in particular to an intelligent vehicle control system and method based on artificial intelligence. Background Art

[0002] With the development of intelligent driving technology, traditional physical control and interaction methods such as voice and gestures have become more dependent on physical control and environmental interference. Brain-computer interfaces (BCIs) enable "mind-driven control" by reading brain waves, offering new possibilities for people with disabilities. However, they still face challenges in controlling new energy vehicles, such as noise interference, low recognition accuracy, and insufficient real-time performance.

[0003] Patent number: CN119271050A, discloses a brain-computer interface device and its control method, medium and electronic device; obtains brain signals generated by a subject under any stimulation state; performs feature recognition on the brain signal to obtain brain response features contained in the brain signal; performs instruction recognition under the brain response features on the brain signal to obtain a target instruction represented by the brain signal; configures a preset execution probability based on the target instruction, and outputs the target instruction based on the execution probability to control an external controlled device. This application can identify the target instruction from the brain signal in real time by identifying the response features of the brain signal and the instructions under the response features, and output the target instruction in real time to control the external controlled device, which can meet the needs of application fields with high real-time requirements.

[0004] In the existing technology, controlling new energy vehicles through brain-computer interfaces has the problems of low command recognition accuracy and low real-time performance, and the control effect of brain-computer signals on new energy vehicles is poor and lacks precision. Summary of the Invention

[0005] The purpose of this invention is to solve the problems of low command recognition accuracy and low real-time performance when controlling new energy vehicles through brain-computer interfaces, and the poor control effect of brain-computer signals on new energy vehicles and lack of precision, and to propose an intelligent automobile control system and method based on artificial intelligence.

[0006] The purpose of the present invention can be achieved through the following technical solutions: First, an intelligent automobile control system based on artificial intelligence is proposed. The system includes: an EEG signal acquisition module, an environment perception module, a new energy vehicle control instruction generation module, and a new energy vehicle control module: The EEG signal acquisition module is configured to acquire an EEG signal of a target driver, perform a preset operation on the EEG signal to obtain a target EEG signal set, fuse the EEG signals in the target EEG signal set to obtain a final EEG signal, and perform intention feature extraction on the final EEG signal to obtain a driving intention instruction; the preset operation is a first preset operation and a second preset operation; and the target EEG signal set includes a first target EEG signal and a second target EEG signal; The environmental perception module is used to acquire multimodal data and extract environmental perception features from the multimodal data to obtain environmental perception parameters; the multimodal data includes: image data, radar data, ultrasonic data and driving data; The new energy vehicle control instruction generation module is used to input the driving intention instruction and the environmental perception parameter into the multimodal large model for training to obtain semantic feature instructions, and input the semantic feature instructions into the intelligent agent decision model to obtain the new energy vehicle control instruction; The new energy vehicle control module is used to verify the correctness of the new energy vehicle control instructions to obtain an evaluation result, and control the new energy vehicle according to the evaluation result.

[0007] Optionally, the new energy vehicle control instruction generation module includes: a real-time state analysis module, a membership determination module, a fuzzy rule generation module and a driving instruction generation module: The real-time state analysis module is used to obtain the real-time state of the new energy vehicle and use the real-time state and the semantic feature instruction as input variables of the fuzzy controller to obtain an output variable; the output variable is the steering wheel angle, and the real-time state includes: lateral deviation and deviation rate; The membership determination module is configured to generate an input fuzzy set based on the input variables, generate an output fuzzy set based on the output variables, and determine membership functions for the input fuzzy set and the output fuzzy set; the membership functions are configured to quantify the membership of the input variables and the output variables in the respective fuzzy sets; The fuzzy rule generation module is used to determine fuzzy rules, perform fuzzy rule matching according to the membership of input variables to obtain multiple matching fuzzy rules, and calculate the output variables according to each matching fuzzy rule to obtain multiple fuzzy values; The driving instruction generation module is used to perform a synthesis operation based on each fuzzy value to obtain a target fuzzy output, and defuzzify the target fuzzy output to obtain a new energy vehicle control instruction; the new energy vehicle control instruction is used to adjust the turning angle of the new energy vehicle.

[0008] Optionally, the new energy vehicle control module includes: a transfer function determination module and an evaluation result calculation module: The transfer function determination module is used to establish the state of the new energy vehicle and construct an event set, and determine the transfer function according to the state of the new energy vehicle and the event set; the state of the new energy vehicle includes: the degree of left turn, the degree of straight driving, and the degree of right turn; the event set includes: controllable events and uncontrollable events; The evaluation result calculation module is used to determine the state of the new energy vehicle and define constraints based on the new energy vehicle control instructions, calculate the evaluation results based on the constraints and the controllable events, and determine the control logic based on the evaluation results; the evaluation results include: correct, wrong and serious error.

[0009] Optionally, the evaluation result calculation module further includes: a first judgment module, a second judgment module and a third judgment module: The first judgment module is configured to determine that the evaluation result is correct if the evaluation result is greater than or equal to a first threshold, and then execute a driving intention instruction; The second judgment module is configured to determine that the evaluation result is wrong if the first threshold value > the evaluation result ≥ the second threshold value, and reduce the weight of the automatic decision; The third judgment module is configured to determine that the evaluation result is a serious error if the second threshold value is greater than the evaluation result, and to perform instruction confirmation through a feedback interface.

[0010] Optionally, the system further includes: a data uploading module and a data updating module: The data upload module is used to periodically obtain local operation data, encrypt the local operation data to obtain an encrypted package, and upload the encrypted package to the cloud; The data update module is used to periodically receive update packages from the cloud and update the multimodal large model.

[0011] An intelligent vehicle control method based on artificial intelligence is proposed, the method comprising: Acquiring an EEG signal of a target driver, performing a preset operation on the EEG signal to obtain a target EEG signal set, fusing the EEG signals in the target EEG signal set to obtain a final EEG signal, and performing intention feature extraction on the final EEG signal to obtain a driving intention instruction; the preset operation being a first preset operation and a second preset operation; and the target EEG signal set including the first target EEG signal and the second target EEG signal; Acquiring multimodal data, and extracting environmental perception features from the multimodal data to obtain environmental perception parameters; the multimodal data includes: image data, radar data, ultrasonic data, and driving data; Inputting the driving intention instruction and the environmental perception parameter into a multimodal large model for training to obtain a semantic feature instruction, and inputting the semantic feature instruction into an intelligent agent decision model to obtain a new energy vehicle control instruction; The new energy vehicle control command is verified for correctness to obtain an evaluation result, and the new energy vehicle is controlled according to the evaluation result.

[0012] Optionally, inputting the semantic feature instruction into the agent decision model to obtain the new energy vehicle control instruction includes: Acquire the real-time state of the new energy vehicle, and use the real-time state and the semantic feature instruction as input variables of a fuzzy controller to obtain an output variable; the output variable is a steering wheel angle, and the real-time state includes: lateral deviation and deviation rate; generating an input fuzzy set based on the input variables and generating an output fuzzy set based on the output variables, and determining membership functions for the input fuzzy set and the output fuzzy set; wherein the membership functions are used to quantify the membership of the input variables and the output variables in the respective fuzzy sets; Determine fuzzy rules, perform fuzzy rule matching according to the membership of input variables to obtain multiple matching fuzzy rules, and calculate the output variables according to each matching fuzzy rule to obtain multiple fuzzy values; A synthesis operation is performed on each fuzzy value to obtain a target fuzzy output, and the target fuzzy output is defuzzified to obtain a new energy vehicle control instruction; the new energy vehicle control instruction is used to adjust the turning angle of the new energy vehicle.

[0013] Optionally, performing correctness verification on the new energy vehicle control instruction to obtain an evaluation result includes: Establishing a new energy vehicle state and constructing an event set, and determining a transfer function based on the new energy vehicle state and the event set; the new energy vehicle state includes: left turn degree, straight driving degree, and right turn degree; and the event set includes: controllable events and uncontrollable events; Determine the state of the new energy vehicle and define constraints according to the new energy vehicle control instructions, calculate an evaluation result according to the constraints and the controllable events, and determine the control logic according to the evaluation result; the evaluation results include: correct, wrong and serious error.

[0014] Optionally, determining the control logic according to the evaluation result includes: If the evaluation result is greater than or equal to the first threshold, the evaluation result is determined to be correct, and the driving intention instruction is executed; If the first threshold value > the evaluation result ≥ the second threshold value, the evaluation result is determined to be wrong, and the weight of the automatic decision is reduced; If the second threshold value is greater than the evaluation result, the evaluation result is determined to be a serious error, and an instruction confirmation is performed through the feedback interface.

[0015] Optionally, after controlling the new energy vehicle according to the evaluation result, the method further includes: Periodically acquiring local operating data, encrypting the local operating data to obtain an encrypted package, and uploading the encrypted package to the cloud; Periodically receive cloud update packages to update large multimodal models.

[0016] Beneficial effects of the present invention: This invention proposes an artificial intelligence-based intelligent vehicle control method. This method obtains the target driver's EEG signals and extracts intent features to obtain driving intention instructions. It also obtains multimodal data and extracts environmental perception features to obtain environmental perception parameters. The driving intention instructions and environmental perception parameters are then input into a large multimodal model for training to obtain semantic feature instructions, which are then input into an intelligent agent decision model to obtain new energy vehicle control instructions. The correctness of the new energy vehicle control instructions is verified to obtain an evaluation result, and the new energy vehicle is controlled based on the evaluation result. A brain-computer interface accurately analyzes the driver's intent, integrates multimodal environmental perception data to achieve high-precision scene understanding, and generates control instructions in conjunction with intelligent agent decision-making. Dynamic optimization and safety redundancy mechanisms ensure instruction reliability, and closed-loop feedback continuously optimizes the model, significantly improving intuitive control, decision-making safety, and system adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic structural diagram of an artificial intelligence-based intelligent vehicle control system provided by an embodiment of the present invention; Figure 2 A flowchart of an artificial intelligence-based intelligent vehicle control method provided by an embodiment of the present invention; Figure 3 A flowchart of brain-computer interface signal processing for an artificial intelligence-based intelligent vehicle control method provided by an embodiment of the present invention; Figure 4 A multimodal feature fusion flow chart of an artificial intelligence-based intelligent vehicle control method provided by an embodiment of the present invention; Figure 5 A control mode flow chart of an artificial intelligence-based intelligent vehicle control method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Based on the same inventive concept, the present invention also provides an intelligent automobile control system based on artificial intelligence. Figure 1 , Figure 1 A schematic diagram of the structure of an artificial intelligence-based intelligent vehicle control system provided by an embodiment of the present invention includes: an EEG signal acquisition module, an environmental perception module, a new energy vehicle control instruction generation module, and a new energy vehicle control module: an EEG signal acquisition module, configured to acquire EEG signals of a target driver, perform a preset operation on the EEG signals to obtain a target EEG signal set, fuse the EEG signals in the target EEG signal set to obtain a final EEG signal, and perform intention feature extraction on the final EEG signal to obtain a driving intention instruction; the preset operation being a first preset operation and a second preset operation; and the target EEG signal set comprising the first target EEG signal and the second target EEG signal; The environmental perception module is used to obtain multimodal data and extract environmental perception features from the multimodal data to obtain environmental perception parameters; A new energy vehicle control instruction generation module is used to input driving intention instructions and environmental perception parameters into a multimodal large model for training to obtain semantic feature instructions, and then input the semantic feature instructions into the intelligent agent decision model to obtain new energy vehicle control instructions; The new energy vehicle control module is used to verify the correctness of the new energy vehicle control instructions to obtain an evaluation result, and control the new energy vehicle according to the evaluation result.

[0020] Multimodal data includes: image data, radar data, ultrasonic data and driving data; An artificial intelligence-based intelligent vehicle control system provided by an embodiment of the present invention accurately analyzes the driver's intentions through a brain-computer interface, integrates multimodal environmental perception data to achieve high-precision scene understanding, and generates control instructions in combination with intelligent agent decision-making; dynamic optimization and safety redundancy mechanisms ensure the reliability of instructions, and closed-loop feedback continuously optimizes the model, significantly improving the intuitiveness of control, decision-making safety and system adaptability.

[0021] In one embodiment, the new energy vehicle control instruction generation module includes: a real-time state analysis module, a membership determination module, a fuzzy rule generation module and a driving instruction generation module: The real-time state analysis module is used to obtain the real-time state of the new energy vehicle and use the real-time state and semantic feature instructions as input variables of the fuzzy controller to obtain the output variable; the output variable is the steering wheel angle, and the real-time state includes: lateral deviation and deviation rate; A membership determination module is used to generate input fuzzy sets based on input variables, generate output fuzzy sets based on output variables, and determine membership functions for the input fuzzy sets and output fuzzy sets; the membership functions are used to quantify the membership of the input variables and output variables in each fuzzy set; A fuzzy rule generation module is used to determine fuzzy rules, perform fuzzy rule matching based on the membership of input variables to obtain multiple matching fuzzy rules, and calculate output variables based on each matching fuzzy rule to obtain multiple fuzzy values; The driving instruction generation module is used to perform a synthesis operation based on each fuzzy value to obtain a target fuzzy output, and defuzzify the target fuzzy output to obtain a new energy vehicle control instruction; the new energy vehicle control instruction is used to adjust the turning angle of the new energy vehicle.

[0022] In one embodiment, the new energy vehicle control module includes: a transfer function determination module and an evaluation result calculation module: A transfer function determination module is used to establish the state of the new energy vehicle and construct an event set, and determine the transfer function based on the new energy vehicle state and the event set; the new energy vehicle state includes: left turn degree, straight driving degree and right turn degree; the event set includes: controllable events and uncontrollable events; The evaluation result calculation module is used to determine the state of the new energy vehicle and define the constraints according to the new energy vehicle control instructions, calculate the evaluation results according to the constraints and controllable events, and determine the control logic according to the evaluation results; the evaluation results include: correct, wrong and serious error.

[0023] In one embodiment, the evaluation result calculation module further includes: a first judgment module, a second judgment module and a third judgment module: A first judgment module is configured to determine that the evaluation result is correct if the evaluation result is greater than or equal to a first threshold, and execute a driving intention instruction; A second judgment module is configured to determine that the evaluation result is wrong if the first threshold value > the evaluation result ≥ the second threshold value, and reduce the weight of the automatic decision; The third judgment module is used to determine that the evaluation result is a serious error if the second threshold value is greater than the evaluation result, and to perform instruction confirmation through the feedback interface.

[0024] In one embodiment, the system further includes: a data uploading module and a data updating module: The data upload module is used to periodically obtain local operation data, encrypt the local operation data to obtain an encrypted package, and upload the encrypted package to the cloud; The data update module is used to periodically receive cloud update packages and update the multimodal large model.

[0025] The embodiment of the present invention provides an intelligent vehicle control method based on artificial intelligence. Figure 2 , Figure 2 A flowchart of an artificial intelligence-based intelligent vehicle control method provided in an embodiment of the present invention. The method includes the following steps: S101, obtaining EEG signals of a target driver, performing a preset operation on the EEG signals to obtain a target EEG signal set, fusing the EEG signals in the target EEG signal set to obtain a final EEG signal, and extracting intention features from the final EEG signal to obtain a driving intention instruction; S102, acquiring multimodal data, and extracting environmental perception features from the multimodal data to obtain environmental perception parameters; S103: Inputting the driving intention instruction and the environmental perception parameters into a multimodal large model for training to obtain a semantic feature instruction, and inputting the semantic feature instruction into an intelligent agent decision model to obtain a new energy vehicle control instruction; S104 , verifying the correctness of the new energy vehicle control command to obtain an evaluation result, and controlling the new energy vehicle according to the evaluation result.

[0026] The preset operation is a first preset operation and a second preset operation; the target EEG signal set includes a first target EEG signal and a second target EEG signal; the multimodal data includes: image data, radar data, ultrasonic data and driving data; An artificial intelligence-based intelligent vehicle control method provided by an embodiment of the present invention accurately analyzes the driver's intentions through a brain-computer interface, integrates multimodal environmental perception data to achieve high-precision scene understanding, and generates control instructions in combination with intelligent agent decision-making; dynamic optimization and safety redundancy mechanisms ensure the reliability of instructions, and closed-loop feedback continuously optimizes the model, significantly improving the intuitiveness of control, decision-making safety and system adaptability.

[0027] In one implementation, preset operations are performed on the EEG signal to obtain a target EEG signal set, such as discrete wavelet transform enhancement and empirical mode decomposition, to obtain a first target EEG signal and a second target EEG signal, respectively; wherein, discrete wavelet transform enhancement: the EEG signal is decomposed to obtain sub-signals of different frequency bands, that is, decomposed into approximate coefficients (low-frequency components, reflecting the overall trend of the signal) and detail coefficients (high-frequency components, reflecting the local fluctuations of the signal), and EEG signals of the same category are combined to obtain target coefficients (the database contains standard approximate coefficients and standard detail coefficients, and the standard approximate coefficients are combined with the target user (driver) detail coefficients, or the standard detail coefficients are combined with the target user (driver) approximate coefficients), and individual differences are addressed by fusing signal features of different subjects; an inverse discrete wavelet transform is performed on the reorganized target coefficients to generate enhanced source subject signals and target subject signals, and finally a first target EEG signal enhanced in the time domain is obtained.

[0028] In one implementation, EMD is used to decompose EEG signals, yielding a set of intrinsic mode functions (IMFs) and residuals. The IMFs represent sub-signals in different frequency bands (reflecting the local characteristics of the signal within these frequency ranges), while the residuals represent the overall trend of the signal. For EEG signals of the same category, standard IMFs and the IMFs of the target user (driver) are combined, integrating time-frequency information across subjects to account for individual differences. The recombined IMFs are summed with the corresponding residuals to generate a second EEG signal enhanced in the time domain. The first target EEG signal and the second EEG signal are then fused to yield the final EEG signal.

[0029] In one implementation, a non-invasive brain-computer interface module is used to collect the driver's EEG signals in real time (e.g., a dry electrode EEG headset or an implantable brain-computer interface system for collecting the driver's brain activity signals in real time, see Figure 3 , Figure 3 This is a brain-computer interface signal processing flow chart for an AI-based intelligent vehicle control method provided by an embodiment of the present invention; wherein Butterworth is a Butterworth filter, and adaptive feature extraction techniques (such as wavelet transform or machine learning algorithms) are used to generate driving intention instructions. This allows drivers to directly control new energy vehicles through brain signals, particularly providing barrier-free driving for people with physical disabilities. It also reduces reliance on physical movements during conventional driving, significantly improving intuitive control.

[0030] In one implementation, see Figure 4 , Figure 4A multimodal feature fusion flow chart for an AI-based intelligent vehicle control method provided by an embodiment of the present invention. Cross-Attention stands for cross-attention, and Perceiver is a general framework for multimodal learning. A multimodal deep learning model (i.e., a "multimodal large model") can be constructed using a neural network architecture such as the Transformer and pre-trained on massive amounts of driving scene data and human brain intention signal data. During operation, the multimodal fusion module takes brain-computer interface (BCI) features and environmental perception features as input, performs feature association and semantic fusion through the model's multi-layer attention mechanism, and outputs a comprehensive semantic representation of the current scene. For example, at a given moment, if BCI features indicate a driver's lane change intention, while environmental perception features indicate a nearby new energy vehicle in the left lane, the multimodal large model can fuse these two pieces of information and output a high-level semantic representation of "driver intends to change lanes left, but there is a vehicle on the left." This representation can include the driver's current decision intention, urgency assessment, key environmental factors related to the intention (such as new energy vehicles, pedestrians, road signs, etc.), and risk prediction.

[0031] In one implementation, multimodal data from on-board cameras, radars, ultrasonic sensors, and other sensors (multiple on-board cameras for acquiring images in front of and around new energy vehicles; millimeter-wave radars and lidars for detecting the distance and speed of obstacles around new energy vehicles; ultrasonic sensors for close-range distance measurement (for example, parking assistance); and new energy vehicle internal state sensors, such as steering wheel angle sensors and pedal position sensors, to obtain the driver's current operating behavior, i.e., driving data) are integrated with deep learning models (such as convolutional neural networks) to extract environmental perception feature instructions (instructions, i.e., feature vectors). This technology achieves all-round dynamic perception of complex driving scenarios, including obstacle distance, lane line recognition, traffic sign detection, etc., ensuring that the system can still accurately understand the environmental status under complex road conditions, providing a reliable data basis for subsequent decision-making.

[0032] In one implementation, a large multimodal model (such as the Transformer architecture) performs semantic fusion of driving intent commands and environmental perception parameters, generating a comprehensive representation that includes intent priority, environmental risks, and scene semantics. Based on this, an agent decision model (such as a reinforcement learning agent) generates control commands that prioritize driver intent while dynamically adjusting strategies based on safety rules. This technology achieves deep collaboration between human and machine intent, preserving the driver's subjective decision-making ability while ensuring the safety and adaptability of the autonomous driving system.

[0033] In one implementation, the correctness of control commands is verified through an evaluation mechanism, and dynamic logic adjustments are made based on the scoring results to generate corrected control commands. This technology combines a redundant control architecture (such as a dual-channel switching mechanism) with a safety monitoring module. In the event of a brain-computer interface anomaly or command conflict, it automatically triggers a safety mode or requests manual takeover, ensuring that new energy vehicles remain under control and significantly improving the system's fault tolerance. This control system is for assisted driving, and all operations require secondary verification with the driver via voice, display, and other devices before execution.

[0034] In one implementation method, the multimodal large model and intelligent agent decision model are continuously optimized through the federated learning mechanism, and data sharing across new energy vehicles is achieved through edge computing and cloud aggregation. The model's adaptability to different drivers' brain signal characteristics and diverse road scenarios is gradually improved, enabling the system to have self-evolution capabilities and maintain technological advancement and personalized service capabilities in the long term.

[0035] In one embodiment, inputting the semantic feature instruction into the agent decision model to obtain the new energy vehicle control instruction includes: The real-time status of the new energy vehicle is obtained, and the real-time status and semantic feature instructions are used as input variables of the fuzzy controller to obtain the output variable; the output variable is the steering wheel angle, and the real-time status includes: lateral deviation and deviation rate; Generate input fuzzy sets based on input variables, generate output fuzzy sets based on output variables, and determine membership functions for input fuzzy sets and output fuzzy sets; membership functions are used to quantify the membership of input variables and output variables in each fuzzy set; Determine the fuzzy rules, perform fuzzy rule matching according to the membership of the input variables to obtain multiple matching fuzzy rules, and calculate the output variables according to each matching fuzzy rule to obtain multiple fuzzy values; A synthesis operation is performed according to each fuzzy value to obtain a target fuzzy output, and the target fuzzy output is defuzzified to obtain a new energy vehicle control instruction; the new energy vehicle control instruction is used to adjust the turning angle of the new energy vehicle.

[0036] In one implementation, the distance that a new energy vehicle deviates from the road centerline (lateral deviation), the deviation rate, and the semantic feature instruction are selected as input variables, and the steering angle is selected as the output variable; the lateral position and deviation rate of the new energy vehicle are key factors affecting driving stability, and these two variables can be used to monitor the driving status of the new energy vehicle in real time; the steering angle is a direct parameter for controlling the driving direction of the new energy vehicle, providing the necessary input information for the fuzzy controller, enabling it to generate corresponding control decisions based on the current status of the new energy vehicle.

[0037] In one implementation, an input fuzzy set is generated based on the input variables, and an output fuzzy set is generated based on the output variables. Fuzzy sets of the input variables f (lateral deviation) and ef (deviation rate) are defined, as well as a fuzzy set of the output variable v (steering angle). For example, the fuzzy set of f is {MC, MO, MT, AP, QT, QO, QC}, which represent "large left deviation", "medium left deviation", "small left deviation", "zero", "small right deviation", "medium right deviation", and "large right deviation", respectively. The fuzzy set of ef is {MC, MO, MT, AP, QT, QO, QC}, which represent "large left deviation change rate", respectively. , "middle left deviation rate of change", "small left deviation rate of change", "zero change rate", "small right deviation rate of change", "middle right deviation rate of change", "large right deviation rate of change"; the fuzzy set of v is: {MC, MO, MT, AP, QT, QO, QC}, which respectively represent "large left turn", "middle left turn", "small left turn", "zero turning angle", "small right turn", "middle right turn", and "large right turn"; fuzzy sets are used to convert precise numerical inputs into fuzzy states so that the fuzzy controller can handle uncertainty and fuzziness, provide a basis for fuzzy reasoning, and enable the controller to generate reasonable control decisions based on fuzzy rules.

[0038] In one implementation, membership functions are determined for input and output fuzzy sets. Membership functions are used to quantify the degree of membership of input and output variables in respective fuzzy sets and are the basis of fuzzy reasoning. Converting precise input values into fuzzy states enables the fuzzy controller to reason according to fuzzy rules. Membership function of f (lateral deviation), for example: Membership MC MO MT AP QT QO QC 0 1 0 0 0 0 0 0 0.2 0 0 0 0 0 1 1 0.4 -2 0 0 2.5 2 5 2 In one implementation, a fuzzy rule is determined, for example: f\ef MC MO MT AP QT QO QC MC QC QC QO QO QT QT AP MO QC QO QO QT QT AP MT MT QO QO QT QT AP MT MT AP QO QT QT AP MT MT MO QT QT QT AP MT MT MO MO QO QT AP MT MT MO MO MC QC AP MT MT MO MO MC MC In one implementation, the fuzzy value of the output variable v is calculated based on the matching fuzzy rules. For example, if the input variables f and ef belong to fuzzy sets B and C, respectively, the fuzzy value D of the output variable v is as follows: D = (B * C) ∘ S, where ∘ represents a compound operation and S represents a fuzzy rule table. A composite operation is performed on each fuzzy value to obtain the target fuzzy output. The maximum-minimum method (max-min) is used to perform a composite operation on all matching fuzzy rules to obtain the final fuzzy output: μV(A)=max{min[μB(x),μC(y)]} Fuzzy reasoning is the core link of the fuzzy controller, which is used to generate fuzzy values of output variables based on fuzzy rules and the fuzzy state of input variables. Through fuzzy reasoning, the controller can generate reasonable control decisions based on the state of input variables and realize dynamic control of new energy vehicles.

[0039] In one implementation, the target fuzzy output is defuzzified to obtain a control decision value, and the fuzzy output value obtained by fuzzy inference is converted into an accurate control decision value: Among them, v represents the precise control decision value after defuzzification, μ V (An) is the output variable v in the fuzzy set A n The membership degree in A n represents an element in the fuzzy set of the output variable v, and N represents the total number of elements in the fuzzy set of the output variable v; The result of fuzzy reasoning is a fuzzy value, while the actual control system requires precise numerical values as control instructions. The fuzzy output value is converted into a precise control decision value so that the controller can output specific control instructions, thereby realizing precise control of the new energy vehicle; according to the precise control decision value v after defuzzification, the steering angle of the new energy vehicle is adjusted. The steering angle is a key parameter for the lateral control of the new energy vehicle. By adjusting the steering angle, the driving trajectory of the new energy vehicle can be precisely controlled. The output of the controller directly affects the driving state of the new energy vehicle. Through precise steering angle adjustment, precise control of the new energy vehicle can be achieved so that it can travel according to the predetermined trajectory.

[0040] In one embodiment, correctness verification of the new energy vehicle control instructions is performed to obtain an evaluation result, including: Establish the state of new energy vehicles and construct an event set, and determine the transfer function based on the state of new energy vehicles and the event set; the state of new energy vehicles includes: left turn degree, straight go degree and right turn degree, and the event set includes: controllable events and uncontrollable events; Determine the state of the new energy vehicle and define constraints based on the new energy vehicle control instructions, calculate the evaluation results based on the constraints and controllable events, and determine the control logic based on the evaluation results; the evaluation results include: correct, wrong and serious error.

[0041] In one implementation, a new energy vehicle state R is established to describe the lateral movement of the brain-controlled new energy vehicle system, R={r L ,r D ,r R}, the new energy vehicle status includes: left turn degree r L , straightness degree r D and right turn degree r R ; Event set, define event set Q, including controllable event Q kk and uncontrollable event Q bk , controllable event Q kk Including receiving the driver's brain control command (turn left, keep direction, turn right), uncontrollable event Q bkIndicates the interval during which no brain control command is received; Q=Q kk ∪Q bk ,Q kk ∩Q bk =∅,Q kk ={q L ,q D ,q R}, where q L ,q D ,q R They respectively represent the reception of brain-controlled commands for turning left, going straight, and turning right; the transfer function is determined according to the state of the new energy vehicle and the event set, the transfer function: R*Q→R'. Through fuzzy reasoning, the system can dynamically adjust the state according to the current state and input events, thereby realizing real-time supervision of brain-controlled commands.

[0042] In one implementation, F(q)=μr(q), where F(q) is a constraint term, μr(q) is the product of the membership function value obtained by fuzzy reasoning of the deviation f and the deviation rate ef, and takes a value between [0,1]; if the evaluation result is ≥0.7 (the first threshold), the evaluation result is judged to be correct; if 0.3≤evaluation result<0.7, the evaluation result is judged to be incorrect; if the evaluation result is <0.3, the evaluation result is judged to be seriously incorrect.

[0043] In one implementation, the evaluation result is calculated based on the constraints and controllable events, for example: R(q)=F(q)∩Q kk (q), where R(q) represents the updated state of the new energy vehicle, F(q) represents the instruction correctness constraint, and Q kk (q) represents the current event (command). The transition function combines the driver's command with the vehicle's real-time status to determine whether to accept the command or trigger a correction mechanism. For example, if a command causes the vehicle to continuously stray from its lane, the transition function might ignore the command and switch to autonomous control mode.

[0044] In one implementation, by establishing a set of new energy vehicle states and events and determining a transfer function based on the state and event set, the system can dynamically adjust the control state of the new energy vehicle based on the current state and input events. This dynamic adjustment capability enables the system to respond quickly when receiving brain control commands and remain stable when no commands are received. If the event set is an uncontrollable event (when the intention information is uncertain or the signal is lost), a voice prompt is issued to the target driver. If no response is received within a preset time period, the vehicle speed is gradually reduced to a stop; In one implementation, see Figure 5 , Figure 5A control mode flow chart of an artificial intelligence-based intelligent vehicle control method provided in an embodiment of the present invention, wherein SNR represents signal-to-noise ratio; the brain-computer interface control channel and the traditional manual control channel exist in parallel in the control of new energy vehicles, and are always on standby to receive control signals directly input by the driver through the steering wheel and pedals. Under normal circumstances, the intelligent decision-making unit controls according to the brain-computer interface and the autonomous driving strategy; once an abnormal brain-computer interface signal is detected (for example, the driver's decreased attention leads to invalid EEG characteristics) or the system detects that the intelligent agent's decision may lead to an unsafe situation, the system will automatically switch the control weight: remind the driver to intervene and take over, allowing the new energy vehicle control unit to temporarily operate in the autonomous driving safety mode (such as slowing down and stopping). This dual-channel redundancy ensures that even if the new interaction method fails, the new energy vehicle can still be controlled by conventional means without the risk of loss of control.

[0045] In one embodiment, determining the control logic according to the evaluation result includes: If the evaluation result is greater than or equal to the first threshold, the evaluation result is determined to be correct, and the driving intention instruction is executed; If the first threshold > the evaluation result ≥ the second threshold, the evaluation result is determined to be wrong, and the weight of the automatic decision is reduced; If the second threshold value is greater than the evaluation result, the evaluation result is determined to be a serious error, and an instruction confirmation is performed through the feedback interface.

[0046] In one implementation, the control logic is executed by an intelligent decision-making unit, and specific new energy vehicle control decisions and interaction strategies are generated based on the fused semantic representation. By combining the driver's intention priority information, the intelligent decision-making unit can quickly generate corresponding instructions when it detects that the driver has a clear control instruction intention, such as a strong braking intention. This mechanism not only improves the system's response speed and accuracy to the driver's intention, but also enhances the ability of human-machine collaborative decision-making, so that the control behavior of new energy vehicles can better reflect the driver's subjective intention. In one implementation, the intelligent decision-making unit considers both driving mission objectives and safety rules when generating control decisions. When human-machine intentions conflict, the intelligent agent can initiate a negotiation strategy, such as asking the driver to confirm an action through a feedback interface, or partially executing the driver's instructions while ensuring safety. This safety-first decision-making mechanism ensures that new energy vehicles adhere to safe driving principles in all circumstances. Even when driver intentions conflict with safety rules, the system can verify or adjust instructions to ensure driving safety, thereby improving system reliability and safety.

[0047] In one implementation, the intelligent decision-making unit can output low-level control instructions, such as steering angle, target speed, and braking intensity, or higher-level strategic decisions (such as lane change recommendations or distance maintenance). This flexibility enables the system to dynamically adjust control strategies based on a comprehensive representation of the current scenario and the driving task objectives. For example, in complex traffic environments, the system can recommend lane changes to avoid potential hazards or adjust speed to maintain a safe distance. This strategic adjustment capability not only improves the system's adaptability but also enhances its performance and reliability in diverse driving scenarios.

[0048] In one embodiment, after controlling the new energy vehicle according to the evaluation result, the method further includes: Periodically obtain local operating data, encrypt the local operating data to obtain an encrypted package, and upload the encrypted package to the cloud; Periodically receive cloud update packages to update large multimodal models.

[0049] In one implementation method, each new energy vehicle regularly uses locally collected brain-computer interface data and driving data to train the model and obtain local model update parameters; the local update parameters are uploaded to the server for aggregation calculation to obtain a global model update; the updated model is sent to each new energy vehicle for subsequent control, thereby gradually improving the model's adaptability to different drivers and diverse road environments.

[0050] In one implementation, federated learning technology continuously updates a large multimodal model and an intelligent agent decision model to adapt to the brain-computer interface signal characteristics of different drivers and road conditions in different regions. Each vehicle, acting as an edge node, regularly collects local operating data for model fine-tuning and training, and uploads encrypted model updates (gradients or parameters) to a cloud-based federated server for aggregation. The federated server aggregates the updates from multiple vehicles, calculates global model improvements, and distributes the updated model parameters to each new energy vehicle, improving system robustness.

[0051] It should be noted that, in this document, terms such as "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements that are inherent to such process, method, article or apparatus.

[0052] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. An intelligent automobile control system based on artificial intelligence, characterized in that: The system includes: an EEG signal acquisition module, an environment perception module, a new energy vehicle control instruction generation module, and a new energy vehicle control module: The EEG signal acquisition module is configured to acquire an EEG signal of a target driver, perform a preset operation on the EEG signal to obtain a target EEG signal set, fuse the EEG signals in the target EEG signal set to obtain a final EEG signal, and perform intention feature extraction on the final EEG signal to obtain a driving intention instruction; the preset operation is a first preset operation and a second preset operation; and the target EEG signal set includes a first target EEG signal and a second target EEG signal; The environmental perception module is used to acquire multimodal data and extract environmental perception features from the multimodal data to obtain environmental perception parameters; the multimodal data includes: image data, radar data, ultrasonic data and driving data; The new energy vehicle control instruction generation module is used to input the driving intention instruction and the environmental perception parameter into the multimodal large model for training to obtain semantic feature instructions, and input the semantic feature instructions into the intelligent agent decision model to obtain the new energy vehicle control instruction; The new energy vehicle control module is used to verify the correctness of the new energy vehicle control instructions to obtain an evaluation result, and control the new energy vehicle according to the evaluation result.

2. The intelligent automobile control system based on artificial intelligence according to claim 1, characterized in that: The new energy vehicle control instruction generation module includes: a real-time state analysis module, a membership determination module, a fuzzy rule generation module and a driving instruction generation module: The real-time state analysis module is used to obtain the real-time state of the new energy vehicle and use the real-time state and the semantic feature instruction as input variables of the fuzzy controller to obtain an output variable; the output variable is the steering wheel angle, and the real-time state includes: lateral deviation and deviation rate; The membership determination module is configured to generate an input fuzzy set based on the input variables, generate an output fuzzy set based on the output variables, and determine membership functions for the input fuzzy set and the output fuzzy set; the membership functions are configured to quantify the membership of the input variables and the output variables in the respective fuzzy sets; The fuzzy rule generation module is used to determine fuzzy rules, perform fuzzy rule matching according to the membership of input variables to obtain multiple matching fuzzy rules, and calculate the output variables according to each matching fuzzy rule to obtain multiple fuzzy values; The driving instruction generation module is used to perform a synthesis operation based on each fuzzy value to obtain a target fuzzy output, and defuzzify the target fuzzy output to obtain a new energy vehicle control instruction; the new energy vehicle control instruction is used to adjust the turning angle of the new energy vehicle.

3. The intelligent automobile control system based on artificial intelligence according to claim 1, characterized in that: The new energy vehicle control module includes: a transfer function determination module and an evaluation result calculation module: The transfer function determination module is used to establish the state of the new energy vehicle and construct an event set, and determine the transfer function according to the state of the new energy vehicle and the event set; the state of the new energy vehicle includes: the degree of left turn, the degree of straight driving, and the degree of right turn; the event set includes: controllable events and uncontrollable events; The evaluation result calculation module is used to determine the state of the new energy vehicle and define constraints based on the new energy vehicle control instructions, calculate the evaluation results based on the constraints and the controllable events, and determine the control logic based on the evaluation results; the evaluation results include: correct, wrong and serious error.

4. The intelligent automobile control system based on artificial intelligence according to claim 3, characterized in that: The evaluation result calculation module includes: a first judgment module, a second judgment module and a third judgment module: The first judgment module is configured to determine that the evaluation result is correct if the evaluation result is greater than or equal to a first threshold, and then execute a driving intention instruction; The second judgment module is configured to determine that the evaluation result is wrong if the first threshold value > the evaluation result ≥ the second threshold value, and reduce the weight of the automatic decision; The third judgment module is configured to determine that the evaluation result is a serious error if the second threshold value is greater than the evaluation result, and to perform instruction confirmation through a feedback interface.

5. The intelligent automobile control system based on artificial intelligence according to claim 1, characterized in that: The system also includes: a data uploading module and a data updating module: The data upload module is used to periodically obtain local operation data, encrypt the local operation data to obtain an encrypted package, and upload the encrypted package to the cloud; The data update module is used to periodically receive update packages from the cloud and update the multimodal large model.

6. An intelligent vehicle control method based on artificial intelligence, characterized in that: The method comprises: Acquiring an EEG signal of a target driver, performing a preset operation on the EEG signal to obtain a target EEG signal set, fusing the EEG signals in the target EEG signal set to obtain a final EEG signal, and performing intention feature extraction on the final EEG signal to obtain a driving intention instruction; the preset operation being a first preset operation and a second preset operation; and the target EEG signal set including the first target EEG signal and the second target EEG signal; Acquiring multimodal data, and extracting environmental perception features from the multimodal data to obtain environmental perception parameters; the multimodal data includes: image data, radar data, ultrasonic data, and driving data; Inputting the driving intention instruction and the environmental perception parameter into a multimodal large model for training to obtain a semantic feature instruction, and inputting the semantic feature instruction into an intelligent agent decision model to obtain a new energy vehicle control instruction; The new energy vehicle control command is verified for correctness to obtain an evaluation result, and the new energy vehicle is controlled according to the evaluation result.

7. The method for controlling an intelligent vehicle based on artificial intelligence according to claim 6, characterized in that: Inputting the semantic feature instructions into the agent decision model to obtain the new energy vehicle control instructions includes: Acquire the real-time state of the new energy vehicle, and use the real-time state and the semantic feature instruction as input variables of a fuzzy controller to obtain an output variable; the output variable is a steering wheel angle, and the real-time state includes: lateral deviation and deviation rate; generating an input fuzzy set based on the input variables and generating an output fuzzy set based on the output variables, and determining membership functions for the input fuzzy set and the output fuzzy set; wherein the membership functions are used to quantify the membership of the input variables and the output variables in the respective fuzzy sets; Determine fuzzy rules, perform fuzzy rule matching according to the membership of input variables to obtain multiple matching fuzzy rules, and calculate the output variables according to each matching fuzzy rule to obtain multiple fuzzy values; A synthesis operation is performed on each fuzzy value to obtain a target fuzzy output, and the target fuzzy output is defuzzified to obtain a new energy vehicle control instruction; the new energy vehicle control instruction is used to adjust the turning angle of the new energy vehicle.

8. The method for controlling an intelligent vehicle based on artificial intelligence according to claim 6, characterized in that: The correctness of the new energy vehicle control instructions is verified to obtain an evaluation result, including: Establishing a new energy vehicle state and constructing an event set, and determining a transfer function based on the new energy vehicle state and the event set; the new energy vehicle state includes: left turn degree, straight driving degree, and right turn degree; and the event set includes: controllable events and uncontrollable events; Determine the state of the new energy vehicle and define constraints according to the new energy vehicle control instructions, calculate an evaluation result according to the constraints and the controllable events, and determine the control logic according to the evaluation result; the evaluation results include: correct, wrong and serious error.

9. The method for controlling an intelligent vehicle based on artificial intelligence according to claim 8, characterized in that: Determining a control logic according to the evaluation result includes: If the evaluation result is greater than or equal to the first threshold, the evaluation result is determined to be correct, and the driving intention instruction is executed; If the first threshold value > the evaluation result ≥ the second threshold value, the evaluation result is determined to be wrong, and the weight of the automatic decision is reduced; If the second threshold value is greater than the evaluation result, the evaluation result is determined to be a serious error, and an instruction confirmation is performed through the feedback interface.

10. The intelligent vehicle control method based on artificial intelligence according to claim 6, characterized in that: After controlling the new energy vehicle according to the evaluation result, the method further includes: Periodically acquiring local operating data, encrypting the local operating data to obtain an encrypted package, and uploading the encrypted package to the cloud; Periodically receive cloud update packages to update large multimodal models.

Images