An electric vehicle chassis pre-control method and system based on artificial intelligence

Abstract

The application provides an electric vehicle chassis pre-control method and system based on artificial intelligence, and relates to the technical field of artificial intelligence, wherein the method comprises the following steps: based on artificial intelligence, determining a pre-control scheme for making the chassis performance achieve global optimal response according to the future driving road conditions of the electric vehicle; and executing the pre-control scheme. The application utilizes artificial intelligence technology to prospectively analyze future road condition information of the electric vehicle about to enter, and calculates a control strategy capable of making the chassis system achieve global performance optimization on the whole predicted journey or key road section, and executes the scheme in advance, so as to overcome the delay and local optimization limitation of traditional reactive control, and improve the performance potential of the chassis.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based method and system for pre-control of electric vehicle chassis. Background Technology

[0002] Currently, existing electric vehicle chassis control mainly relies on real-time sensors for reactive adjustment, which has significant lag and can only achieve local instantaneous optimization. It cannot make overall planning and proactive adjustments to the overall performance (comprehensive comfort, handling, energy consumption, and safety) based on continuous and complex road conditions (such as undulations, curves, and combinations of bumps) over medium to long distances (such as hundreds of meters), thus limiting the realization of the chassis performance potential. Summary of the Invention

[0003] One of the objectives of this invention is to provide an artificial intelligence-based pre-control method and system for electric vehicle chassis to solve the problems mentioned in the background art.

[0004] In a first aspect, embodiments of the present invention provide an artificial intelligence-based pre-control method for electric vehicle chassis, comprising:

[0005] Based on artificial intelligence, a pre-control scheme is determined to achieve the optimal global response of the chassis performance according to the future driving conditions of the electric vehicle.

[0006] Implement the aforementioned pre-control scheme.

[0007] Optionally, the pre-control scheme based on artificial intelligence, determining the chassis performance to achieve globally optimal response according to the future driving conditions of the electric vehicle, includes:

[0008] Assign response weights to different performance objectives;

[0009] By inputting different performance targets after assigning response weights and the future driving conditions of the electric vehicle into the pre-control decision AI model, a pre-control scheme that enables the chassis performance to achieve the global optimal response is obtained.

[0010] Optionally, the performance targets include:

[0011] Objective 1: To maintain occupant comfort levels consistently within a comfort range; whereby the comfort range is adapted to and matched with the comfort sensitivity predicted based on occupant profiles.

[0012] Objective 2: To maintain the chassis's multiple key stability parameters within their respective safety threshold ranges.

[0013] Objective 3: While simultaneously satisfying Objective 1 and Objective 2, maintain optimal energy consumption of the chassis.

[0014] Optionally, the step of obtaining the occupant's profile includes:

[0015] The vehicle's infotainment system obtains a map of the electric vehicle's driving environment and basic passenger information for a previously preset time period.

[0016] The pressure change distribution over a previously preset time period is obtained through a pressure sensor array under the occupant seat;

[0017] Based on the driving environment map and pressure change distribution, the degree of occupants' attention and immersion in the scenery outside the window was determined.

[0018] The passenger's level of focus and immersion in the scenery outside the window, along with basic passenger information, is used to create a passenger profile.

[0019] Optionally, determining the occupant's level of attention and immersion in the external environment based on the driving environment map and pressure change distribution includes:

[0020] For each historical time window, the change in the visible environment range outside the window corresponding to the occupant's seat is determined on the driving environment map within that historical time window. The various possible standard local pressure change distributions that the pressure sensor array should receive when the occupant focuses on different targets in the change of the visible environment range outside the window within that historical time window are analyzed. Each standard local pressure change distribution is matched with the pressure change distribution within the local pressure change distribution of that historical time window. The maximum matching degree is taken as the occupant's local immersion in the external environment outside the window within that historical time window.

[0021] Assign computational weights to different historical time windows;

[0022] Based on the assigned calculation weights, the local attention immersion of different historical time windows is weighted and calculated to obtain the occupant's attention immersion to the environment outside the car window.

[0023] Optionally, the step of predicting the comfort sensitivity includes:

[0024] Construct feature vectors for occupant profiles;

[0025] Based on the feature vector, the corresponding comfort sensitivity is matched from the comfort sensitivity table.

[0026] Optionally, the step of obtaining the comfort range includes:

[0027] Determine the comfort range corresponding to the sensitivity to comfort from the comfort range table.

[0028] Optionally, executing the pre-control scheme includes:

[0029] The pre-control scheme is then sent to the chassis control center platform of the electric vehicle.

[0030] Optionally, the artificial intelligence-based pre-control method for electric vehicle chassis also includes:

[0031] The pre-control decision AI model is periodically improved and trained.

[0032] Secondly, an embodiment of the present invention provides an artificial intelligence-based electric vehicle chassis pre-control system, comprising:

[0033] The artificial intelligence module is used to determine a pre-control scheme that enables the chassis performance to achieve the global optimal response based on artificial intelligence and the future driving conditions of the electric vehicle.

[0034] The chassis pre-control module is used to execute the pre-control scheme. This invention achieves the following beneficial effects:

[0035] By leveraging artificial intelligence technology, we can proactively analyze future road conditions that electric vehicles will soon be able to navigate, and calculate control strategies that enable the chassis system to achieve optimal global performance throughout the entire predicted journey or key road segments. We can then execute these strategies in advance to overcome the delays and limitations of local optimization in traditional reactive control, and enhance the full potential of chassis performance.

[0036] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0037] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0038] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0039] Figure 1 This is a schematic diagram of an artificial intelligence-based pre-control method for electric vehicle chassis in an embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram of an artificial intelligence-based electric vehicle chassis pre-control system according to an embodiment of the present invention. Detailed Implementation

[0041] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0042] The research and development approach of this application is to optimize the control strategy of electric vehicle chassis by combining artificial intelligence technology, so as to achieve more precise pre-control and response.

[0043] Figure 1 A flowchart of an artificial intelligence-based pre-control method for electric vehicle chassis is provided in this application embodiment, as shown below. Figure 1 As shown, the method includes:

[0044] 101. Based on artificial intelligence, determine a pre-control scheme that enables the chassis performance to achieve the optimal global response according to the future driving conditions of the electric vehicle.

[0045] In this step, future road condition characteristics are transformed into global optimization control commands for the chassis system through multi-objective weight allocation and AI decision-making models. Specifically, this includes the following sub-steps:

[0046] 201. Assign response weights to different performance objectives. The performance objectives include:

[0047] Objective 1: To maintain occupant comfort levels within a defined comfort range, where the comfort range is matched to the comfort sensitivity predicted based on occupant profiles.

[0048] In this objective, the sensitivity to comfort is predicted based on the occupant profile (such as age, physical characteristics, and historical motion sickness data), and the tolerance threshold range of parameters such as vibration frequency and acceleration change rate is dynamically set to form a comfort range.

[0049] Here is a specific implementation example: the vertical vibration frequency threshold range for elderly occupants (image tag: osteoporosis) is set to 2-5Hz (the conventional threshold is 1-8Hz) to avoid low-frequency resonance causing skeletal discomfort; the upper limit of the rate of change of acceleration is set to 0.3g / s (the conventional value is 0.5g / s).

[0050] Objective 2: To maintain the chassis’s multiple key stability parameters within their respective safety threshold ranges.

[0051] The key stability parameters for this objective include yaw rate deviation, lateral acceleration, tire slip ratio, and suspension travel margin. Different key stability parameters have pre-set safety threshold ranges, such as ±3° / s for yaw rate deviation and ±0.4g for lateral acceleration.

[0052] Here is a specific implementation example: if an electric vehicle is driving on an icy or snowy road, the safety threshold range for lateral acceleration on icy or snowy roads is tightened to ±0.2g.

[0053] Objective 3: While simultaneously satisfying Objective 1 and Objective 2, maintain optimal energy consumption of the chassis.

[0054] In this objective, the energy consumption of the chassis can be counted as the integral of the power battery output power. The optimal solution is to minimize this integral.

[0055] The above objectives one, two, and three represent performance targets for comfort, stability, and energy consumption, respectively. When assigning response weights to different performance targets, these weights can be dynamically allocated as needed based on the vehicle's usage scenarios. A specific implementation example is provided below:

[0056] In high-speed cruising scenarios, the stability weight is set to 0.6, the comfort weight to 0.3, and the energy consumption weight to 0.1; in urban congestion scenarios, the comfort weight increases to 0.5, the energy consumption weight is set to 0.4, and the stability weight is set to 0.1.

[0057] In addition, the weights can be constrained by the following condition: the sum of all weights is 1, and each weight is greater than or equal to 0.1 and less than or equal to 0.8, to avoid single-objective dominance.

[0058] 202. Input the different performance targets after assigning response weights and the future driving conditions of the electric vehicle into the pre-control decision AI model to obtain a pre-control scheme that enables the chassis performance to achieve the global optimal response.

[0059] In this step, pre-control schemes that enable the chassis to achieve the globally optimal response are determined in advance through experiments for different performance targets after assigning response weights and various possible future driving conditions of the electric vehicle. These schemes are used as training samples to train the artificial intelligence model, enabling it to determine the pre-control schemes that enable the chassis to achieve the globally optimal response based on different performance targets after assigning response weights and the future driving conditions of the electric vehicle.

[0060] Here's a concrete implementation example: First, in a chassis performance testing platform or high-precision simulation environment, conduct numerous experiments targeting various driving conditions that electric vehicles may encounter in the future, along with different pre-set combinations of performance target response weights. Under each experimental condition, based on expert experience, determine and record specific pre-control schemes (such as active suspension actuation commands, torque vector distribution strategies, braking pre-compression, etc.) that achieve the globally optimal chassis performance response, forming a training sample dataset. Use data management software to associate, label, and store the driving condition features, performance target weight vectors, and corresponding optimal pre-control schemes in the samples.

[0061] Next, the training environment for the pre-control decision-making AI model was set up. Python version 3.11 was used, and PyTorch version 2.3.1 was selected as the deep learning framework, with CUDA 12.4 acceleration library configured. After configuring the necessary scientific computing libraries (such as NumPy, SciPy, etc.) and dependencies, a neural network architecture suitable for sequence decision-making and multi-objective optimization (such as an attention-based encoder-decoder structure or a deep reinforcement learning network DQN / DDPG variant) was selected as the base model. The prepared dataset was divided into training, validation, and test sets in a 7:1.5:1.5 ratio. The model was iteratively trained using the training set, and the model performance was monitored using the validation set (such as the matching degree between the predicted control scheme and the global optimal scheme, and the value of the multi-objective comprehensive loss function) and hyperparameters were tuned (such as learning rate, batch size, and number of network layers). Finally, the parameter file of the best-performing model on the validation set (such as best_model.pth) was saved.

[0062] Finally, in the online application phase, the specific performance target vector after real-time allocation of response weights, along with the predicted or perceived sequence of future driving conditions for the electric vehicle, are input into the trained pre-control decision AI model. The model performs computational inference based on the learned mapping relationships and outputs a sequence of pre-control instructions that achieves the globally optimal response of the chassis performance under given conditions.

[0063] 102. Execute the aforementioned pre-control scheme.

[0064] In this step, after the pre-control scheme is determined, the pre-control scheme is executed on the electric vehicle chassis.

[0065] In summary, through steps 101 to 102, artificial intelligence technology is used to proactively analyze future road conditions that the electric vehicle will soon enter, and based on this, a control strategy (pre-control scheme) that enables the chassis system (suspension, steering, braking) to achieve optimal global performance (comprehensively considering smoothness, handling stability, energy consumption, safety, etc.) throughout the entire predicted journey or key road segments is calculated. This scheme is then executed in advance to overcome the delay and local optimization limitations of traditional reactive control and to enhance the performance potential of the chassis.

[0066] The first objective (dynamically adapting comfort ranges based on occupant profiles) significantly enhances the personalized riding experience, avoiding over-constraint or insufficient protection caused by uniform thresholds. The second objective (dynamic safety thresholds for key stability parameters) proactively defends against instability risks such as rollovers and fishtailing in complex road conditions, strengthening the chassis safety boundary. The third objective (optimal continuous energy consumption under multi-objective constraints) maximizes energy utilization efficiency while ensuring both comfort and safety.

[0067] In some embodiments, the step of obtaining the occupant's image includes:

[0068] 301. Obtain the driving environment map and basic passenger information of the electric vehicle within a preset time period through the vehicle's infotainment system.

[0069] In this step, the previously preset time can be set to 10 minutes; the driving environment map is a map that includes the environmental conditions along the historical driving route of the electric vehicle and the changes in the vehicle's position; the basic information of the occupants includes: age, gender, medical history tags, and the number of times they have experienced motion sickness.

[0070] 302. Obtain the pressure change distribution within a previously preset time period through the pressure sensor array under the passenger seat.

[0071] In this step, a 64-unit piezoresistive sensor grid (spatial resolution 2cm×2cm) can be built into the seat cushion to capture pressure spatiotemporal distribution data at a sampling frequency of 100Hz. The historical sampling results corresponding to this data within a previously preset time period are used as the pressure change distribution.

[0072] 303. Based on the driving environment map and pressure change distribution, determine the occupants' level of attention and immersion in the scenery outside the car window.

[0073] In this step, when occupants focus on the scenery outside the window, their posture typically changes (e.g., if they see an object outside the window but it will soon be behind the vehicle as it moves, they will slightly turn their head and lean to the side to continue viewing the object). The pressure distribution received by the pressure sensor array will also change. Therefore, based on the driving environment map and the pressure change distribution, the occupant's level of immersion in the scenery outside the window can be determined. The level of immersion represents the degree to which the occupant is fully engaged in focusing on the scenery outside the window.

[0074] Here's a concrete implementation example: Based on the pressure change distribution, with the pelvic position (the pressure point the pelvis contacts when the occupant is seated) as the origin, record the displacement sequence of the pressure center point in the XY plane to obtain the pressure center offset trajectory. Then, use the Shannon entropy formula to calculate the entropy value of the pressure change distribution. The entropy value represents the disorder of the pressure change distribution; the higher the entropy value, the more frequently the occupant adjusts their posture. Taking the vehicle approaching a scenic spot as an example, the pressure center offset trajectory reflects that the pressure center is continuously biased towards the window side (angle > 15°), and the entropy value is below 0.3, indicating a relatively high immersion level of 0.75. Conversely, if the pressure distribution is disordered (entropy value > 0.6) or the center of gravity is offset in the opposite direction, the immersion level is determined to be relatively low at 0.3.

[0075] 304. The degree of occupant's focus on and immersion in the scenery outside the car window, along with the occupant's basic information, is used to create an occupant profile.

[0076] In summary, steps 301 to 304, by combining the occupant's level of immersion in the scenery outside the window with basic occupant information to create an occupant profile, allow for a focus on this specific factor when predicting comfort sensitivity. This provides occupants with an experience that allows them to comfortably focus on the scenery outside the window.

[0077] In some embodiments, step 303, determining the occupant's level of attention and immersion in the external environment based on a driving environment map and pressure change distribution, includes:

[0078] 401. For each historical time window, determine the change in the visible environment range outside the window corresponding to the occupant's seat on the driving environment map within that historical time window. Analyze the various possible standard local pressure change distributions that the pressure sensor array should receive when the occupant focuses on different targets in the change of the visible environment range outside the window within that historical time window. Match each standard local pressure change distribution with the pressure change distribution within the local pressure change distribution of that historical time window, and take the maximum matching degree as the occupant's local immersion in the external environment outside the window within that historical time window.

[0079] In this step, a sliding window mechanism is used to divide the preset time, with a window length of 5 seconds and a step size of 1 second, forming overlapping time segments, thus creating multiple historical time windows. Changes in the visible environment outside the window refer to the changes in the visible range of the environment viewed by the occupant from their seat through the window when the vehicle's posture changes accordingly within the historical time window. Beforehand, experiments are conducted to determine the pressure change distribution that the pressure sensor array should receive when the occupant focuses on different targets within the changing visible environment outside the window, under various possible changes in the visible environment outside the window and the vehicle's posture. This distribution is then used as the standard local pressure change distribution. Each standard local pressure change distribution is matched with the local pressure change distribution within the historical time window. The maximum matching degree reflects the actual situation of the occupant focusing on different targets, and is thus taken as the occupant's local immersion in the external environment outside the window within that historical time window.

[0080] Here is a specific implementation example: Based on the vehicle's pose (GPS+IMU data) and the occupant's seat coordinates (relative to the vehicle center), a dynamic visual cone is calculated with the occupant's eye point as the origin. This dynamic visual cone has a horizontal viewing angle of 60° and a vertical viewing angle of 30° (compliant with SAE J941 standards). Combined with high-precision map terrain elevation and building occlusion data (such as LiDAR point clouds), a real-time visible area raster map (resolution 0.5°×0.5°) is generated, which serves as the range of changes in the visible environment outside the window. Continuing with the above specific implementation example, the pressure change distribution is analyzed to determine the pressure center of gravity offset trajectory. If a target is located outside the occupant's left window, and the occupant is focused on it, the corresponding standard local pressure change distribution is a negative pressure center of gravity offset of 4-8cm and a 30%-50% increase in pressure on the left hip. The pressure center of gravity offset trajectory determined within this time window is used as the local pressure change distribution. When matching the two, the degree to which the characteristics reflected by the pressure center of gravity offset trajectory conform to the standard local pressure change distribution is determined, which is used as the matching degree and as the local focus immersion degree.

[0081] 402. Assign calculation weights to different historical time windows.

[0082] In this step, calculation weights can be assigned to different historical time windows as needed. For example, if the significance of targets within the range of changes in the visible environment outside the window is high within a certain historical time window, then a higher weight is assigned to the corresponding historical time window. Or, if the signal-to-noise ratio of the data distribution of local pressure changes within a certain historical time window is <15dB, the weight is forced to zero to avoid including low-confidence data in the calculation.

[0083] 403. Based on the assigned calculation weights, the local attention immersion of different historical time windows is weighted and calculated to obtain the occupant's attention immersion to the environment outside the window.

[0084] In this step, finally, based on the assigned calculation weights, the local attention immersion of different historical time windows is weighted and calculated to obtain the attention immersion.

[0085] In summary, steps 401 to 403 accurately analyze the changes in the visible environment outside the window within each historical time window, determine multiple possible standard local pressure change distributions, and match them with the local pressure change distributions within the corresponding historical time window. The maximum match is taken as the occupant's local attention immersion in the external environment of the vehicle window in the corresponding historical time window. Then, calculation weights are assigned to different historical time windows. Based on the assigned calculation weights, the local attention immersion in different historical time windows is calculated in a weighted manner to finally obtain the attention immersion. This greatly improves the accuracy, comprehensiveness, and applicability of the attention immersion analysis, and enhances the accuracy of predicting comfort sensitivity by incorporating it into the occupant profile, thus taking into account the special factor of the occupant's attention immersion in the scenery outside the vehicle window.

[0086] In some embodiments, the step of predicting the comfort sensitivity includes:

[0087] Construct feature vectors for occupant profiles;

[0088] Based on the feature vector, the corresponding comfort sensitivity is matched from the comfort sensitivity table.

[0089] The comfort sensitivity corresponding to the feature vectors of different occupant profiles can be pre-set as needed to form a comfort sensitivity table. When predicting comfort sensitivity, the feature vectors of the occupant profiles can be directly obtained by looking up the table.

[0090] In some embodiments, the step of obtaining the comfort range includes:

[0091] Determine the comfort range corresponding to the sensitivity to comfort from the comfort range table.

[0092] Similarly, comfort ranges corresponding to different levels of comfort sensitivity can be pre-set as needed to form a comfort range table.

[0093] In some embodiments, executing the pre-control scheme includes:

[0094] The pre-control scheme is then sent to the chassis control center platform of the electric vehicle.

[0095] In actual implementation, the pre-control commands in the pre-control scheme can be sent to the domain controllers (chassis domain controller and powertrain domain controller) via the CAN-FD bus.

[0096] In some embodiments, the artificial intelligence-based pre-control method for electric vehicle chassis further includes:

[0097] The pre-control decision AI model is periodically improved and trained.

[0098] In addition, the pre-control decision AI model can be improved and trained periodically to enhance its performance.

[0099] Figure 2 This application provides a schematic diagram of an artificial intelligence-based electric vehicle chassis pre-control system, as shown in the embodiment. Figure 2 As shown, the system includes:

[0100] Artificial intelligence module 100 is used to determine a pre-control scheme that enables the chassis performance to achieve the global optimal response based on artificial intelligence and the future driving conditions of the electric vehicle.

[0101] The chassis pre-control module 200 is used to execute the pre-control scheme.

[0102] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A pre-control method for electric vehicle chassis based on artificial intelligence, characterized in that, include: Based on artificial intelligence, a pre-control scheme is determined to achieve the optimal global response of the chassis performance according to the future driving conditions of the electric vehicle. Execute the aforementioned pre-control scheme; The pre-control scheme, based on artificial intelligence and determined according to the future driving conditions of the electric vehicle, to achieve the globally optimal response of the chassis performance, includes: Assign response weights to different performance objectives; The different performance targets after assigning response weights and the future driving conditions of the electric vehicle are input into the pre-control decision AI model to obtain a pre-control scheme that enables the chassis performance to achieve the global optimal response. The performance targets include: Objective 1: To maintain occupant comfort levels consistently within a comfort range; whereby the comfort range is adapted to and matched with the comfort sensitivity predicted based on occupant profiles. Objective 2: To maintain the chassis's multiple key stability parameters within their respective safety threshold ranges. Objective 3: While simultaneously satisfying Objective 1 and Objective 2, maintain optimal energy consumption of the chassis. The steps for obtaining the occupant's image include: The vehicle's infotainment system obtains a map of the electric vehicle's driving environment and basic passenger information for a previously preset time period. The pressure change distribution over a previously preset time period is obtained through a pressure sensor array under the occupant seat; Based on the driving environment map and pressure change distribution, the degree of occupants' attention and immersion in the scenery outside the window was determined. The passenger's level of focus and immersion in the scenery outside the window, along with the passenger's basic information, is used to create a passenger profile; The determination of occupants' level of attention and immersion in the external environment based on driving environment maps and pressure change distribution includes: For each historical time window, the change in the visible environment range outside the window corresponding to the occupant's seat is determined on the driving environment map within that historical time window. The various possible standard local pressure change distributions that the pressure sensor array should receive when the occupant focuses on different targets in the change of the visible environment range outside the window within that historical time window are analyzed. Each standard local pressure change distribution is matched with the pressure change distribution within the local pressure change distribution of that historical time window. The maximum matching degree is taken as the occupant's local immersion in the external environment outside the window within that historical time window. Assign computational weights to different historical time windows; Based on the assigned calculation weights, the local attention immersion of different historical time windows is weighted and calculated to obtain the occupant's attention immersion to the environment outside the car window.

2. The artificial intelligence-based pre-control method for electric vehicle chassis as described in claim 1, characterized in that, The steps for predicting comfort sensitivity include: Construct feature vectors for occupant profiles; Based on the feature vector, the corresponding comfort sensitivity is matched from the comfort sensitivity table.

3. The artificial intelligence-based pre-control method for electric vehicle chassis as described in claim 1, characterized in that, The steps for obtaining the comfort range include: Determine the comfort range corresponding to the sensitivity to comfort from the comfort range table.

4. The artificial intelligence-based pre-control method for electric vehicle chassis as described in claim 1, characterized in that, The execution of the pre-control scheme includes: The pre-control scheme is then sent to the chassis control center platform of the electric vehicle.

5. The artificial intelligence-based pre-control method for electric vehicle chassis as described in claim 1, characterized in that, Also includes: The pre-control decision AI model is periodically improved and trained.

6. An artificial intelligence-based pre-control system for electric vehicle chassis, characterized in that, include: The artificial intelligence module is used to determine a pre-control scheme that enables the chassis performance to achieve the global optimal response based on artificial intelligence and the future driving conditions of the electric vehicle. The chassis pre-control module is used to execute the pre-control scheme; The pre-control scheme, based on artificial intelligence and determined according to the future driving conditions of the electric vehicle, to achieve the globally optimal response of the chassis performance, includes: Assign response weights to different performance objectives; The different performance targets after assigning response weights and the future driving conditions of the electric vehicle are input into the pre-control decision AI model to obtain a pre-control scheme that enables the chassis performance to achieve the global optimal response. The performance targets include: Objective 1: To maintain occupant comfort levels consistently within a comfort range; whereby the comfort range is adapted to and matched with the comfort sensitivity predicted based on occupant profiles. Objective 2: To maintain the chassis's multiple key stability parameters within their respective safety threshold ranges. Objective 3: While simultaneously satisfying Objective 1 and Objective 2, maintain optimal energy consumption of the chassis. The steps for obtaining the occupant's image include: The vehicle's infotainment system obtains a map of the electric vehicle's driving environment and basic passenger information for a previously preset time period. The pressure change distribution over a previously preset time period is obtained through a pressure sensor array under the occupant seat; Based on the driving environment map and pressure change distribution, the degree of occupants' attention and immersion in the scenery outside the window was determined. The passenger's level of focus and immersion in the scenery outside the window, along with the passenger's basic information, is used to create a passenger profile; The determination of occupants' level of attention and immersion in the external environment based on driving environment maps and pressure change distribution includes: For each historical time window, the change in the visible environment range outside the window corresponding to the occupant's seat is determined on the driving environment map within that historical time window. The various possible standard local pressure change distributions that the pressure sensor array should receive when the occupant focuses on different targets in the change of the visible environment range outside the window within that historical time window are analyzed. Each standard local pressure change distribution is matched with the pressure change distribution within the local pressure change distribution of that historical time window. The maximum matching degree is taken as the occupant's local immersion in the external environment outside the window within that historical time window. Assign computational weights to different historical time windows; Based on the assigned calculation weights, the local attention immersion of different historical time windows is weighted and calculated to obtain the occupant's attention immersion to the environment outside the car window.

Images