Cloud elastic computing resource scheduling method and system for emotion cockpit

Abstract

According to the emotion cabin-oriented cloud elastic computing resource scheduling method and system provided by the invention, by dynamically adjusting cloud computing resources, the real-time performance and accuracy of emotion recognition and in-vehicle environment adjustment are ensured according to the real-time emotion change and environment adjustment requirements of a vehicle owner; by introducing task priority scheduling, load balancing and automatic expansion and contraction mechanisms, the system can efficiently utilize computing resources, resource waste is avoided, and meanwhile the operation and maintenance cost is reduced; besides, all data processing adopts a localized encryption technology, so that the privacy and data security of the vehicle owner are ensured, and modern privacy protection laws and regulations are met; in addition, the system can provide personalized computing resource scheduling through continuous learning and intelligent optimization, and the personalized experience and comfort level of the user are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of resource scheduling, and in particular to a cloud-based elastic computing resource scheduling method and system for an emotional cockpit. Background Art

[0002] In the areas of smart cockpits and in-vehicle emotion recognition, existing cloud computing resource scheduling methods generally suffer from insufficient real-time responsiveness, inefficient resource allocation, and poor adaptability to sudden mood swings. Existing systems often rely heavily on historical operation logs for resource forecasting and lack dynamic response mechanisms to sudden demands, such as mood swings, leading to resource waste or delays. Furthermore, existing methods also lack load balancing and task priority management, making it difficult to flexibly adjust resources based on the driver's real-time mood and environmental needs.

[0003] Patent application CN119718682A discloses a server resource scheduling system that combines AI and edge computing. This system optimizes resource usage by real-time monitoring of server nodes, predicting task demands, and dynamically adjusting resource allocation. The system includes multiple modules, including computing nodes, task traversal, and resource allocation, and utilizes operation logs for demand forecasting and task scheduling. However, this patent relies heavily on operation logs and monitoring data, and data anomalies can affect scheduling decisions. The system also suffers from a complex structure and high implementation and maintenance costs. Response to sudden loads is limited, and the scheduling algorithm is static, lacking the ability to dynamically adjust priorities.

[0004] Patent application document CN119668883A discloses a computing resource scheduling method based on the Internet of Vehicles. By real-time monitoring of computing node status, analyzing task resource requirements and response time trends, it dynamically selects computing terminals for task offloading to improve computing efficiency. However, this patent judges the adaptability of computing tasks based on time change trends, but may fail to provide sufficient emergency response mechanisms when faced with emergencies or large-scale data changes. Dynamic computing resource allocation may be interfered with by sudden loads, resulting in delayed system response.

[0005] Patent application CN119652754A utilizes AI models to dynamically schedule and optimize network resources, including traffic forecasting, anomaly detection, route optimization, and resource allocation. This approach uses edge computing to improve the computing efficiency and stability of in-vehicle networks. While this approach proposes a dynamic allocation strategy for computing resources, ensuring timely and efficient scheduling of network resources to cope with the heavy workload under extreme conditions remains a challenge requiring further optimization.

[0006] Patent application document CN119829293A proposes a heterogeneous resource scheduling method for deep learning tasks that combines task bundling, duration prediction, and resource scheduling. Scheduling optimization is achieved through AI models to improve the execution efficiency of computing tasks. However, the current resource scheduling scheme in this patent relies primarily on the prediction and scheduling of existing resources and may not be able to fully adapt to the dynamic changes in resource allocation between devices. In particular, when computing resources are relatively tight or computing tasks are relatively complex, there may be a shortage of resources. Summary of the Invention

[0007] In view of the defects in the prior art, the purpose of the present invention is to provide a cloud-based elastic computing resource scheduling method and system for emotional cockpits.

[0008] The cloud-based elastic computing resource scheduling method for an emotional cockpit provided by the present invention includes:

[0009] Step S1: Cloud elastic computing resource scheduling demand analysis;

[0010] Step S1.1: Based on the actual needs of the emotional cockpit system, calculate the data volume of each task per unit time and, combined with the computational complexity of the algorithm model, predict the overall computing load per unit time. Establish a resource demand model by task type, clarifying the configuration standards and elastic expansion limits for CPU, GPU, memory, and bandwidth.

[0011] Step S1.2: Analyze the data transmission bandwidth and latency requirements between the vehicle system and the cloud to ensure low latency and high bandwidth requirements for high real-time tasks;

[0012] Step S2: elastic computing resource model design;

[0013] Step S2.1: Dynamically adjust cloud resources based on the resource demand prediction function;

[0014] Step S2.2: Allocate resources based on resource scheduling algorithm;

[0015] Step S2.3: Build an automatic expansion and contraction mechanism to add cloud nodes when the predicted resource demand exceeds the expansion threshold and recycle idle nodes when it falls below the contraction threshold;

[0016] Step S3: resource scheduling algorithm implementation and optimization;

[0017] Step S3.1: Scheduling resources based on task priorities;

[0018] Step S3.2: Dynamically adjust resource allocation by combining real-time emotion data and environmental data, and implement resource allocation through a load balancing algorithm;

[0019] Step S4: real-time monitoring and dynamic adjustment;

[0020] Step S4.1: Monitor computing resource usage in real time and trigger resource allocation adjustments when the load approaches a preset threshold;

[0021] Step S4.2: Dynamically adjust resources based on mood fluctuations and environmental changes, reducing resource consumption when emotions are stable and increasing resources when emotions fluctuate violently;

[0022] Step S5: data privacy and security protection;

[0023] Step S5.1: The driver's emotional data, facial expressions, and voice information are locally encrypted and transmitted to the cloud;

[0024] Step S5.2: Process sensitive information using data desensitization technology and comply with privacy protection regulations;

[0025] Step S6: Intelligent and continuous optimization of cloud resource scheduling;

[0026] Step S6.1: Utilize AI algorithms to analyze vehicle owner behavior data and environmental adjustment history to optimize resource scheduling strategies;

[0027] Step S6.2: Learn the emotion change pattern and computing requirements in real time, and optimize the scheduling algorithm to improve the accuracy of resource allocation.

[0028] Preferably, the resource demand prediction function is:

[0029]

[0030] Among them, R(t) is the total resource demand at time t, D i (t) represents the data processing amount or computing intensity of the i-th task at time t, w i Represents the weight coefficient of the corresponding task, and n is the total number of tasks.

[0031] Preferably, the resource scheduling algorithm is:

[0032]

[0033] Among them, A(t) is the comprehensive score of the resources required for allocation at time t; α, β, and γ are weight coefficients that control the influence of priority, complexity, and real-time performance in the allocation decision; P i 、C i 、R i Represents the priority, complexity, and real-time level value of task i.

[0034] Preferably, the task priority assessment formula is:

[0035] Priority i =λ·U i+(1-λ)·C i

[0036] Among them, Priority i Score the priority of task i, U i is the task urgency, normalized to 0-1; λ is the weight parameter, which controls the relative importance of urgency and complexity.

[0037] Preferably, the load balancing algorithm adopts a weighted minimum connection number algorithm, which is expressed as:

[0038]

[0039] Among them, WLC i is the weighted connection index of node i, N i is the current number of connections of node i, W i is the performance weight of node i.

[0040] The cloud-based elastic computing resource scheduling system for the emotional cockpit provided by the present invention includes:

[0041] Module M1: Cloud elastic computing resource scheduling demand analysis;

[0042] Module M1.1: Based on the actual needs of the emotional cockpit system, the data volume of various tasks per unit time is counted. Combined with the computational complexity of the algorithm model, the overall computing load per unit time is predicted. A resource demand model is established by task type, clarifying the configuration standards and elastic expansion limits of CPU, GPU, memory, and bandwidth.

[0043] Module M1.2: Analyzes the data transmission bandwidth and latency requirements between the vehicle system and the cloud to ensure low latency and high bandwidth requirements for high real-time tasks;

[0044] Module M2: Design of elastic computing resource model;

[0045] Module M2.1: Dynamically adjust cloud resources based on resource demand prediction function;

[0046] Module M2.2: Allocate resources based on resource scheduling algorithms;

[0047] Module M2.3: Build an automatic expansion and contraction mechanism to add cloud nodes when the predicted resource demand exceeds the expansion threshold and recycle idle nodes when it falls below the contraction threshold;

[0048] Module M3: Resource scheduling algorithm implementation and optimization;

[0049] Module M3.1: Scheduling resources based on task priority;

[0050] Module M3.2: Dynamically adjust resource allocation by combining real-time sentiment data and environmental data, and implement resource allocation through load balancing algorithms;

[0051] Module M4: real-time monitoring and dynamic adjustment;

[0052] Module M4.1: monitors computing resource usage in real time and triggers resource allocation adjustments when the load approaches a preset threshold;

[0053] Module M4.2: Dynamically adjust resources based on mood swings and environmental changes, reducing resource consumption when emotions are stable and increasing resources when emotions fluctuate violently;

[0054] Module M5: Data Privacy and Security;

[0055] Module M5.1: Locally encrypts the driver's emotional data, facial expressions, and voice information and transmits them to the cloud;

[0056] Module M5.2: Process sensitive information through data desensitization technology and comply with privacy protection regulations;

[0057] Module M6: Intelligent and continuous optimization of cloud resource scheduling;

[0058] Module M6.1: Use AI algorithms to analyze driver behavior data and environmental adjustment history to optimize resource scheduling strategies;

[0059] Module M6.2: Learn emotion change patterns and computing requirements in real time, and optimize scheduling algorithms to improve resource allocation accuracy.

[0060] Preferably, the resource demand prediction function is:

[0061]

[0062] Among them, R(t) is the total resource demand at time t, D i (t) represents the data processing amount or computing intensity of the i-th task at time t, w i Represents the weight coefficient of the corresponding task, and n is the total number of tasks.

[0063] Preferably, the resource scheduling algorithm is:

[0064]

[0065] Among them, A(t) is the comprehensive score of the resources required for allocation at time t; α, β, and γ are weight coefficients that control the influence of priority, complexity, and real-time performance in the allocation decision; P i 、C i 、R i Represents the priority, complexity, and real-time level value of task i.

[0066] Preferably, the task priority assessment formula is:

[0067] Priority i =λ·U i +(1-λ)·C i

[0068] Among them, Priority i Score the priority of task i, U i is the task urgency, normalized to 0-1; λ is the weight parameter, which controls the relative importance of urgency and complexity.

[0069] Preferably, the load balancing algorithm adopts a weighted minimum connection number algorithm, which is expressed as:

[0070]

[0071] Among them, WLC i is the weighted connection index of node i, N i is the current number of connections of node i, W i is the performance weight of node i.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] (1) The present invention dynamically adjusts cloud computing resources based on the driver's real-time emotional changes and environmental adjustment needs to ensure the real-time and accurate emotional recognition and in-vehicle environmental adjustment.

[0074] (2) By introducing task priority scheduling, load balancing, and automatic expansion and contraction mechanisms, the present invention enables the system to efficiently utilize computing resources, avoid resource waste, and reduce operation and maintenance costs;

[0075] (3) All data processing in the present invention uses local encryption technology to ensure the privacy and data security of the car owner and comply with modern privacy protection regulations;

[0076] (4) The present invention can provide personalized computing resource scheduling through continuous learning and intelligent optimization, thereby improving the user's personalized experience and comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0078] Figure 1 This is a flow chart of the cloud-based elastic computing resource scheduling method for the emotional cockpit of the present invention. DETAILED DESCRIPTION

[0079] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0080] Example

[0081] like Figure 1 The present invention provides a cloud-based elastic computing resource scheduling method for an emotional cockpit, comprising:

[0082] Step S1: Analyze the demand for cloud elastic computing resource scheduling.

[0083] Step S1.1: Based on the actual needs of the emotional cockpit system, analyze the computing resources required for system operation, including the computational complexity of tasks such as emotion recognition, environmental adjustment, and real-time data processing. Each task has different computing resource requirements at different times. In particular, computing resource requirements may increase dramatically during peak periods of emotion recognition and environmental adjustment.

[0084] The specific analysis process is as follows:

[0085] First, count the amount of data generated by each task per unit time (per second) (for example, the number of image frames and speech sampling rate required for emotion recognition, and the amount of sensor data required for environmental adjustment).

[0086] Secondly, the computing resources required for a single processing are estimated based on the computational complexity (floating-point operations FLOPs) of the algorithm model used (such as deep learning emotion recognition network, environmental control optimization algorithm).

[0087] Then, the overall computing load per unit time is predicted by combining the triggering frequency and real-time requirements of different tasks in different time periods.

[0088] Finally, a resource requirement model is established by task type to clarify the minimum configuration standards and elastic expansion limits of CPU, GPU, memory, and bandwidth, which serve as the basis for subsequent elastic scheduling.

[0089] Step S1.2: Analyze the data transmission bandwidth and latency requirements between the in-vehicle system and the cloud. For high-precision emotion recognition and environmental adjustment, the system needs to respond in real time, so low latency and high bandwidth requirements need to be considered when scheduling cloud resources.

[0090] Step S2: Design of elastic computing resource model.

[0091] Step S2.1: Based on the computing requirements of the emotion cockpit, design a flexible computing resource model that can dynamically adjust the configuration of cloud computing resources to ensure that the required computing resources can be quickly provided during peak hours of emotion recognition and environmental adjustment tasks. The specific process is as follows:

[0092] 1. Introduce resource demand prediction function: Among them, R(t) is the total resource demand at time t, D i (t) represents the data processing amount or computing intensity of the i-th task at time t, w i Indicates the weight coefficient of the corresponding task (determined according to task priority and real-time performance).

[0093] 2. The resource scheduling algorithm automatically and elastically expands or contracts based on the prediction results. Specifically:

[0094] Demand assessment: Collect the current load data of each task in real time and calculate the predicted resource demand R(t+Δt), which is the demand at the next moment;

[0095] Resource allocation decision: setting the expansion threshold T up and the contraction threshold T down , when R(t+Δt)>T up When the number of cloud nodes increases or the node specifications are improved, the opposite is true.

[0096] Resource allocation adjustment: Dynamically modify resource pool configuration through the scheduling controller, such as increasing the number of virtual machines and container instances, and adjusting the CPU cores and GPU computing power allocated to the emotion recognition / environment adjustment module.

[0097] For example, when it detects that the driver's emotions are fluctuating drastically, causing the predicted load of the emotion recognition module to increase by more than 30%, the system automatically triggers the expansion mechanism based on the calculation results of the resource demand prediction function, adding 2 computing nodes within 1 minute, and adjusting the load balancer's allocation strategy to ensure that the emotion recognition task gets priority in obtaining the newly added resources.

[0098] Step S2.2: Design resource scheduling algorithm;

[0099]

[0100] Among them, A(t) is the comprehensive score of the resources required for allocation at time t; α, β, and γ are weight coefficients that control the influence of priority, complexity, and real-time performance in the allocation decision; P i 、C i 、R i Represents the priority, complexity, and real-time level value of task i.

[0101] It automatically adjusts the allocation of cloud resources based on the current task's priority, computational complexity, and real-time requirements. For example, if a driver's emotional state fluctuates significantly, the system needs to quickly allocate more computing resources to adjust the environment.

[0102] Step S2.3: Build an automatic expansion and contraction mechanism for cloud computing resources. When demand for computing resources increases, the system can automatically expand cloud computing resources, and when demand decreases, it can automatically reclaim resources, thereby improving the system's resource utilization and cost-effectiveness.

[0103] Expansion mechanism: The system dynamically adds cloud computing instances / container nodes through the container orchestration platform (Kubernetes);

[0104] The load balancing system (such as the service gateway) is updated synchronously to incorporate the newly added resources into the scheduling pool;

[0105] High-priority tasks are automatically reallocated to new nodes, ensuring that high-load tasks receive priority access to expanded resources.

[0106] Shrinkage mechanism: detects the number and time of idle resource nodes;

[0107] If some nodes are idle for more than a set time threshold (e.g. no task scheduling for 5 minutes), node recycling is triggered;

[0108] Before recycling, ensure that no critical tasks are running to avoid interruption.

[0109] Step S3: Implementation and optimization of resource scheduling algorithm.

[0110] Step S3.1: Implement a computing resource scheduling algorithm based on task priority. Each task (e.g., emotion recognition, environmental adjustment, etc.) is assigned a different priority based on its urgency and computational complexity. High-priority tasks receive preferential computing resources, ensuring the system can respond as quickly as possible during peak periods of emotional fluctuations and environmental adjustment.

[0111] Priority assessment is based on urgency and computational complexity to produce a priority assessment score;

[0112] Priority i =λ·U i +(1-λ)·C i

[0113] Among them, U iThe urgency of the task (such as the driver's emotional fluctuation or the level of environmental mutation) is normalized to 0-1. λ is a weight parameter that controls the relative importance of urgency and complexity (for example, setting it to 0.7 means more emphasis on urgency). For example, if the emotion recognition module currently detects a sharp increase in the driver's emotional abnormality (such as a +30% fluctuation in the face recognition score), then its U≈1. Even if the computational complexity is high, it will still be prioritized due to its high urgency.

[0114] Step S3.2: Dynamically adjust the allocation of computing resources based on the driver's real-time emotional data, environmental data, and the task queue. For example, if the driver's emotions fluctuate dramatically, the system automatically increases computing resources to ensure the real-time and accurate emotion recognition. This real-time feedback mechanism allows the system to continuously optimize the allocation of computing resources.

[0115] When the driver's emotions fluctuate dramatically (e.g., significant changes in expression recognition scores or noticeable fluctuations in voice emotion), the system determines that the urgency of the emotion recognition and environment adjustment modules increases, and prioritizes more computing resources for these tasks to ensure real-time emotion recognition and timely response to environment adjustment.

[0116] When there are significant changes in in-vehicle environmental data (such as temperature, humidity, and light), the priority of the environmental adjustment task increases, and more resources are allocated accordingly;

[0117] At the same time, each task in the task queue is assigned a different priority based on urgency and complexity. The system will automatically sort the tasks according to their priority and allocate more resources to high-priority tasks.

[0118] In order to ensure the timeliness and accuracy of resource allocation, the system has designed a real-time feedback mechanism:

[0119] The system sets a fixed time interval (e.g., every 5 seconds) to detect the owner's emotional state, environmental changes, and task queue status;

[0120] After each test, the allocation ratio of computing resources is dynamically adjusted based on the latest data;

[0121] If it is detected that the resource utilization rate, task response time or load situation deviates from the expected value by more than a set threshold (such as 15%), the system will immediately readjust the resource allocation strategy;

[0122] Through this continuous real-time feedback and adjustment, resources are always prioritized to support critical tasks, improving the overall response speed and stability of the system.

[0123] Step S3.3: Optimize the load balancing of computing resources to ensure a reasonable distribution of load across computing nodes in the cloud, avoiding overloading of any node and resulting in system performance degradation. Use a distributed load balancing algorithm to achieve efficient resource allocation and parallel processing of tasks.

[0124] The Weighted Least Connection (WLC) algorithm is used here, and the expression is: WLC i is the weighted connection index of node i, N i is the current number of connections of node i, W i is the performance weight of node i (such as CPU performance, memory size, etc.).

[0125] A personality trait model is constructed using a neural network decision tree. Data training is performed based on contrastive learning of ciphertext similarity. Encrypted multimodal features are input and the probability distribution of the five personality dimensions is output. Specifically:

[0126] Model input is encrypted multimodal features Among them: F multi is the unencrypted joint feature vector of face, physiological and voice; E(·) represents the encryption operation based on Paillier homomorphic encryption;

[0127] The model structure adopts a hybrid neural network decision tree, and the expression is as follows:

[0128] Decision tree path selection:

[0129] G=ReLU(W gate ·E(F muiti )+b gate )

[0130] in, is the gating weight matrix; is the path selection probability, m is the number of decision tree branches; b gate is the gate bias term; RELU() is the activation function;

[0131] Leaf node prediction:

[0132]

[0133] in, is the weight matrix of the i-th leaf node; is the probability distribution of the five personality dimensions, which are extraversion, conscientiousness, openness, agreeableness, and neuroticism; G i is the path selection probability of leaf node i; is the bias term of leaf node i;

[0134] After the general personality model is generated in the cloud, it is transmitted to the edge node through the TEE secure channel;

[0135] Edge nodes regularly upload encrypted local optimization logs, and the cloud performs federated averaging to update the global personality model. When a user's personality-emotion combination deviates from the cluster center, model retraining is triggered.

[0136] The specific process of data training is as follows:

[0137] Initialize global personality model parameters Θ on the cloud golbal ;

[0138] The edge node downloads the initial model and stores the user encrypted data E(F multi ) and marked with personality label Y personality ;

[0139] Optimize the local model based on the contrastive loss function:

[0140]

[0141] Among them, SimCLR is the contrastive learning loss, which enhances the ciphertext similarity of similar personality traits; λ is the balance hyperparameter; N is the number of edge nodes participating in federated learning; Y i For labels; is the probability distribution of the five personality dimensions of leaf node i; F' multi Unencrypted facial, physiological, and voice joint features;

[0142] The edge node uploads the encrypted model gradient E(ΔΘ) at a preset time interval local );

[0143] The cloud aggregates gradients and updates the global model;

[0144]

[0145] Where D(·) is the Paillier decryption operation; K is the number of edge nodes participating in the aggregation; η is the learning rate; k is the index variable; is the global personality model parameter of the tth iteration;

[0146] Calculate the Mahalanobis distance between the user's personality-emotion combination and the cluster center:

[0147]

[0148] Where x is the user feature vector; μ, Σ are the cluster center mean and covariance matrix;

[0149] If D Mahalanobis >2σ, it triggers the cloud to retrain the user's model, where σ is the standard deviation.

[0150] Step S4: Real-time monitoring and dynamic adjustment.

[0151] Step S4.1: Implement a real-time monitoring system for cloud computing resources to continuously track the usage of computing resources (such as CPU, memory, bandwidth, etc.). When the system load approaches the preset threshold, adjust resource allocation in real time to ensure efficient resource utilization and system stability.

[0152] Step S4.2: Through a feedback mechanism, computing resources are dynamically adjusted based on the driver's real-time emotional changes and environmental adjustment needs. For example, when the driver's emotions are relatively stable, the system can reduce computing resource consumption; however, when emotions fluctuate drastically, the system automatically increases computing resources to process more data.

[0153] Step S5: Data privacy and security protection.

[0154] Step S5.1: Ensure data privacy when scheduling cloud computing resources. Sensitive data such as the owner's emotional data, facial expressions, and voice information are encrypted before being transmitted to the cloud to ensure that the data will not be leaked or tampered with during the cloud processing process.

[0155] Step S5.2: The system complies with existing data privacy regulations (such as GDPR and CCPA). All data is processed only within the scope of the vehicle owner's authorization. Data desensitization technology is used to ensure the security of the vehicle owner's identity information during cloud processing.

[0156] Step S6: Intelligent and continuous optimization of cloud resource scheduling.

[0157] Step S6.1: AI algorithms analyze the driver's long-term behavioral data and environmental adjustment history to continuously optimize cloud resource scheduling strategies. The system intelligently adjusts computing resource allocation based on the driver's individual needs and emotional state, improving system responsiveness and resource utilization efficiency.

[0158] Step S6.2: During system operation, the system learns the owner's emotional change patterns and computing needs in real time, optimizes the scheduling algorithm, makes the allocation of cloud computing resources more accurate, reduces the waste of computing resources, and ensures that the owner always experiences timely and comfortable emotional cockpit services.

[0159] Those skilled in the art will appreciate that, in addition to implementing the system, device, and various modules provided by the present invention in purely computer-readable program code, it is entirely possible to implement the same program in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, embedded microcontrollers, and the like by logically programming the method steps. Therefore, the system, device, and various modules provided by the present invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; the modules for implementing various functions can also be considered both software programs for implementing the method and structures within the hardware component.

[0160] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. The embodiments of this application and the features in the embodiments may be combined with each other in any manner unless there is a conflict.

Claims

1. A cloud-based elastic computing resource scheduling method for an emotional cockpit, characterized in that: include: Step S1: Cloud elastic computing resource scheduling demand analysis; Step S1.1: Based on the actual needs of the emotional cockpit system, calculate the data volume of each task per unit time and, combined with the computational complexity of the algorithm model, predict the overall computing load per unit time. Establish a resource demand model by task type, clarifying the configuration standards and elastic expansion limits for CPU, GPU, memory, and bandwidth. Step S1.2: Analyze the data transmission bandwidth and latency requirements between the vehicle system and the cloud to ensure low latency and high bandwidth requirements for high real-time tasks; Step S2: elastic computing resource model design; Step S2.1: Dynamically adjust cloud resources based on the resource demand prediction function; Step S2.2: Allocate resources based on resource scheduling algorithm; Step S2.3: Build an automatic expansion and contraction mechanism to add cloud nodes when the predicted resource demand exceeds the expansion threshold and recycle idle nodes when it falls below the contraction threshold; Step S3: resource scheduling algorithm implementation and optimization; Step S3.1: Scheduling resources based on task priorities; Step S3.2: Dynamically adjust resource allocation by combining real-time emotion data and environmental data, and implement resource allocation through a load balancing algorithm; Step S4: real-time monitoring and dynamic adjustment; Step S4.1: Monitor computing resource usage in real time and trigger resource allocation adjustments when the load approaches a preset threshold; Step S4.2: Dynamically adjust resources based on mood fluctuations and environmental changes, reducing resource consumption when emotions are stable and increasing resources when emotions fluctuate violently; Step S5: data privacy and security protection; Step S5.1: The driver's emotional data, facial expressions, and voice information are locally encrypted and transmitted to the cloud; Step S5.2: Process sensitive information using data desensitization technology and comply with privacy protection regulations; Step S6: Intelligent and continuous optimization of cloud resource scheduling; Step S6.1: Utilize AI algorithms to analyze vehicle owner behavior data and environmental adjustment history to optimize resource scheduling strategies; Step S6.2: Learn the emotion change pattern and computing requirements in real time, and optimize the scheduling algorithm to improve the accuracy of resource allocation.

2. The cloud-based elastic computing resource scheduling method for the emotional cockpit according to claim 1 is characterized in that: The resource demand prediction function is: Among them, R(t) is the total resource demand at time t, D i (t) represents the data processing amount or computing intensity of the i-th task at time t, w i Represents the weight coefficient of the corresponding task, and n is the total number of tasks.

3. The cloud-based elastic computing resource scheduling method for the emotional cockpit according to claim 2 is characterized in that: The resource scheduling algorithm is: Among them, A(t) is the comprehensive score of the resources required for allocation at time t; α, β, and γ are weight coefficients that control the influence of priority, complexity, and real-time performance in the allocation decision; P i 、C i 、R i Represents the priority, complexity, and real-time level value of task i.

4. The cloud-based elastic computing resource scheduling method for the emotional cockpit according to claim 3 is characterized in that: The task priority assessment formula is: Priority i =λ·U i +(1-λ)·C i Among them, Priority i Score the priority of task i, U i is the task urgency, normalized to 0-1; λ is the weight parameter, which controls the relative importance of urgency and complexity.

5. The cloud-based elastic computing resource scheduling method for the emotional cockpit according to claim 1 is characterized in that: The load balancing algorithm uses the weighted minimum connection number algorithm, which is expressed as: Among them, WLC i is the weighted connection index of node i, N i is the current number of connections of node i, W i is the performance weight of node i.

6. A cloud-based elastic computing resource scheduling system for emotional cockpits, characterized by: include: Module M1: Cloud elastic computing resource scheduling demand analysis; Module M1.1: Based on the actual needs of the emotional cockpit system, the data volume of various tasks per unit time is counted. Combined with the computational complexity of the algorithm model, the overall computing load per unit time is predicted. A resource demand model is established by task type, clarifying the configuration standards and elastic expansion limits of CPU, GPU, memory, and bandwidth. Module M1.2: Analyzes the data transmission bandwidth and latency requirements between the vehicle system and the cloud to ensure low latency and high bandwidth requirements for high real-time tasks; Module M2: Design of elastic computing resource model; Module M2.1: Dynamically adjust cloud resources based on resource demand prediction function; Module M2.2: Allocate resources based on resource scheduling algorithms; Module M2.3: Build an automatic expansion and contraction mechanism to add cloud nodes when the predicted resource demand exceeds the expansion threshold and recycle idle nodes when it falls below the contraction threshold; Module M3: Resource scheduling algorithm implementation and optimization; Module M3.1: Scheduling resources based on task priority; Module M3.2: Dynamically adjust resource allocation by combining real-time sentiment data and environmental data, and implement resource allocation through load balancing algorithms; Module M4: real-time monitoring and dynamic adjustment; Module M4.1: monitors computing resource usage in real time and triggers resource allocation adjustments when the load approaches a preset threshold; Module M4.2: Dynamically adjust resources based on mood swings and environmental changes, reducing resource consumption when emotions are stable and increasing resources when emotions fluctuate violently; Module M5: Data Privacy and Security; Module M5.1: Locally encrypts the driver's emotional data, facial expressions, and voice information and transmits them to the cloud; Module M5.2: Process sensitive information through data desensitization technology and comply with privacy protection regulations; Module M6: Intelligent and continuous optimization of cloud resource scheduling; Module M6.1: Use AI algorithms to analyze driver behavior data and environmental adjustment history to optimize resource scheduling strategies; Module M6.2: Learn emotion change patterns and computing requirements in real time, and optimize scheduling algorithms to improve resource allocation accuracy.

7. The cloud-based elastic computing resource scheduling system for emotional cockpits according to claim 6 is characterized in that: The resource demand prediction function is: Among them, R(t) is the total resource demand at time t, D i (t) represents the data processing amount or computing intensity of the i-th task at time t, w i Represents the weight coefficient of the corresponding task, and n is the total number of tasks.

8. The cloud-based elastic computing resource scheduling system for emotional cockpits according to claim 7 is characterized in that: The resource scheduling algorithm is: Among them, A(t) is the comprehensive score of the resources required for allocation at time t; α, β, and γ are weight coefficients that control the influence of priority, complexity, and real-time performance in the allocation decision; P i 、C i 、R i Represents the priority, complexity, and real-time level value of task i.

9. The cloud-based elastic computing resource scheduling system for emotional cockpits according to claim 8, characterized in that: The task priority assessment formula is: Priority i =λ·U i +(1-λ)·C i Among them, Priority i Score the priority of task i, U i is the task urgency, normalized to 0-1; λ is the weight parameter, which controls the relative importance of urgency and complexity.

10. The cloud-based elastic computing resource scheduling system for emotional cockpits according to claim 6, characterized in that: The load balancing algorithm uses the weighted minimum connection number algorithm, which is expressed as: Among them, WLC i is the weighted connection index of node i, N i is the current number of connections of node i, W i is the performance weight of node i.

Images