A collaborative control method and system for intelligent driving test equipment

Abstract

The present invention provides a collaborative control method and system for intelligent driving test equipment, which selects the first template collaborative control demand that best matches the current collaborative control demand from multiple template collaborative control demands through precise matching of collaborative knowledge data, effectively improving the accuracy and efficiency of demand identification. By utilizing a deep learning network, combined with the first template collaborative control demand and its template response knowledge data, deep learning and response feedback are performed on the original collaborative control demand to generate a first guidance result, and further specific response knowledge data is generated through decision-making, which not only enhances the intelligence level of collaborative control, but also significantly improves the response speed and control accuracy. Ultimately, based on the response knowledge data, coordinated control operations are performed on the intelligent driving test equipment, achieving efficient and precise control of the test equipment, providing data support for the testing and verification of intelligent driving technology, and thus greatly promoting the application of intelligent driving research and development.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving technology, and in particular to a collaborative control method and system for intelligent driving test equipment. Background Art

[0002] With the rapid development of intelligent driving technology, intelligent driving test equipment plays a vital role in technology research, development, and verification. However, the coordinated control of intelligent driving test equipment faces numerous challenges. Traditional control methods often rely on manually set rules and parameters, making them difficult to adapt to complex and changing test scenarios and requirements. This results in limited control effectiveness and a lack of flexibility and intelligence.

[0003] During intelligent driving testing, collaborative control requirements are often diverse and complex, with different test scenarios and tasks placing varying demands on control strategies. To effectively address these demands, the traditional approach involves manually analyzing test scenarios and developing corresponding control strategies based on experience. However, this approach is not only time-consuming and labor-intensive, but also struggles to ensure the accuracy and optimality of control strategies.

[0004] Furthermore, with the continuous advancement of intelligent driving technology, the functionality and performance of test equipment are also constantly improving, placing higher demands on the intelligence level of collaborative control methods. Traditional rule-based control methods are no longer able to meet this demand. There is an urgent need for a collaborative control method that can automatically adapt to different test scenarios and requirements and possesses deep learning and decision-making capabilities. Summary of the Invention

[0005] In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a collaborative control method for intelligent driving test equipment, the method comprising:

[0006] Extracting a first template collaborative control requirement from the multiple template collaborative control requirements based on a degree of matching between collaborative knowledge data of the collaborative control requirement of the intelligent driving test equipment and collaborative knowledge data of multiple template collaborative control requirements, wherein the collaborative knowledge data is used to express a requirement knowledge label corresponding to the collaborative control requirement or a knowledge element associated with the corresponding collaborative control requirement, and the matching degree corresponding to the first template collaborative control requirement is greater than the matching degrees corresponding to the remaining template collaborative control requirements in the multiple template collaborative control requirements;

[0007] Generate a first guidance result based on the collaborative control requirement, the first template collaborative control requirement, and the template response knowledge data of the first template collaborative control requirement, wherein the first guidance result is used to express that the deep learning network responds to the collaborative control requirement based on the first template collaborative control requirement and the template response knowledge data;

[0008] Using the deep learning network, making a decision on the first guidance result to generate response knowledge data of the collaborative control requirement;

[0009] The intelligent driving test equipment is subjected to coordinated control operations based on the response knowledge data of the coordinated control requirements.

[0010] On the other hand, an embodiment of the present invention also provides a collaborative control system for intelligent driving test equipment, including a processor and a machine-readable storage medium, wherein the machine-readable storage medium is connected to the processor, the machine-readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine-readable storage medium to implement the above method.

[0011] Based on the above aspects, the embodiment of the present application selects the first template collaborative control demand that best matches the current collaborative control demand from multiple template collaborative control demands by accurately matching collaborative knowledge data, thereby effectively improving the accuracy and efficiency of demand identification. By utilizing a deep learning network, combined with the first template collaborative control demand and its template response knowledge data, deep learning and response feedback are performed on the original collaborative control demand to generate a first guidance result, and further specific response knowledge data is generated through decision-making, which not only enhances the intelligence level of collaborative control, but also significantly improves the response speed and control accuracy. Ultimately, coordinated control operations are performed on intelligent driving test equipment based on response knowledge data, achieving efficient and precise control of the test equipment, providing data support for the testing and verification of intelligent driving technology, and thus greatly promoting the application of intelligent driving research and development. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a schematic diagram of the execution flow of the collaborative control method for intelligent driving test equipment provided by an embodiment of the present invention.

[0013] Figure 2 This is a schematic diagram of the hardware architecture of a collaborative control system for intelligent driving test equipment provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The present invention will be described in detail below with reference to the accompanying drawings. Figure 1 This is a flow chart of a collaborative control method for intelligent driving test equipment provided by an embodiment of the present invention. The collaborative control method for intelligent driving test equipment is introduced in detail below.

[0015] Step S110, based on the matching degree between the collaborative knowledge data of the collaborative control requirements of the intelligent driving test equipment and the collaborative knowledge data of multiple template collaborative control requirements, extract the first template collaborative control requirement from the multiple template collaborative control requirements, the collaborative knowledge data is used to express the requirement knowledge label of the corresponding collaborative control requirement or the knowledge element associated with the corresponding collaborative control requirement, and the matching degree corresponding to the first template collaborative control requirement is greater than the matching degree corresponding to the remaining template collaborative control requirements in the multiple template collaborative control requirements.

[0016] In this embodiment, during the process of performing collaborative control related work of the intelligent driving test equipment, the intelligent driving test equipment needs to simultaneously coordinate the work of multiple sensors (such as lidar, cameras, millimeter-wave radar, etc.) and different actuators (such as steering systems, braking systems, acceleration systems, etc.).

[0017] The collaborative knowledge data required for collaborative control includes multiple aspects. For example, the first-dimensional collaborative knowledge data (used to express the knowledge labels corresponding to the collaborative control requirements) may include labels such as "multi-sensor collaboration under complex road conditions" and "actuator priority in emergency response." The second-dimensional collaborative knowledge data (used to express the knowledge elements corresponding to the collaborative control requirements) may include various specific road conditions (such as intersections, curves, uphill and downhill slopes, etc.), the characteristics of different sensors (such as the detection range of lidar, the resolution and field of view of the camera, etc.), the response time and accuracy of the actuator, etc.

[0018] The server stores multiple template collaborative control requirements and their collaborative knowledge data. For example, Template Collaborative Control Requirement 1 might be for the coordination of sensors and actuators during simple straight-line driving. Its first dimension of collaborative knowledge data is labeled "straight-line driving coordination," and its second dimension of collaborative knowledge data includes some basic information about the straight-line (such as road width, vehicle speed range, etc.). Template Collaborative Control Requirement 2 might be for the coordination of driving on a curved road in rainy weather. Its first dimension of collaborative knowledge data is labeled "rainy-day curved road coordination," and its second dimension of collaborative knowledge data includes factors such as the curvature of the curve and the impact of rainfall on sensors.

[0019] There are multiple ways for the server to calculate the matching degree between the collaborative control requirement and the collaborative control requirement of each template.

[0020] If we match the collaborative knowledge data based on the first dimension first:

[0021] The server first compares the first-dimensional collaborative knowledge data of the intelligent driving test equipment collaborative control requirement with the first-dimensional collaborative knowledge data of each template collaborative control requirement. For example, a collaborative control requirement labeled "Emergency Collaboration of Multiple Sensors and Actuators at Complex Intersections" would have a low match with the "Straight-Road Collaboration" tag of Template Collaborative Control Requirement 1 because the scenarios and tasks are completely different. However, the first-dimensional collaborative knowledge data of Template Collaborative Control Requirement 3 (assuming it's "Regular Collaboration at Intersections") would have a relatively high match because both involve intersection scenarios. This initially extracts multiple second template collaborative control requirements from the multiple template collaborative control requirements (here assuming Template Collaborative Control Requirement 3 is extracted as one of the second template collaborative control requirements). The server then performs further screening based on the match between the second-dimensional collaborative knowledge data of the collaborative control requirement and the second-dimensional collaborative knowledge data of each second template collaborative control requirement. For example, the intersection in the second-dimensional collaborative knowledge data of the collaborative control requirement is a busy four-way intersection, while the intersection in the second-dimensional collaborative knowledge data of Template Collaborative Control Requirement 3 is a simpler two-way intersection with low traffic volume, resulting in a lower match. If there are other second template collaborative control requirements, continue to compare and finally extract the first template collaborative control requirement. This first template collaborative control requirement has the highest matching degree with the collaborative control requirement of the intelligent driving test equipment among all template collaborative control requirements.

[0022] If we first match the collaborative knowledge data based on the second dimension:

[0023] The server first compares the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each template collaborative control requirement. For example, for the sensor characteristic part in the collaborative control requirement, such as the detection requirement of the laser radar in a complex environment, it is compared with the sensor-related knowledge elements in each template collaborative control requirement. If the part about the laser radar in the second dimension collaborative knowledge data of the template collaborative control requirement 4 has a certain similarity with the collaborative control requirement (for example, both involve the detection and processing of the laser radar under partial occlusion), it is extracted as one of the multiple third template collaborative control requirements. Then, it is screened based on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each third template collaborative control requirement. Assuming that the first dimension collaborative knowledge data of the collaborative control requirement is "multi-sensor collaborative response to emergencies in complex environments", and the first dimension collaborative knowledge data of the template collaborative control requirement 4 is "collaboration between laser radar and other sensors in specific scenarios", after comparison, the first template collaborative control requirement is determined from multiple third template collaborative control requirements.

[0024] There is also a weighted calculation method:

[0025] The server performs weighted calculations on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each template collaborative control requirement, and the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each template collaborative control requirement. For example, the first dimension matching degree is weighted as 0.4, and the second dimension matching degree is weighted as 0.6. For template collaborative control requirement 5, the first dimension collaborative knowledge data matching degree is calculated to be 0.3 (full score 1), and the second dimension collaborative knowledge data matching degree is calculated to be 0.5 (full score 1), then the first matching degree is obtained by weighted calculation (0.3×0.4+0.5×0.6=0.42). Such calculations are performed on all template collaborative control requirements, and finally the first template collaborative control requirement is extracted based on the first matching degrees corresponding to multiple template collaborative control requirements.

[0026] In addition, you can also combine the collaborative control requirements themselves with the template collaborative control requirements to match:

[0027] Based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, multiple fourth template collaborative control requirements are extracted from multiple template collaborative control requirements. For example, the knowledge elements in the collaborative control requirement involve the collaborative working time interval of multiple sensors under complex road conditions and other contents, which have a certain similarity with the time logic of the sensor collaborative work in template collaborative control requirement 6, so template collaborative control requirement 6 is extracted as one of the fourth template collaborative control requirements. Then, based on the matching degree between the collaborative control requirement and each fourth template collaborative control requirement, the first template collaborative control requirement is extracted from multiple fourth template collaborative control requirements. For example, the actuator response sequence in the collaborative control requirement is compared with the actuator operation logic in template collaborative control requirement 6. If the matching degree is higher than that of other fourth template collaborative control requirements, then template collaborative control requirement 6 is determined to be the first template collaborative control requirement.

[0028] Alternatively, the matching degree between collaborative control requirements and template collaborative control requirements can be compared first, and then the matching degree between collaborative knowledge data can be compared:

[0029] Based on the matching degree between the collaborative control requirement and each template collaborative control requirement, multiple fifth template collaborative control requirements are extracted from multiple template collaborative control requirements. For example, if the overall task logic in the collaborative control requirement (such as first collecting sensor data, and then judging the actuator operation sequence based on the data) is partially similar to the overall task logic of template collaborative control requirement 7, template collaborative control requirement 7 is extracted as one of the fifth template collaborative control requirements. Then, based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each fifth template collaborative control requirement, the first template collaborative control requirement is extracted from multiple fifth template collaborative control requirements. For example, the sensor accuracy requirements in the collaborative knowledge data of the collaborative control requirement are compared with the sensor-related elements in the collaborative knowledge data of template collaborative control requirement 7. If the matching degree is the highest, then template collaborative control requirement 7 becomes the first template collaborative control requirement.

[0030] There is another way to think about the cost of collaborative control:

[0031] The collaborative control cost corresponding to the template collaborative control requirement is used to express the logical node span of the template response knowledge data. For example, the logical node span of the template response knowledge data of template collaborative control requirement 8 is large, which means that when processing this requirement, more logical judgments and operation steps are involved, and the collaborative control cost is high. The server extracts multiple sixth template collaborative control requirements from multiple template collaborative control requirements based on the collaborative control cost corresponding to each template collaborative control requirement. Assume that a threshold value of the collaborative control cost is set, and the template collaborative control requirements above this threshold are extracted as the sixth template collaborative control requirements. Then, based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each sixth template collaborative control requirement, the first template collaborative control requirement is extracted from the multiple sixth template collaborative control requirements. For example, the road condition information in the collaborative knowledge data of the collaborative control requirement is compared with the relevant road condition information in the collaborative knowledge data of the template collaborative control requirement 8. If the matching degree is higher than that of other sixth template collaborative control requirements, then the template collaborative control requirement 8 is determined to be the first template collaborative control requirement.

[0032] Or first extract based on the matching degree of collaborative knowledge data, and then filter based on the collaborative control cost:

[0033] Based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, multiple fourth template collaborative control requirements are extracted from the multiple template collaborative control requirements. Then, based on the collaborative control cost corresponding to each fourth template collaborative control requirement, the first template collaborative control requirement is extracted from the multiple fourth template collaborative control requirements. For example, among the multiple fourth template collaborative control requirements, the collaborative control cost of template collaborative control requirement 9 is relatively high (the logical node span of its template response knowledge data is large, and may involve more complex interaction logic between sensors and actuators). If its matching degree with the collaborative knowledge data of the collaborative control requirement is the highest among these fourth template collaborative control requirements, then template collaborative control requirement 9 is determined to be the first template collaborative control requirement.

[0034] Finally, we can also use weighted calculation method:

[0035] The matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the collaborative control cost corresponding to each template collaborative control requirement are weighted and calculated to generate the third matching degree corresponding to each template collaborative control requirement. For example, the collaborative knowledge data matching degree is weighted as 0.7, and the collaborative control cost is weighted as 0.3. For template collaborative control requirement 10, the collaborative knowledge data matching degree is calculated to be 0.6 (full score 1), and the collaborative control cost is at a high level (assuming the corresponding value is 0.8, full score 1). The third matching degree is obtained by weighted calculation (0.6×0.7+0.8×0.3=0.66). Based on the third matching degrees corresponding to multiple template collaborative control requirements, the first template collaborative control requirement is extracted.

[0036] Step S120: Generate a first guidance result based on the collaborative control requirement, the first template collaborative control requirement, and the template response knowledge data of the first template collaborative control requirement. The first guidance result is used to express that the deep learning network responds to and provides feedback on the collaborative control requirement based on the first template collaborative control requirement and the template response knowledge data.

[0037] In this embodiment, in the intelligent driving test scenario, the collaborative control requirement is that in a complex urban road environment, the vehicle needs to coordinate multiple sensors (lidar, cameras, etc.) under different traffic rules (such as traffic restrictions at different times, speed limits on special sections of roads, etc.) to accurately perceive the surrounding environment (including vehicles, pedestrians, traffic signs, etc.), and at the same time control actuators (steering, braking, acceleration, etc.) to ensure safe and efficient driving.

[0038] The first template collaborative control requirement addresses specific scenarios in complex urban road scenarios, such as those affecting complex traffic conditions in specific areas (e.g., commercial districts). The template response knowledge data includes specific sensor operating modes (e.g., high frame rate camera acquisition in high-traffic areas) and actuator operation logic (e.g., varying degrees of braking based on obstacle distances) for these scenarios.

[0039] The server first analyzes the key elements of the collaborative control requirements, including various traffic rules, different types of road users (vehicles, pedestrians, etc.), the types and number of sensors, and types of actuators. Constraints include maximum vehicle speed limits and sensor blind spots. The goal is safe and efficient driving on urban roads. These elements are then structured into a collaborative control requirement description. For example, the key element set includes {traffic rule set, road user type set, sensor set, actuator set}, the constraint set includes {speed limit set, blind spot set}, and the goal requirement set is {safe driving, efficient driving}.

[0040] Then, the template structure of the first template collaborative control requirement is parsed, such as a hierarchical control structure (sensor data processing first, then decision-making layer, and finally actuator operation). The key element template includes common traffic elements in a specific area (such as the layout of traffic lights in a specific commercial area, etc.), constraint templates (such as special speed limits in the area) and target requirement templates (such as ensuring pedestrian priority while ensuring normal vehicle traffic in the area). The parsed first template collaborative control requirement is generated, including a template structure description, a key element template set, a constraint template set and a target requirement template set.

[0041] Then, the response logic in the template response knowledge data is parsed, such as the judgment logic based on distance and speed (determining whether to brake based on the distance and relative speed to the vehicle in front), response rules (such as when the distance is less than a certain value, the actuator brakes at a certain proportion) and response parameters (such as the specific value of the braking ratio, etc.), to generate the parsed template response knowledge data, including the response logic description, response rule set and response parameter set.

[0042] The key elements, constraints, and target requirements of the collaborative control requirement are matched with the key element template, constraint template, and target requirement template of the first collaborative control requirement template. For example, the traffic rules in the collaborative control requirement are compared with the traffic rules template of the first collaborative control requirement template to calculate the degree of match. If there are some special traffic rules in the collaborative control requirement that are not fully covered in the template, the corresponding matching degree calculation results and the corresponding relationship description between the collaborative control requirement and the collaborative control requirement of the first template will be obtained.

[0043] Based on the matching results and the corresponding relationship description, the template's response knowledge data is evaluated for its adaptability to the collaborative control requirements. If the sensor operating mode in the template doesn't fully match the sensor conditions in the collaborative control requirements, then this portion of the response knowledge data needs to be adjusted. If the template doesn't address actuator operations under certain special traffic rules in the collaborative control requirements, then this portion needs to be supplemented. Parts that fully match the template are directly applicable response knowledge data. The resulting adaptability evaluation results include directly applicable response knowledge data, those requiring adjustment, and those requiring supplementation.

[0044] For the response knowledge data that needs to be adjusted, adjustments are made based on the collaborative control requirements of the collaborative control needs. For example, the response rules for braking operations in the template may need to be adjusted based on the stricter speed limits in the collaborative control needs. For the response knowledge data that needs to be supplemented, a supplementary design is performed based on the demand characteristics of the collaborative control needs and the structure of the template response knowledge data. For example, for operations under special traffic rules in the collaborative control needs, the corresponding actuator operation logic is supplemented based on the logical structure in the template. The directly applicable response knowledge data, the adjusted response knowledge data, and the supplemented response knowledge data are then integrated to form an integrated response knowledge data set.

[0045] Based on the integrated response knowledge data set and collaborative control requirements, the learning parameters of the deep learning network are configured. For example, the learning rate is set according to the complexity of the data and the real-time requirements of the collaborative control needs, the number of iterations is determined according to the expected convergence speed, and the appropriate optimization algorithm (such as stochastic gradient descent) is selected according to the characteristics of the response knowledge data.

[0046] Finally, using the first template collaborative control requirement and the parsed template response knowledge data as the learning basis, the configured deep learning network conducts preliminary learning of the collaborative control requirement, generating a first guidance result. This first guidance result includes the deep learning network's preliminary understanding of the collaborative control requirement (such as its understanding of the complex traffic rules and the relationship between sensors and actuators), the initially set response logic (such as preliminary operation logic based on the current traffic conditions), and the response parameters (such as preliminary actuator operation parameters).

[0047] Step S130: Utilize the deep learning network to make a decision on the first guidance result and generate knowledge data that responds to the collaborative control requirement.

[0048] Continuing with the aforementioned intelligent driving test scenario, the server uses a deep learning network to classify collaborative control requirements based on the first guidance result. For example, based on the initial understanding description, response logic, and response parameters in the first guidance result, the collaborative control requirements are classified as "normal driving collaborative control in complex urban traffic" or "special situation collaborative control in complex urban traffic."

[0049] Based on the first template collaborative control requirement and template response knowledge data, the collaborative control requirement is responded to and fed back according to the collaborative control response category corresponding to the collaborative control requirement. For example, if it is a "normal driving collaborative control category in complex urban traffic", the deep learning network generates response knowledge data for this collaborative control response category based on the sensor data processing method (such as camera and lidar data fusion processing) and actuator operation logic (such as smooth acceleration or deceleration according to traffic flow) during normal driving in the first template collaborative control requirement, combined with the specific response rules and parameters in the template response knowledge data. This response knowledge data may include more accurate sensor data processing algorithms, actuator operation instructions that are more optimized for current traffic conditions, etc.

[0050] Step S140 : performing coordinated control operations on the intelligent driving test equipment based on the response knowledge data of the coordinated control requirements.

[0051] In the above-mentioned intelligent driving test scenario, the intelligent driving test equipment is coordinated and controlled based on the response knowledge data of the collaborative control requirements. The server sends the generated response knowledge data to the control system of the intelligent driving test equipment.

[0052] For sensors, the operating modes of the LiDAR and camera are adjusted based on the sensor data processing algorithms in the response knowledge data. For example, if the response knowledge data requires improved detection accuracy for small targets at long distances, the server will send instructions to the LiDAR system to adjust parameters such as its transmit power and scanning frequency. For the camera system, parameters such as its image acquisition resolution and frame rate may be adjusted to better meet collaborative control requirements.

[0053] The actuators control the steering, braking, and acceleration systems based on the actuator operation instructions in the response knowledge data. For example, if the response knowledge data indicates slow-moving traffic ahead, the server sends instructions to the braking system to gradually reduce the vehicle's speed. If a turn is required, the server sends instructions to the steering system based on the steering angle and speed requirements in the response knowledge data to ensure accurate steering. This coordinated control ensures that intelligent driving test equipment operates safely and efficiently in complex urban road environments, in accordance with collaborative control requirements.

[0054] Based on the above steps, the embodiment of the present application selects the first template collaborative control demand that best matches the current collaborative control demand from multiple template collaborative control demands by accurately matching the collaborative knowledge data, thereby effectively improving the accuracy and efficiency of demand identification. By utilizing a deep learning network, combined with the first template collaborative control demand and its template response knowledge data, deep learning and response feedback are performed on the original collaborative control demand to generate a first guidance result, and further specific response knowledge data is generated through decision-making, which not only enhances the intelligence level of collaborative control, but also significantly improves the response speed and control accuracy. Ultimately, based on the response knowledge data, coordinated control operations are performed on the intelligent driving test equipment to achieve efficient and precise control of the test equipment, providing data support for the testing and verification of intelligent driving technology, thereby greatly promoting the application of intelligent driving research and development.

[0055] In one possible implementation, the collaborative knowledge data includes first-dimensional collaborative knowledge data and second-dimensional collaborative knowledge data, wherein the first-dimensional collaborative knowledge data is used to express the demand knowledge tags corresponding to the collaborative control requirements, and the second-dimensional collaborative knowledge data is used to express the knowledge elements corresponding to the collaborative control requirements. Step S110 may include:

[0056] Extract multiple second template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each template collaborative control requirement. Extract the first template collaborative control requirement from the multiple second template collaborative control requirements based on the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each second template collaborative control requirement. Or,

[0057] Extract multiple third template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each template collaborative control requirement. Extract the first template collaborative control requirement from the multiple third template collaborative control requirements based on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each third template collaborative control requirement. Or,

[0058] A weighted calculation is performed on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each template collaborative control requirement, and the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each template collaborative control requirement to generate a first matching degree corresponding to each template collaborative control requirement. Based on the first matching degrees corresponding to the multiple template collaborative control requirements, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements.

[0059] In this embodiment, multiple sensors (lidar, cameras, millimeter-wave radar, etc.) and actuators (steering, braking, acceleration systems, etc.) collaborate under various special road conditions (such as slippery roads, narrow curves, and temporary construction areas). The collaborative knowledge data here consists of two dimensions. The first dimension expresses the required knowledge tags, such as "multi-sensor actuator collaboration under complex and special road conditions." The second dimension contains knowledge elements, such as the friction coefficient range of slippery roads, the curvature of narrow curves, the location of temporary construction areas, and their impact on the sensor's field of view.

[0060] The server stores multiple template collaborative control requirements and their collaborative knowledge data. In accordance with the matching method based on the first dimension collaborative knowledge data, the server first compares the first dimension collaborative knowledge data of the collaborative control requirement with the first dimension collaborative knowledge data of each template collaborative control requirement. For example, the first dimension collaborative knowledge data label of template collaborative control requirement 1 is "basic collaboration under normal road conditions", which has a low matching degree with the label of the collaborative control requirement "multi-sensor actuator collaboration under complex and special road conditions"; while the label of template collaborative control requirement 3 is "partial sensor actuator collaboration under special road conditions", which has a relatively high matching degree. Through such comparison, multiple second template collaborative control requirements are extracted from multiple template collaborative control requirements, such as template collaborative control requirement 3. Next, the server matches the second dimension collaborative knowledge data of the collaborative control requirement with the second dimension collaborative knowledge data of each second template collaborative control requirement. For example, for template collaborative control requirement 3, the special road conditions in its second-dimensional collaborative knowledge data may be just a simple slippery road surface, while the special road conditions in the collaborative control requirements also include narrow curves and temporary construction areas. Through a detailed comparison of these knowledge elements, the first template collaborative control requirement is finally determined from multiple second template collaborative control requirements. This first template collaborative control requirement has the highest matching degree with the collaborative control requirement in this comparison process.

[0061] According to the method of first matching the collaborative knowledge data of the second dimension, the server first compares the collaborative knowledge data of the second dimension of the collaborative control requirement with the collaborative knowledge data of the second dimension of each template collaborative control requirement. For example, the knowledge elements such as the range of friction coefficient of slippery roads and curvature of narrow curves in the collaborative control requirement are compared with the relevant elements in the second dimension collaborative knowledge data of template collaborative control requirement 5. If there is a certain similarity, the template collaborative control requirement 5 is extracted as one of multiple third template collaborative control requirements. Then, the server compares the collaborative knowledge data of the first dimension of the collaborative control requirement with the first dimension collaborative knowledge data of each third template collaborative control requirement. For example, the "multi-sensor actuator collaboration under complex and special road conditions" label of the collaborative control requirement is compared with the first dimension collaborative knowledge data label of template collaborative control requirement 5. If the matching degree is higher than that of other third template collaborative control requirements, then template collaborative control requirement 5 is determined to be the first template collaborative control requirement.

[0062] There is also a weighted calculation method, in which the server performs weighted calculation on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each template collaborative control requirement, and the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each template collaborative control requirement. For example, the weight of the first dimension matching is set to 0.3, and the weight of the second dimension matching is set to 0.7. For template collaborative control requirement 7, the matching degree of the first dimension collaborative knowledge data is calculated to be 0.2, and the matching degree of the second dimension collaborative knowledge data is 0.4. The first matching degree is obtained by weighted calculation (0.2×0.3+0.4×0.7=0.34). Such calculation is performed on all template collaborative control requirements, and then the first template collaborative control requirement is extracted based on the first matching degrees corresponding to multiple template collaborative control requirements. The first matching degree of this first template collaborative control requirement is the highest among all template collaborative control requirements. Through the above different methods, the server can accurately extract the first template collaborative control requirement that best matches the collaborative control requirement of intelligent driving test equipment from many template collaborative control requirements.

[0063] In a possible implementation, before step S110, the method further includes:

[0064] Based on the collaborative control requirement, a second guidance result is generated, where the second guidance result is used to express the decision made by the deep learning network on the collaborative control requirement to generate a data result for expressing the collaborative control requirement.

[0065] The deep learning network is used to make a decision on the second guidance result to generate collaborative knowledge data of the collaborative control requirement.

[0066] In this embodiment, the collaborative control requirements include many aspects. For example, in changeable weather conditions (such as rain, fog, snow, etc.) and complex road types (such as highways, narrow urban streets, mountain curves, etc.), multiple sensors (lidar, cameras, millimeter-wave radars, etc.) are required to accurately collect environmental information, and at the same time, actuators (steering, braking, acceleration systems, etc.) are required to perform precise operations based on the feedback from the sensors. The server inputs such complex collaborative control requirements into the deep learning network. The deep learning network analyzes and processes this collaborative control requirement based on its internal algorithm structure and pre-trained model to generate a second guidance result. This second guidance result is intended to express the deep learning network's decision-making on the collaborative control requirements to generate data results used to express the collaborative control requirements. Specifically, the deep learning network may preliminarily determine some key data points based on its understanding of sensor performance under different weather conditions and the requirements of vehicle driving on different road types, such as the approximate range of camera visibility in rainy and foggy weather, the operating frequency range of the steering system when driving on mountain curves, etc. These data points constitute the second guidance result.

[0067] Next, after receiving the second guidance result, the deep learning network conducts further analysis. For example, for the previously determined approximate camera viewing distance range in rainy and foggy weather, the network incorporates additional factors for precise calculation, such as the effect of rain and fog concentration on light propagation and the characteristics of the camera lens. This allows for more accurate camera viewing distance values ​​under different rain and fog concentrations, which become part of the collaborative knowledge data. Similarly, for the steering system's operating frequency range when driving on a mountain curve, the deep learning network considers factors such as curve curvature, vehicle speed, and tire-road friction to accurately calculate the exact steering system operating frequency for different curve curvatures and vehicle speeds, which is also incorporated into the collaborative knowledge data. Furthermore, the deep learning network comprehensively considers the collaborative relationships between sensors, such as how lidar and cameras complement each other in collecting data in different environments, as well as the coordination logic between actuators, such as the sequencing and coordination between braking and steering systems during emergency avoidance maneuvers, and integrates these into the collaborative knowledge data required for collaborative control. Therefore, with the help of deep learning networks, comprehensive and accurate collaborative knowledge data is generated based on collaborative control requirements, preparing for the subsequent extraction of the first template collaborative control requirement from multiple template collaborative control requirements.

[0068] In a possible implementation, before step S110, the method further includes:

[0069] Step A110 : Acquire the multiple template collaborative control requirements and candidate response knowledge data for each template collaborative control requirement.

[0070] Step A120: utilizing the deep learning network, based on each template collaborative control requirement, adapting and optimizing the candidate response knowledge data of each template collaborative control requirement, and generating the template response knowledge data of each template collaborative control requirement.

[0071] In a possible implementation, before step A120, the method further includes: generating a third guidance result based on the template collaborative control requirement and the candidate response knowledge data of the template collaborative control requirement, and the third guidance result is used to express that the deep learning network adapts and optimizes the candidate response knowledge data of the template collaborative control requirement based on the template collaborative control requirement.

[0072] Step A120 includes:

[0073] Step A121: Utilize the deep learning network to make a decision on the third guidance result and generate iterative response knowledge data of the template collaborative control requirement.

[0074] Step A122: When the iterative response knowledge data of the template collaborative control requirement and the candidate response knowledge data of the template collaborative control requirement express the same response result, the iterative response knowledge data of the template collaborative control requirement is output as the template response knowledge data of the template collaborative control requirement.

[0075] In a possible implementation, the plurality of template collaborative control requirements and the template response knowledge data of each template collaborative control requirement are stored in a pre-built sample library. After step A120, the method further includes:

[0076] Step A130, based on any template collaborative control requirement and the template response knowledge data of the template collaborative control requirement, generates a fourth guidance result, and the fourth guidance result is used to express that the deep learning network uses the template collaborative control requirement and the template response knowledge data of the template collaborative control requirement as the learning basis to generate iterative collaborative control requirements and response knowledge data.

[0077] Step A140: Utilize the deep learning network to make a decision on the fourth guidance result, and generate iterative template collaborative control requirements and template response knowledge data of the iterative template collaborative control requirements.

[0078] Step A150: adding the iterative template collaborative control requirement and the template response knowledge data of the iterative template collaborative control requirement to the sample library.

[0079] In this embodiment, first, before extracting the first template collaborative control requirement based on the matching degree between the collaborative knowledge data of the collaborative control requirement of the intelligent driving test equipment and the collaborative knowledge data of multiple template collaborative control requirements, the server needs to obtain multiple template collaborative control requirements and candidate response knowledge data for each template collaborative control requirement. In the field of intelligent driving, template collaborative control requirements cover a variety of situations. For example, for different road scenarios (such as urban congested roads, rural simple roads, highways, etc.), different traffic conditions (such as peak hours, off-peak hours, traffic control caused by special events, etc.) and different vehicle states (such as full load, empty load, specific fault simulation, etc.), there are corresponding template collaborative control requirements. The candidate response knowledge data for each template collaborative control requirement is the knowledge related to possible response solutions generated based on past experience or preliminary algorithms. Taking the urban congested road scenario as an example, its template collaborative control requirements may include sensors (such as lidar, cameras, etc.) needing to collect vehicle and pedestrian information within a close range more frequently, and actuators (such as braking and acceleration systems) needing to perform fine-tuning operations more accurately. The corresponding candidate response knowledge data may include the preliminary setting value of the camera acquisition frequency, the preliminary force range of braking and acceleration operations, etc.

[0080] Next, the server uses the deep learning network to adapt and optimize the candidate response knowledge data of each template collaborative control requirement based on each template collaborative control requirement to generate template response knowledge data for each template collaborative control requirement. But before that, the server must generate a third guidance result based on the template collaborative control requirement and the candidate response knowledge data of the template collaborative control requirement. For example, for the template collaborative control requirements of urban congested roads and their corresponding candidate response knowledge data, the deep learning network will analyze the logical relationship therein, such as comprehensively considering preliminary settings such as camera acquisition frequency and braking and acceleration operation force range based on factors such as traffic flow and vehicle spacing, and generate a third guidance result. This third guidance result expresses the direction and strategy of the deep learning network to adapt and optimize the candidate response knowledge data based on the template collaborative control requirements.

[0081] Then, the server uses the deep learning network to make a decision on the third guidance result and generate iterative response knowledge data for the template collaborative control requirement. Taking urban congested roads as an example, the deep learning network will make more precise adjustments to the camera acquisition frequency and the braking and acceleration operation force range based on more accurate traffic flow models, vehicle dynamic behavior analysis, etc., to obtain iterative response knowledge data. If this iterative response knowledge data is the same as the response result expressed by the candidate response knowledge data of the template collaborative control requirement, this iterative response knowledge data is output as the template response knowledge data of the template collaborative control requirement. For example, after optimization by the deep learning network, the final result of the camera acquisition frequency and the braking and acceleration operation force range is consistent with the initial candidate response knowledge data in actual effect, then this final result becomes the template response knowledge data of the template collaborative control requirement.

[0082] In the intelligent driving test, multiple template collaborative control requirements and the template response knowledge data of each template collaborative control requirement are stored in a pre-built sample library. After using the deep learning network to adapt and optimize the candidate response knowledge data of each template collaborative control requirement based on each template collaborative control requirement to generate the template response knowledge data of each template collaborative control requirement, the server shall generate a fourth guidance result based on any template collaborative control requirement and the template response knowledge data of the template collaborative control requirement. For example, the template collaborative control requirement and its template response knowledge data of the urban congested road are selected. The deep learning network will use this template collaborative control requirement and template response knowledge data as the basis for learning, taking into account more actual changes (such as the impact of different weather conditions on traffic, the addition of special vehicle types, etc.), and generate the fourth guidance result. This result is used to express the process of the deep learning network generating iterative collaborative control requirements and response knowledge data.

[0083] Finally, the server uses a deep learning network to make decisions on the fourth guidance result, generating iterative template collaborative control requirements and template response knowledge data for iterative template collaborative control requirements. For example, after taking into account the weather and special vehicle types, new iterative template collaborative control requirements for urban congested roads are generated. For example, on rainy days, sensors need to pay extra attention to waterlogged areas on the road, and actuators need to adjust braking and acceleration strategies according to the slippery road conditions, etc., and corresponding template response knowledge data are also generated. Then, the iterative template collaborative control requirements and template response knowledge data for iterative template collaborative control requirements are added to the sample library, so that in subsequent intelligent driving tests, more comprehensive and accurate template references can be provided for more collaborative control requirements matching. Through such a series of operations, the server continuously optimizes and expands the template collaborative control requirements and template response knowledge data in the sample library, improving the accuracy and effectiveness of the collaborative control of intelligent driving test equipment.

[0084] In a possible implementation, step S110 may further include:

[0085] Extracting multiple fourth template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement. Extracting the first template collaborative control requirement from the multiple fourth template collaborative control requirements based on the matching degree between the collaborative control requirement and each fourth template collaborative control requirement. Or,

[0086] Extracting multiple fifth template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the collaborative control requirement and each template collaborative control requirement. Extracting the first template collaborative control requirement from the multiple fifth template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each fifth template collaborative control requirement. Or,

[0087] A weighted calculation is performed on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the matching degree between the collaborative control requirement and each template collaborative control requirement to generate a second matching degree corresponding to each template collaborative control requirement. Based on the second matching degrees corresponding to the multiple template collaborative control requirements, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements.

[0088] In this embodiment, let's first look at the first method. The server extracts multiple fourth template collaborative control requirements from multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement. For example, in the complex scenario of intelligent driving, the collaborative knowledge data of the collaborative control requirement includes multiple knowledge elements such as the collaborative working mode of multiple sensors (lidar, camera, millimeter wave radar, etc.) under specific complex road conditions (such as roads with many bends in mountainous areas and some sections under construction), as well as the operating logic of actuators (steering, braking, acceleration systems, etc.). The server compares these collaborative knowledge data with the collaborative knowledge data of each template collaborative control requirement. For template collaborative control requirement 1, its collaborative knowledge data is for the collaborative situation of sensors and actuators on ordinary urban roads, and has a low matching degree with the collaborative knowledge data under complex mountainous road conditions of the current collaborative control requirement; while for template collaborative control requirement 3, its collaborative knowledge data involves the situation of some mountainous roads, which has a certain similarity with the current collaborative control requirement. The server will extract template collaborative control requirement 3 as one of the fourth template collaborative control requirements. Then, the server extracts the first template collaborative control requirement from multiple fourth template collaborative control requirements based on the matching degree between the collaborative control requirement and each fourth template collaborative control requirement. For example, the collaborative control requirement requires precise steering operations on mountain bends and special sensor data processing methods on construction sections. For the fourth template collaborative control requirements extracted previously, such as template collaborative control requirement 3, the server will further compare these requirements in the collaborative control requirements with the specific content in template collaborative control requirement 3. If the matching degree between the steering operation logic and sensor data processing method in template collaborative control requirement 3 and the collaborative control requirement is the highest among these fourth template collaborative control requirements, then template collaborative control requirement 3 will be determined as the first template collaborative control requirement.

[0089] Looking at the second approach, the server extracts multiple fifth template collaborative control requirements from multiple template collaborative control requirements based on the matching degree between the collaborative control requirement and each template collaborative control requirement. Taking intelligent driving testing as an example, a collaborative control requirement might be for intelligent driving operations in specific weather conditions (such as heavy rain), including how sensors accurately acquire information in rainwater and how actuators safely operate on slippery roads based on sensor information. The server compares this collaborative control requirement with each template collaborative control requirement. Template collaborative control requirement 5 primarily addresses driving collaborative requirements on dry roads on sunny days, and has a low matching degree with the collaborative control requirements under current heavy rain conditions. Template collaborative control requirement 7, however, covers some driving situations under specific weather conditions and has a certain relevance to the current collaborative control requirement. Therefore, the server extracts template collaborative control requirement 7 as one of the fifth template collaborative control requirements. Next, the server extracts the first template collaborative control requirement from the multiple fifth template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each fifth template collaborative control requirement. For example, the collaborative knowledge data of a collaborative control requirement might include special detection modes for lidar and special operating logic for the braking system under heavy rain conditions. For collaborative control requirement template 7, the server compares its collaborative knowledge data with the collaborative knowledge data of the current collaborative control requirement. If the degree of match between the relevant knowledge data in collaborative control requirement template 7 and the collaborative knowledge data of the collaborative control requirement is the highest among the fifth collaborative control requirement templates, then collaborative control requirement template 7 is determined as the first collaborative control requirement template.

[0090] Finally, there is the weighted calculation method. The server performs weighted calculations on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the matching degree between the collaborative control requirement and each template collaborative control requirement, generating a second matching degree corresponding to each template collaborative control requirement. Based on the second matching degrees corresponding to multiple template collaborative control requirements, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements. For example, the matching degree weight between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of the template collaborative control requirement is set to 0.6, and the matching degree weight between the collaborative control requirement and the template collaborative control requirement is set to 0.4. For template collaborative control requirement 9, the matching degree between its collaborative knowledge data and the collaborative knowledge data of the collaborative control requirement is calculated to be 0.5, and its matching degree with the collaborative control requirement itself is calculated to be 0.3. Then, the second matching degree is obtained by weighted calculation (0.5×0.6+0.3×0.4=0.42). The server performs such weighted calculations on all template collaborative control requirements, and then extracts the template collaborative control requirement with the highest second matching degree as the first template collaborative control requirement based on the second matching degrees corresponding to these template collaborative control requirements. Through the above different methods, the server can accurately extract the first template collaborative control requirement that best matches the collaborative control requirement of the intelligent driving test equipment from a large number of template collaborative control requirements.

[0091] In a possible implementation, step S110 may further include:

[0092] Based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the collaborative control cost corresponding to each template collaborative control requirement, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements, and the collaborative control cost corresponding to the template collaborative control requirement is used to express the logical node span of the template response knowledge data of the template collaborative control requirement.

[0093] The extracting of the first template collaborative control requirement from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement and the collaborative control cost corresponding to each template collaborative control requirement includes:

[0094] According to the collaborative control cost corresponding to each template collaborative control requirement, multiple sixth template collaborative control requirements are extracted from the multiple template collaborative control requirements. According to the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each sixth template collaborative control requirement, the first template collaborative control requirement is extracted from the multiple sixth template collaborative control requirements, and the collaborative control cost corresponding to the sixth template collaborative control requirement is greater than the collaborative control cost corresponding to the remaining template collaborative control requirements in the multiple template collaborative control requirements. Or,

[0095] Extract multiple fourth template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement. Extract the first template collaborative control requirement from the multiple fourth template collaborative control requirements based on the collaborative control cost corresponding to each fourth template collaborative control requirement, and the collaborative control cost corresponding to the first template collaborative control requirement is greater than the collaborative control cost corresponding to the remaining template collaborative control requirements in the multiple fourth template collaborative control requirements. Or,

[0096] A weighted calculation is performed on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the collaborative control cost corresponding to each template collaborative control requirement to generate a third matching degree corresponding to each template collaborative control requirement. Based on the third matching degrees corresponding to the multiple template collaborative control requirements, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements.

[0097] In this embodiment, the first method is first considered. The server extracts multiple sixth template collaborative control requirements from multiple template collaborative control requirements based on the collaborative control cost corresponding to each template collaborative control requirement. In the intelligent driving scenario, collaborative control requirements are a very complex concept. They cover the collaborative working mode of multiple sensors (lidar, camera, millimeter wave radar, etc.) and actuators (steering system, braking system, acceleration system, etc.) in the intelligent driving test equipment under different road conditions (urban congested roads, rugged mountain roads, highways, etc.), different weather conditions (rain, snow, fog, etc.), and different traffic conditions (peak hours, off-peak hours, traffic accident sections, etc.). Each template collaborative control requirement has its corresponding collaborative control cost, which reflects the logical node span of the template response knowledge data. For example, if a template collaborative control requirement is for intelligent driving under complex and congested road conditions in the city, it may involve more sensor data fusion processing, more traffic rule judgments, and more frequent fine-tuning operations of the actuator, which results in a larger logical node span of its template response knowledge data and a relatively high collaborative control cost.

[0098] The server evaluates the collaborative control cost of each template collaborative control requirement and extracts those with higher collaborative control costs as multiple sixth template collaborative control requirements. For example, let's say that template collaborative control requirement 1 is for normal driving on ordinary highways and has a relatively low collaborative control cost, as the coordinated operation of sensors and actuators in this case is relatively simple and has fewer logical nodes. Meanwhile, template collaborative control requirement 3 is for intelligent driving in congested urban areas during rush hour, involving extensive vehicle distance monitoring, frequent braking and acceleration adjustments, and complex traffic light judgment. Its collaborative control cost is higher and it may be extracted as one of the sixth template collaborative control requirements.

[0099] Next, the server extracts the first template collaborative control requirement from multiple sixth template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each sixth template collaborative control requirement. The collaborative knowledge data of the collaborative control requirement contains various requirements in specific scenarios. For example, on specific mountain roads, the laser radar may be required to focus on monitoring the road conditions at bends, and the camera must accurately identify road signs, etc. For the sixth template collaborative control requirement, such as the template collaborative control requirement 3 mentioned earlier, the server will compare its collaborative knowledge data with the collaborative knowledge data of the collaborative control requirement. If the sensor working mode, actuator operation logic, etc. in the template collaborative control requirement 3 have the highest matching degree with the relevant requirements in the collaborative control requirement among all the sixth template collaborative control requirements, and its collaborative control cost is also high, then the template collaborative control requirement 3 will be determined as the first template collaborative control requirement.

[0100] Looking at the second method, the server extracts multiple fourth template collaborative control requirements from multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement. For example, the collaborative control requirement is intelligent driving in bad weather (such as heavy snow), which requires the sensor to accurately detect the road conditions on the snow-covered road, and the actuator to safely control the vehicle according to the sensor information. The server compares the collaborative knowledge data of this collaborative control requirement with the collaborative knowledge data of each template collaborative control requirement. For template collaborative control requirement 5, its collaborative knowledge data is for intelligent driving in light rain weather, and has a low matching degree with the collaborative knowledge data of the collaborative control requirement in heavy snow weather; while the collaborative knowledge data of template collaborative control requirement 7 involves some driving conditions in bad weather, and has a certain similarity with the current collaborative control requirement. The server will extract template collaborative control requirement 7 as one of the fourth template collaborative control requirements.

[0101] Then, the server extracts the first template collaborative control requirement from multiple fourth template collaborative control requirements based on the collaborative control cost corresponding to each fourth template collaborative control requirement. Taking the fourth template collaborative control requirement extracted previously as an example, assuming that the logical node span of the template response knowledge data of template collaborative control requirement 7 is large when dealing with intelligent driving in bad weather, it means that its collaborative control cost is high. For example, it may require complex fusion processing of data from multiple sensors to cope with the impact of heavy snow on the sensor's field of view. At the same time, the operating logic of the actuator is also relatively complex and requires multiple adjustments according to different road conditions and vehicle states. If, among these fourth template collaborative control requirements, the collaborative control cost of template collaborative control requirement 7 is higher than that of other requirements, and its matching degree with the collaborative knowledge data of the collaborative control requirement is also relatively high, then template collaborative control requirement 7 will be determined as the first template collaborative control requirement.

[0102] Finally, there is the weighted calculation method. The server performs weighted calculation on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the collaborative control cost corresponding to each template collaborative control requirement, to generate a third matching degree corresponding to each template collaborative control requirement. For example, the matching degree weight between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of the template collaborative control requirement is set to 0.7, and the collaborative control cost weight corresponding to each template collaborative control requirement is 0.3. For template collaborative control requirement 9, assuming that the matching degree between its collaborative knowledge data and the collaborative knowledge data of the collaborative control requirement is calculated to be 0.6, and its collaborative control cost is evaluated to be at a high level (assuming the corresponding value is 0.8, the value here is obtained by comprehensive evaluation based on factors such as the logical node span), then the third matching degree is obtained through weighted calculation (0.6×0.7+0.8×0.3=0.66).

[0103] The server performs such a weighted calculation on all template collaborative control requirements, and then extracts the first template collaborative control requirement from multiple template collaborative control requirements based on the third matching degree corresponding to the multiple template collaborative control requirements. That is, the template collaborative control requirement with the highest third matching degree is extracted as the first template collaborative control requirement. In this way, the server comprehensively considers the two important factors of the matching degree of collaborative knowledge data and the collaborative control cost, and can more accurately screen out the first template collaborative control requirement that best meets the collaborative control requirements of intelligent driving test equipment from a large number of template collaborative control requirements, thereby providing more effective support for subsequent intelligent driving test work.

[0104] In a possible implementation, step S120 includes:

[0105] Step S121 , parsing the key elements, constraints and target requirements in the collaborative control requirement to form a structured collaborative control requirement description, wherein the structured collaborative control requirement description includes a set of key elements, a set of constraints and a set of target requirements.

[0106] Step S122, parse the template structure, key element template, constraint template and target requirement template of the first template collaborative control requirement, and generate the parsed first template collaborative control requirement, wherein the parsed first template collaborative control requirement includes a template structure description, a key element template set, a constraint template set and a target requirement template set.

[0107] Step S123 , parsing the response logic, response rules and response parameters in the template response knowledge data to generate parsed template response knowledge data, wherein the parsed template response knowledge data includes a response logic description, a response rule set and a response parameter set.

[0108] Step S124, match the key elements, constraints and target requirements of the collaborative control requirement with the key element template, constraint template and target requirement template of the first template collaborative control requirement, generate a matching degree calculation result and a description of the corresponding relationship between the collaborative control requirement and the first template collaborative control requirement.

[0109] Step S125, based on the matching degree calculation result and the corresponding relationship description, evaluate the adaptability of the template response knowledge data to the collaborative control requirements, and generate an adaptability evaluation result, which includes the response knowledge data part that can be directly applied, the response knowledge data part that needs to be adjusted, and the response knowledge data part that needs to be supplemented.

[0110] Step S126: For the response knowledge data part that needs to be adjusted, it is adjusted according to the collaborative control requirements of the collaborative control needs. For the response knowledge data part that needs to be supplemented, it is supplemented and designed in combination with the demand characteristics of the collaborative control needs and the structure of the template response knowledge data. In addition, the response knowledge data part that can be directly applied, the adjusted response knowledge data and the supplemented response knowledge data are integrated to form an integrated response knowledge data set.

[0111] Step S127: configuring the learning parameters of the deep learning network according to the integrated response knowledge data set and the collaborative control requirements, wherein the learning parameters include a learning rate, a number of iterations, and an optimization algorithm.

[0112] Step S128, using the first template collaborative control requirement and the parsed template response knowledge data as the learning basis, using the configured deep learning network to conduct preliminary learning on the collaborative control requirement, and generate a first guidance result, which includes the deep learning network's preliminary understanding description of the collaborative control requirement, the preliminary set response logic and response parameters.

[0113] Step S130 includes: utilizing the deep learning network to classify the collaborative control requirements according to the first guidance result, generating a collaborative control response category corresponding to the collaborative control requirements, using the first template collaborative control requirements and the template response knowledge data as a learning basis, responding to the collaborative control requirements according to the collaborative control response category, and generating response knowledge data of the collaborative control response category.

[0114] In this embodiment, in the intelligent driving test, the collaborative control requirements cover many complex aspects. For example, in a complex urban traffic scenario, key factors may include the layout and performance parameters of various sensors (such as lidar, cameras, millimeter wave radar, etc.), the characteristics of actuators (such as steering systems, braking systems, acceleration systems, etc.), and road environments (such as whether the road type is a narrow urban street or a main road, the layout of traffic signs and markings, etc.). In terms of constraints, there may be statutory speed limits, physical performance limitations of vehicles (such as maximum steering angle, shortest braking distance, etc.), and working limitations of sensors in different environments (such as the visual range of cameras under low light conditions, etc.). The target requirement is to ensure that the vehicle drives safely, efficiently and legally in this complex environment, such as avoiding collisions with other vehicles and pedestrians, and passing intersections accurately in accordance with traffic rules. Organizing these contents into a set of key elements, a set of constraints and a set of target requirements constitutes a structured description of collaborative control requirements.

[0115] Next, the server parses the first template collaborative control requirement. This first template collaborative control requirement is selected from multiple template collaborative control requirements and is the most compatible with the current collaborative control requirement. Its template structure, key element template, constraint template, and target requirement template are parsed to generate a parsed first template collaborative control requirement. For example, a first template collaborative control requirement for traffic in a specific urban area (such as a commercial district) might have a hierarchical structure, first collecting and preprocessing sensor data, then making decisions based on this data, and finally converting the decision results into actuator operation instructions. The key element template might include information about the common road layout, pedestrian flow characteristics, and vehicle type distribution in the commercial district. The constraint template might include information about speed limits (possibly lower than those on ordinary roads) and turning restrictions specific to the area. The target requirement template ensures smooth vehicle traffic while ensuring the safety of people in the commercial district. This parsing process yields a parsed first template collaborative control requirement, including a template structure description, a set of key element templates, a set of constraint templates, and a set of target requirement templates.

[0116] Subsequently, the server parses the response logic, response rules, and response parameters in the template response knowledge data of the first template collaborative control requirement to generate the parsed template response knowledge data. In this example, the response logic may be based on sensor data (such as the distance of the vehicle in front detected by the lidar, the traffic signs recognized by the camera, etc.) and make judgments in a certain logical order. For example, first determine whether it is close to the intersection, and then determine whether there is a pedestrian crosswalk. The response rules may include performing a braking operation if the distance of the vehicle in front is less than a certain safety threshold, and the braking force is determined according to a certain functional relationship based on the distance. The response parameters are the specific values ​​in these rules, such as the specific distance value of the safety threshold, the coefficient in the function of the relationship between braking force and distance, etc. These contents constitute the parsed template response knowledge data of the response logic description, response rule set, and response parameter set.

[0117] The server then matches the key elements, constraints, and target requirements of the collaborative control requirements with the key element template, constraint template, and target requirement template of the first template collaborative control requirements. For example, the sensor layout and performance parameters in the collaborative control requirements are compared with the relevant content in the key element template of the first template collaborative control requirements. If the lidar detection range in the collaborative control requirements is different from that in the template, these differences are recorded. For constraints, such as the statutory speed limit, compare it with the speed limit in a specific area in the template. In terms of target requirements, compare the degree of fit between safe, efficient, and legal driving and the smooth passage of vehicles under the premise of ensuring the safety of people in commercial areas. Through such a detailed comparison, the matching calculation results and the description of the corresponding relationship between the collaborative control requirements and the collaborative control requirements of the first template are generated.

[0118] Based on the above-mentioned matching degree calculation results and corresponding relationship descriptions, the server evaluates the adaptability of the template response knowledge data to the collaborative control requirements and generates an adaptability evaluation result. If in the template response knowledge data of the first template collaborative control requirement, the response rule of the braking operation is based on a relatively simple speed and distance relationship, and the speed limit in the collaborative control requirement is stricter and the road conditions are more complicated, then this part of the response knowledge data may be the response knowledge data part that needs to be adjusted. If the first template collaborative control requirement does not involve operations under certain special traffic rules in the collaborative control requirement (such as one-way rules in a specific time period), this is the response knowledge data part that needs to be supplemented. And those parts that are completely matched in the two requirements, such as the basic data acquisition logic of the sensor, are the response knowledge data parts that can be directly applied. In this way, an adaptability evaluation result is generated, including the response knowledge data part that can be directly applied, the response knowledge data part that needs to be adjusted, and the response knowledge data part that needs to be supplemented.

[0119] For the response knowledge data parts that need to be adjusted, the server adjusts them according to the collaborative control requirements of the collaborative control needs. For example, for the response rules of braking operations, according to the stricter speed limits and complex road conditions in the collaborative control requirements, the function of the relationship between braking force and distance is redesigned, and the coefficients therein are adjusted to adapt to the new requirements. For the response knowledge data parts that need to be supplemented, a supplementary design is carried out in combination with the demand characteristics of the collaborative control requirements and the structure of the template response knowledge data. For example, for the one-way rule in a specific time period, combined with the existing traffic rule judgment logic in the template, a judgment branch for the one-way street is added, and the operation logic of the actuator in this case is determined. Then the directly applicable response knowledge data parts, the adjusted response knowledge data, and the supplemented response knowledge data are integrated to form an integrated response knowledge data set.

[0120] Based on the integrated response knowledge data set and collaborative control requirements, the server configures the learning parameters of the deep learning network. If the integrated response knowledge data set is more complex and the collaborative control requirements have high real-time requirements, a higher learning rate may be set to speed up the learning speed, but at the same time, in order to avoid overfitting, the number of iterations will be adjusted appropriately. In the selection of optimization algorithms, stochastic gradient descent algorithms may be selected based on the characteristics of the response knowledge data. The configuration of these learning parameters (learning rate, number of iterations, optimization algorithm, etc.) is to enable the deep learning network to better learn the collaborative control requirements based on the first template collaborative control requirements and the parsed template response knowledge data.

[0121] Based on the first template collaborative control requirements and the parsed template response knowledge data, the server uses the configured deep learning network to conduct preliminary learning of the collaborative control requirements and generate a first guidance result. This first guidance result includes the deep learning network's preliminary understanding of the collaborative control requirements, such as the understanding of the relationship between complex traffic rules and sensor actuators, such as the network's understanding of how to coordinate actuator operations based on traffic signs and sensor data at specific intersections. Preliminary response logic, such as the preliminary operation logic set based on current sensor data and traffic rules when encountering a pedestrian crossing and there are pedestrians waiting. And response parameters, such as preliminarily determined actuator operation parameters (such as steering angle, braking force, etc.).

[0122] After generating the first guidance result, the server uses a deep learning network to classify the collaborative control request based on the first guidance result. For example, based on the initial understanding description, response logic, and response parameters in the first guidance result, the collaborative control request is classified as "Normal Driving Collaborative Control in Complex Urban Traffic" or "Special Situation Collaborative Control in Complex Urban Traffic." Using the first template collaborative control request and template response knowledge data as a learning basis, the server responds to the collaborative control request based on the corresponding collaborative control response category, generating response knowledge data for the collaborative control response category. For the "Normal Driving Collaborative Control in Complex Urban Traffic" category, the deep learning network generates response knowledge data for this collaborative control response category based on the sensor data processing method (such as camera and lidar data fusion processing) and actuator operation logic (such as smooth acceleration or deceleration based on traffic flow) during normal driving in the first template collaborative control request, combined with the specific response rules and parameters in the template response knowledge data. This response knowledge data may include more precise sensor data processing algorithms and actuator operation instructions that are more optimized for current traffic conditions. Through such a rigorous series of steps, the server effectively starts from the collaborative control requirements in the intelligent driving test scenario, and after interactive processing with the first template collaborative control requirements and its template response knowledge data, it finally generates response knowledge data that meets the collaborative control requirements, providing an important basis for the accurate control of intelligent driving test equipment.

[0123] Figure 2 The hardware structure of the collaborative control system 100 for intelligent driving test equipment provided by the embodiment of the present invention for implementing the above-mentioned collaborative control method for intelligent driving test equipment is shown as follows: Figure 2 As shown, the collaborative control system 100 for intelligent driving test equipment may include a processor 110 , a machine-readable storage medium 120 , a bus 130 , and a communication unit 140 .

[0124] The machine-readable storage medium 120 may store data and / or instructions. In some embodiments, the machine-readable storage medium 120 may store data obtained from an external terminal. In some embodiments, the machine-readable storage medium 120 may store data and / or instructions for the collaborative control system 100 of the intelligent driving test equipment to execute or use to implement the exemplary methods described herein.

[0125] During the specific implementation process, one or more processors 110 execute computer-executable instructions stored in the machine-readable storage medium 120, so that the processor 110 can execute the collaborative control method for intelligent driving test equipment in the above method embodiment. The processor 110, the machine-readable storage medium 120 and the communication unit 140 are connected through the bus 130, and the processor 110 can be used to control the sending and receiving actions of the communication unit 140.

[0126] The specific implementation process of the processor 110 can be found in the various method embodiments executed by the collaborative control system 100 for intelligent driving test equipment mentioned above. The implementation principles and technical effects are similar and will not be repeated here in this embodiment.

[0127] In addition, an embodiment of the present invention further provides a readable storage medium, in which computer-executable instructions are preset. When a processor executes the computer-executable instructions, the above-mentioned collaborative control method for intelligent driving test equipment is implemented.

[0128] It should be noted that in order to simplify the description of the present invention and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present invention, multiple features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. A collaborative control method for intelligent driving test equipment, characterized in that: The method comprises: Extracting a first template collaborative control requirement from the multiple template collaborative control requirements based on a degree of matching between collaborative knowledge data of the collaborative control requirement of the intelligent driving test equipment and collaborative knowledge data of multiple template collaborative control requirements, wherein the collaborative knowledge data is used to express a requirement knowledge label corresponding to the collaborative control requirement or a knowledge element associated with the corresponding collaborative control requirement, and the matching degree corresponding to the first template collaborative control requirement is greater than the matching degrees corresponding to the remaining template collaborative control requirements in the multiple template collaborative control requirements; Generate a first guidance result based on the collaborative control requirement, the first template collaborative control requirement, and the template response knowledge data of the first template collaborative control requirement, wherein the first guidance result is used to express that the deep learning network responds to the collaborative control requirement based on the first template collaborative control requirement and the template response knowledge data; Using the deep learning network, making a decision on the first guidance result to generate response knowledge data of the collaborative control requirement; Performing a coordinated control operation on the intelligent driving test equipment based on the response knowledge data of the coordinated control requirement, wherein the first guidance result includes a preliminary understanding description of the coordinated control requirement by the deep learning network, and a preliminary set response logic and response parameters; The utilizing the deep learning network to make a decision on the first guidance result and generate response knowledge data of the collaborative control requirement includes: Utilizing the deep learning network, the collaborative control requirements are classified according to the first guidance result, and a collaborative control response category corresponding to the collaborative control requirement is generated. Taking the first template collaborative control requirement and the template response knowledge data as the learning basis, the collaborative control requirement is responded to and fed back according to the collaborative control response category to generate response knowledge data of the collaborative control response category.

2. The collaborative control method for intelligent driving test equipment according to claim 1, characterized in that: The collaborative knowledge data includes first-dimensional collaborative knowledge data and second-dimensional collaborative knowledge data, wherein the first-dimensional collaborative knowledge data is used to express a demand knowledge label corresponding to the collaborative control demand, and the second-dimensional collaborative knowledge data is used to express a knowledge element corresponding to the collaborative control demand; extracting a first template collaborative control demand from the multiple template collaborative control demands based on a matching degree between the collaborative knowledge data of the collaborative control demand of the intelligent driving test equipment and the collaborative knowledge data of the multiple template collaborative control demands includes: Extracting multiple second template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each template collaborative control requirement; extracting the first template collaborative control requirement from the multiple second template collaborative control requirements based on the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each second template collaborative control requirement; or, Extracting multiple third template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each template collaborative control requirement; extracting the first template collaborative control requirement from the multiple third template collaborative control requirements based on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each third template collaborative control requirement; or, A weighted calculation is performed on the matching degree between the first dimension collaborative knowledge data of the collaborative control requirement and the first dimension collaborative knowledge data of each template collaborative control requirement, and the matching degree between the second dimension collaborative knowledge data of the collaborative control requirement and the second dimension collaborative knowledge data of each template collaborative control requirement to generate a first matching degree corresponding to each template collaborative control requirement. Based on the first matching degrees corresponding to the multiple template collaborative control requirements, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements.

3. The collaborative control method for intelligent driving test equipment according to claim 1, characterized in that: Before extracting the first template collaborative control requirement from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement of the intelligent driving test equipment and the collaborative knowledge data of the multiple template collaborative control requirements, the method further includes: Generate a second guidance result according to the collaborative control requirement, where the second guidance result is used to express the data result generated by the deep learning network in response to the collaborative control requirement; The deep learning network is used to make a decision on the second guidance result to generate collaborative knowledge data of the collaborative control requirement.

4. The collaborative control method for intelligent driving test equipment according to claim 1, characterized in that: Before extracting the first template collaborative control requirement from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement of the intelligent driving test equipment and the collaborative knowledge data of the multiple template collaborative control requirements, the method further includes: Acquire the plurality of template collaborative control requirements and candidate response knowledge data for each template collaborative control requirement; Utilizing the deep learning network, the candidate response knowledge data of each template collaborative control requirement is adapted and optimized according to each template collaborative control requirement, and the template response knowledge data of each template collaborative control requirement is generated.

5. The collaborative control method for intelligent driving test equipment according to claim 4, characterized in that: Before utilizing the deep learning network to adaptively optimize candidate response knowledge data for each template collaborative control requirement based on each template collaborative control requirement and generating template response knowledge data for each template collaborative control requirement, the method further includes: generating a third guidance result based on the template collaborative control requirement and the candidate response knowledge data of the template collaborative control requirement, wherein the third guidance result is used to express that the deep learning network performs adaptive optimization on the candidate response knowledge data of the template collaborative control requirement based on the template collaborative control requirement; The utilizing the deep learning network to adapt and optimize the candidate response knowledge data of each template collaborative control requirement according to each template collaborative control requirement to generate the template response knowledge data of each template collaborative control requirement includes: Using the deep learning network, making a decision on the third guidance result to generate iterative response knowledge data of the template collaborative control requirement; When the iterative response knowledge data of the template collaborative control requirement and the candidate response knowledge data of the template collaborative control requirement express the same response result, the iterative response knowledge data of the template collaborative control requirement is output as the template response knowledge data of the template collaborative control requirement.

6. The collaborative control method for intelligent driving test equipment according to claim 4, characterized in that: The multiple template collaborative control requirements and the template response knowledge data of each template collaborative control requirement are stored in a pre-built sample library; after using the deep learning network to adapt and optimize the candidate response knowledge data of each template collaborative control requirement according to each template collaborative control requirement and generate the template response knowledge data of each template collaborative control requirement, the method further includes: Generate a fourth guidance result based on any template collaborative control requirement and the template response knowledge data of the template collaborative control requirement, wherein the fourth guidance result is used to express that the deep learning network uses the template collaborative control requirement and the template response knowledge data of the template collaborative control requirement as a learning basis to generate iterative collaborative control requirements and response knowledge data; Using the deep learning network, making a decision on the fourth guidance result to generate an iterative template collaborative control requirement and template response knowledge data of the iterative template collaborative control requirement; The iterative template collaborative control requirement and the template response knowledge data of the iterative template collaborative control requirement are added to the sample library.

7. The collaborative control method for intelligent driving test equipment according to claim 1, characterized in that: The extracting a first template collaborative control requirement from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement of the intelligent driving test equipment and the collaborative knowledge data of the multiple template collaborative control requirements includes: Extracting multiple fourth template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement; extracting the first template collaborative control requirement from the multiple fourth template collaborative control requirements based on the matching degree between the collaborative control requirement and each fourth template collaborative control requirement; or, Extracting multiple fifth template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the collaborative control requirement and the collaborative control requirement of each template; extracting the first template collaborative control requirement from the multiple fifth template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each fifth template collaborative control requirement; or, A weighted calculation is performed on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the matching degree between the collaborative control requirement and each template collaborative control requirement to generate a second matching degree corresponding to each template collaborative control requirement. Based on the second matching degrees corresponding to the multiple template collaborative control requirements, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements.

8. The collaborative control method for intelligent driving test equipment according to claim 1, characterized in that: The extracting a first template collaborative control requirement from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement of the intelligent driving test equipment and the collaborative knowledge data of the multiple template collaborative control requirements includes: Extracting the first template collaborative control requirement from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the collaborative control cost corresponding to each template collaborative control requirement, wherein the collaborative control cost corresponding to the template collaborative control requirement is used to express the logical node span of the template response knowledge data of the template collaborative control requirement; The extracting of the first template collaborative control requirement from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement and the collaborative control cost corresponding to each template collaborative control requirement includes: Extract multiple sixth template collaborative control requirements from the multiple template collaborative control requirements based on the collaborative control cost corresponding to each template collaborative control requirement; extract the first template collaborative control requirement from the multiple sixth template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each sixth template collaborative control requirement, and the collaborative control cost corresponding to the sixth template collaborative control requirement is greater than the collaborative control cost corresponding to the remaining template collaborative control requirements in the multiple template collaborative control requirements; or Extracting multiple fourth template collaborative control requirements from the multiple template collaborative control requirements based on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement; extracting the first template collaborative control requirement from the multiple fourth template collaborative control requirements based on the collaborative control cost corresponding to each fourth template collaborative control requirement, the collaborative control cost corresponding to the first template collaborative control requirement being greater than the collaborative control cost corresponding to the remaining template collaborative control requirements in the multiple fourth template collaborative control requirements; or A weighted calculation is performed on the matching degree between the collaborative knowledge data of the collaborative control requirement and the collaborative knowledge data of each template collaborative control requirement, and the collaborative control cost corresponding to each template collaborative control requirement to generate a third matching degree corresponding to each template collaborative control requirement. Based on the third matching degrees corresponding to the multiple template collaborative control requirements, the first template collaborative control requirement is extracted from the multiple template collaborative control requirements.

9. The collaborative control method for intelligent driving test equipment according to claim 1, characterized in that: The step of generating a first guidance result based on the collaborative control requirement, the first template collaborative control requirement, and the template response knowledge data of the first template collaborative control requirement includes: Analyzing the key elements, constraints, and target requirements in the collaborative control requirement to form a structured collaborative control requirement description, wherein the structured collaborative control requirement description includes a set of key elements, a set of constraints, and a set of target requirements; Parsing the template structure, key element template, constraint condition template, and target requirement template of the first template collaborative control requirement to generate a parsed first template collaborative control requirement, wherein the parsed first template collaborative control requirement includes a template structure description, a key element template set, a constraint condition template set, and a target requirement template set; Parsing the response logic, response rules, and response parameters in the template response knowledge data to generate parsed template response knowledge data, wherein the parsed template response knowledge data includes a response logic description, a response rule set, and a response parameter set; Matching the key elements, constraints, and target requirements of the collaborative control requirement with the key element template, constraint template, and target requirement template of the first template collaborative control requirement, generating a matching degree calculation result and a description of the corresponding relationship between the collaborative control requirement and the first template collaborative control requirement; Based on the matching degree calculation result and the corresponding relationship description, the adaptability of the template response knowledge data to the collaborative control requirements is evaluated to generate an adaptability evaluation result, wherein the adaptability evaluation result includes a portion of the response knowledge data that can be directly applied, a portion of the response knowledge data that needs to be adjusted, and a portion of the response knowledge data that needs to be supplemented; For the response knowledge data portion that needs to be adjusted, the adjustment is performed according to the collaborative control requirement of the collaborative control demand; for the response knowledge data portion that needs to be supplemented, a supplementary design is performed in combination with the demand characteristics of the collaborative control demand and the structure of the template response knowledge data; and the directly applicable response knowledge data portion, the adjusted response knowledge data, and the supplemented response knowledge data are integrated to form an integrated response knowledge data set; Configuring learning parameters of a deep learning network according to the integrated response knowledge data set and the collaborative control requirements, wherein the learning parameters include a learning rate, a number of iterations, and an optimization algorithm; Taking the first template collaborative control requirement and the parsed template response knowledge data as the learning basis, the configured deep learning network is used to perform preliminary learning on the collaborative control requirement to generate a first guidance result.

10. A collaborative control system for intelligent driving test equipment, characterized in that: The collaborative control system for intelligent driving test equipment includes a processor and a memory, the memory is connected to the processor, the memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to implement the collaborative control method for intelligent driving test equipment described in any one of claims 1 to 9 above.

Images