CN-122019383-A - Intelligent cabin testing method and device, electronic equipment and readable storage medium

CN122019383ACN 122019383 ACN122019383 ACN 122019383ACN-122019383-A

Abstract

The invention provides a testing method and device of an intelligent cabin, electronic equipment and a computer readable storage medium, and belongs to the technical field of intelligent cabins; the method comprises the steps of obtaining a test case aiming at an intelligent cabin, wherein the test case comprises a non-standard instruction, executing the test case in the intelligent cabin, obtaining a response result of the intelligent cabin to the test case, obtaining quantitative scores of the intelligent cabin in different evaluation dimensions based on the response result, and obtaining an evaluation result of the intelligent cabin based on the quantitative scores of the different evaluation dimensions. The method provided by the invention can at least solve the problems that the current testing method leads to correct functions of the intelligent cabin and poor experience, and finally influences the user experience of the intelligent cabin.

Inventors

CHENG QIAN
LI JI
WANG CHENYANG
PENG JIANG
TANG RUYI

Assignees

重庆赛力斯凤凰智创科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260128

Claims (10)

1. A method for testing an intelligent cockpit, the method comprising: acquiring a test case aiming at an intelligent cabin, wherein the test case comprises a non-standard instruction; Executing the test case in the intelligent cabin; acquiring a response result of the intelligent cabin to the test case; Based on the response result, obtaining quantitative scores of the intelligent cabin in different evaluation dimensions; and acquiring an evaluation result of the intelligent cabin based on the quantitative scores of the different evaluation dimensions.
2. The method of claim 1, wherein the non-standard instructions comprise at least one of semantically ambiguous or semantically contradictory instructions, formally conflicting instructions, and continuously varying instruction sequences.
3. The method of claim 2, wherein the obtaining test cases for the intelligent cockpit comprises: decomposing the text of the standard instruction into semantic elements of different types, wherein the semantic elements comprise objects of the standard instruction and limiting conditions of the standard instruction; Modifying original semantic elements based on a rule base or adding additional semantic elements based on the rule base, wherein the rule base comprises fuzzy rules, logic contradiction rules and complex negation rules, the fuzzy rules are used for replacing objects of the standard instructions with fuzzy pronouns, the logic contradiction rules are used for introducing first limiting conditions contradicting the text semantics of the standard instructions into the text of the standard instructions, and the complex negation rules are used for introducing second limiting conditions representing multiple negations, partial negations or range limitations into the text of the standard instructions; And recombining the semantic elements to generate the command of semantic ambiguity or semantic contradiction.
4. The method of claim 2, wherein the conflicting instructions comprise a first conflicting instruction and a second conflicting instruction, and wherein the obtaining test cases for the intelligent cockpit comprises: The method comprises the steps of generating a first conflict instruction based on a preset target table, wherein the target table is used for recording a corresponding relation between semantic types and emotion tones, the first conflict instruction is provided with a first emotion label for recording the emotion tones, the emotion tones recorded by the first emotion label are not matched with the semantic types of the first conflict instruction, and the first emotion label is used for playing texts of the first conflict instruction by using the emotion tones recorded by the first emotion label when the test case is executed; Or generating the second conflict instruction based on a preset position coordinate library, wherein the position coordinate library is used for storing the coordinates of the interactable position in the intelligent shelter, the second conflict instruction is provided with an action label, the coordinates of the target position of the action recorded by the action label are different from the position coordinates of the object of the second conflict instruction, and the action label is used for executing the action recorded by the action label when the test case is executed.
5. The method of claim 2, wherein the obtaining test cases for the intelligent cockpit comprises: modifying at least one of the object, the action and the parameter of the initial instruction to obtain a target instruction; Taking the target instruction as a new initial instruction, entering at least one of an object, an action and a parameter for modifying the initial instruction, and obtaining a target instruction; And combining the time sequence of the initial instruction and the target instruction to be used as the continuously-changing instruction sequence, wherein the target instruction is executed in a preset time period after the initial instruction.
6. The method of claim 1, wherein the evaluation dimensions include intent resolution accuracy, multi-form arbitration rationality, interaction policy effectiveness, emotion consistency, software pressure tolerance, and wherein the obtaining the quantitative scores of the intelligent capsule in different evaluation dimensions based on the response results comprises: determining a quantization score of the intention analysis accuracy based on a preset standard response result, wherein the intention analysis accuracy is used for representing the understanding accuracy degree of the intelligent cabin to the core intention of the user instruction; Determining a quantization score of the multi-form arbitration rationality based on a preset priority rule, wherein the multi-form arbitration rationality is used for representing the rationality of decision making of the intelligent cabin under the condition of inputting different forms of conflicting instructions; determining a quantization score of the interaction strategy effectiveness based on the context and risk level of the instruction in the test case, wherein the interaction strategy effectiveness is used for representing the effectiveness degree of the interaction strategy adopted by the intelligent cabin to the user; Determining a quantization score of emotion consistency based on a second emotion label of the response result, wherein the emotion consistency is used for representing the matching degree of the response result of the intelligent cabin and an interaction scene or user emotion; And determining a quantitative score of the software pressure tolerance based on the response delay and the response error rate of the intelligent cabin, wherein the software pressure tolerance is used for representing the performance attenuation degree of the intelligent cabin under the condition of receiving a continuously-changing instruction sequence.
7. The method of claim 1, wherein different ones of the evaluation dimensions have respective weights, wherein the obtaining the intelligent cockpit evaluation result based on the quantitative scores of the different ones of the evaluation dimensions comprises: taking the weighted sum of the evaluation dimensions as the overall score of the intelligent cockpit; and outputting an evaluation result of the intelligent cabin based on the overall score.
8. A test device for an intelligent cockpit, said device comprising: The system comprises a case acquisition module, a control module and a control module, wherein the case acquisition module is used for acquiring a test case aiming at an intelligent cabin, and the test case comprises a non-standard instruction; The case execution module is used for executing the test case in the intelligent seat cabin; the response acquisition module is used for acquiring a response result of the intelligent cabin to the test case; The quantization score module is used for acquiring quantization scores of the intelligent cabin in different evaluation dimensions based on the response result; and the evaluation analysis module is used for acquiring an evaluation result of the intelligent cabin based on the quantitative scores of the different evaluation dimensions.
9. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus; a memory for storing a computer program; A processor for implementing the steps in the method for testing a smart capsule according to any one of claims 1 to 7 when executing a program stored on a memory.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the intelligent cockpit method according to any one of claims 1 to 7.

Description

Intelligent cabin testing method and device, electronic equipment and readable storage medium Technical Field The present invention relates to the technical field of intelligent cabins, and in particular, to a method and apparatus for testing an intelligent cabin, an electronic device, and a computer readable storage medium. Background The intelligent cabin is an integrated system for realizing intelligent interaction between a user and a carrier, and along with the improvement of complexity and integration degree of the system, the requirements on the reliability and the stability of the intelligent cabin are higher and higher, and the automatic test of the intelligent cabin is more and more focused by manufacturers. In the related art, the automatic test methods mainly focus on verification of functional correctness of the intelligent cabin, such as recognition accuracy of voice instructions by the intelligent cabin, success rate of executing the instructions by the intelligent cabin and the like, and in the automatic test methods, users issuing the instructions are considered rational, so that the test cases are generally clear and unambiguous instructions. However, in real scenes, the user may be irrational, the instruction given by the user may be ambiguous or even contradictory, and the test cases and the matched evaluating means in the related art are difficult to cover the scenes, which easily results in correct functions of the intelligent cabin but poor experience, and influences the user experience of the final intelligent cabin. Disclosure of Invention In view of the above, the present invention provides a method, apparatus, electronic device and storage medium for testing an intelligent cabin, so as to at least solve the problem that the current testing method results in correct functions of the intelligent cabin but poor experience, and finally affects the user experience of the intelligent cabin. In order to achieve the above purpose, the technical scheme of the invention is realized as follows: The invention provides in a first aspect a method for testing an intelligent cabin, the method comprising: acquiring a test case aiming at an intelligent cabin, wherein the test case comprises a non-standard instruction; Executing the test case in the intelligent cabin; acquiring a response result of the intelligent cabin to the test case; Based on the response result, obtaining quantitative scores of the intelligent cabin in different evaluation dimensions; and acquiring an evaluation result of the intelligent cabin based on the quantitative scores of the different evaluation dimensions. Optionally, the non-standard instructions include at least one of semantically ambiguous or semantically contradictory instructions, form conflicting instructions, and continuously varying instruction sequences. Optionally, the acquiring the test case for the intelligent cabin includes: decomposing the text of the standard instruction into semantic elements of different types, wherein the semantic elements comprise objects of the standard instruction and limiting conditions of the standard instruction; Modifying original semantic elements based on a rule base or adding additional semantic elements based on the rule base, wherein the rule base comprises fuzzy rules, logic contradiction rules and complex negation rules, the fuzzy rules are used for replacing objects of the standard instructions with fuzzy pronouns, the logic contradiction rules are used for introducing first limiting conditions contradicting the text semantics of the standard instructions into the text of the standard instructions, and the complex negation rules are used for introducing second limiting conditions representing multiple negations, partial negations or range limitations into the text of the standard instructions; And recombining the semantic elements to generate the command of semantic ambiguity or semantic contradiction. Optionally, the instructions for the form conflict comprise a first conflict instruction and a second conflict instruction, and the steps of obtaining the test case for the intelligent cabin comprise the following steps: The method comprises the steps of generating a first conflict instruction based on a preset target table, wherein the target table is used for recording a corresponding relation between semantic types and emotion tones, the first conflict instruction is provided with a first emotion label for recording the emotion tones, the emotion tones recorded by the first emotion label are not matched with the semantic types of the first conflict instruction, and the first emotion label is used for playing texts of the first conflict instruction by using the emotion tones recorded by the first emotion label when the test case is executed; Or generating the second conflict instruction based on a preset position coordinate library, wherein the position coordinate library is used for storing the coordinates of the inte