CN-121995930-A - Unmanned ship multi-stage task reliability assessment method based on given confidence

CN121995930ACN 121995930 ACN121995930 ACN 121995930ACN-121995930-A

Abstract

The invention provides an unmanned ship multi-stage task reliability assessment method based on given confidence, which relates to the technical field of unmanned ship reliability analysis and comprises the following steps of S1, defining system composition equipment, fault distribution and series-parallel connection relation; the method comprises the steps of S2, multi-level data fusion, S3, calculating the confidence coefficient of a single-stage task according to the target confidence coefficient corresponding to the target reliability of the multi-stage task, S4, setting multi-stage task reliability aggregation under the confidence coefficient of the single-stage task, and S5, judging whether the reliability evaluation result of the multi-stage task of the unmanned ship meets the reliability requirement under the target confidence coefficient. The invention provides a unmanned ship multi-layer data fusion method, which realizes unmanned ship task reliability verification method based on the combination of a small amount of task reliability test data and performance function verification, and provides a task reliability calculation method of an unmanned ship under set confidence level based on task reliability assessment results of various stages of the unmanned ship through a multi-stage task reliability aggregation technology.

Inventors

ZHANG YIMING
LIANG ZHAOLEI
HOU CHENGUANG
XU YUNTIAN
Zhao Daiyi
TIAN JIEWEI
Qi Huizhi
CAO DALI

Assignees

中国航空综合技术研究所

Dates

Publication Date: 20260508
Application Date: 20251229

Claims (9)

1. The unmanned ship multi-stage task reliability assessment method based on the given confidence is characterized by comprising the following steps of: S1, defining system composition equipment, fault distribution and serial-parallel connection relations, namely dividing an unmanned ship system into a power system, an autonomous navigation system, a control system, an identification system and a communication system, and determining the fault distribution and the serial-parallel connection relations of unmanned ship equipment according to the system and equipment composition of the unmanned ship and a stage task reliability model; S2, multi-level data fusion, namely classifying fault distribution and serial-parallel connection relation of unmanned ship equipment, and fusing working time and fault data of the equipment to a system level according to classification results to obtain working time and fault data of the system; s3, determining the confidence coefficient of the single-stage task, namely determining the confidence coefficient required for checking the reliability of the single-stage task according to the set target confidence coefficient and the target task reliability of the multi-stage task check: calculating confidence in reliability of multi-stage task : ; Wherein: ; ; ; in the formula, M is a fixed value selected according to the numerical calculation precision requirement, M is the number of multi-stage tasks; Representing the ith stage task, in Confidence value under precision operator; Confidence in multi-stage task reliability The confidence coefficient is equal to the set target confidence coefficient, and single-stage task confidence coefficient is obtained through calculation; S4, setting multi-stage task reliability aggregation under the single-stage task confidence, namely according to the single-stage task confidence set in S3, combining the classification result in S2, the working time of the system and fault data, and calculating to obtain the reliability confidence lower limit of each stage task; And S5, judging whether the evaluation result meets the reliability requirement under the target confidence coefficient, namely comparing the evaluation result obtained in the S4 with the set target task reliability, and if the evaluation result reaches the target task reliability threshold, checking the unmanned ship through the task reliability.
2. The unmanned aerial vehicle multi-stage task reliability assessment method based on the given confidence level, as set forth in claim 1, wherein in S2, classifying the fault distribution and the series-parallel relationship of the unmanned aerial vehicle equipment comprises: The serial system data fusion method is composed of different index distribution devices; the serial system data fusion method consists of the same index distribution equipment; the parallel system data fusion method consists of the same index distribution equipment; The K/N system data fusion method consists of the same index distribution equipment.
3. The unmanned ship multi-stage task reliability assessment method based on the given confidence level according to claim 2, wherein in S4, the serial system task reliability lower limit calculation method composed of different index distribution devices comprises the following steps: When the constituent devices of the system fail: ; ; The sub-confidence is The lower confidence limit of the time task reliability is as follows: ; When the constituent devices of the system do not fail: ; Confidence is of The lower confidence limit of the time task reliability is as follows: ; in the formula, Is normally distributed Dividing the number of bits; indicating that the system comprises the number of devices; representing the equivalent running time of each unit device; Indicating the failure number of each unit device: representing the system test time after fusion; Representing the failure times of the fused system; Representing a time of a task; Representing the equivalent number of tasks.
4. The unmanned ship multi-stage task reliability assessment method based on the given confidence level according to claim 2, wherein in S4, the serial system task reliability lower limit calculation method composed of the same index distribution equipment comprises the following steps: when the constituent equipment of the system fails, the sub-confidence is that The lower confidence limit of the time task reliability is as follows: ; Wherein, the ; In the formula, Is of degree of freedom of And confidence is as follows A kind of electronic device A quantile of the distribution; when the component equipment of the system fails, the sub-confidence is that The lower confidence limit of the time task reliability is as follows: ; in the formula, Indicating the test time of the device; Representing the test time of the fused system; Representing the failure times of the fused system; representing the confidence level; Representing a time of a task; Representing the equivalent task number; Representing the equivalent task number after fusion; Indicating the number of task failures.
5. The unmanned ship multi-stage task reliability assessment method based on the given confidence level according to claim 2, wherein in S4, the parallel system task reliability lower limit calculation method composed of the same index distribution equipment comprises the following steps: Confidence is then The lower confidence limit of the time task reliability is as follows: ; Wherein, the ; In the middle of Is of degree of freedom of And confidence is as follows A kind of electronic device A quantile of the distribution; When the component equipment of the system fails, the confidence is that The lower confidence limit of the time task reliability is as follows: ; in the formula, Indicating that the system comprises the number of devices; indicating the total test time of the device; Representing the test time of the fused system; Representing the failure times of the fused system; representing the confidence level; Representing a time of a task; Representing the equivalent task number; And representing the equivalent task number after fusion.
6. The unmanned ship multi-stage task reliability assessment method based on the given confidence level according to claim 2, wherein in S4, the calculation method of the K/N system task reliability lower limit composed of the same index distribution equipment comprises the following steps: When the constituent devices of the system fail: System MTBF sub-confidence level is Lower confidence limit for time: ; When the constituent devices of the system do not fail: The sub-confidence is The lower confidence limit of the time task reliability is as follows: ; Wherein X is the number of devices which can work normally; representing the number of devices contained in the system; representing a minimum number of working devices; Representing the total test time of the system; Representing the test time of the fused system; Representing the failure times of the fused system; Representing a time of a task; Representing the equivalent task number; Indicating the confidence level.
7. The method for evaluating the reliability of an unmanned ship multi-stage task based on a given confidence level according to claim 1, wherein in S4, when the multi-stage task is composed of m stage tasks which are distributed identically and independently, the i-th stage task has a confidence level of The confidence lower limit of the reliability is that Multi-stage task at confidence Lower task reliability confidence limit The method comprises the following steps: 。
8. The unmanned ship multi-stage task reliability assessment method based on given confidence level as set forth in claim 1, wherein in S4, when the multi-stage task is composed of a plurality of single-stage tasks which are not completely identical in distribution and are independent of each other, each single-stage task is classified into a plurality of single-stage tasks with different distributions The lower confidence limit of the task reliability is Multi-stage task at confidence Lower task reliability confidence limit The method comprises the following steps: 。
9. The unmanned aerial vehicle multistage task reliability assessment method based on a given confidence level of any of claims 1-8, wherein S4 further comprises calculating a lower confidence limit for the first stage task reliability at the determined confidence level by the confidence level of the single stage task determined in S3: ; in the formula, A confidence lower limit for the first stage task reliability; For the total time of the equivalent trial of the first phase task, For the number of equivalent failures, Is of the degree of freedom of The determined sub-confidence is When (1) Distribution values.

Description

Unmanned ship multi-stage task reliability assessment method based on given confidence Technical Field The invention relates to the technical field of unmanned ship reliability analysis, in particular to an unmanned ship multi-stage task reliability assessment method based on given confidence. Background Task reliability refers to the ability of a product to perform a specified function within a specified task profile (i.e., the full cycle of task execution), which evaluation includes not only overall system stability, but also the guarantee of task continuity by redundant design. The reliability verification of the task in the equipment identification process mainly comprises the steps of determining the number of serious faults through statistics analysis of working time and number of faults of the equipment in the identification process, determining the number of the serious faults based on the definition of the serious faults, calculating the confidence lower limit of average serious fault interval time under the given confidence, and then calculating the confidence lower limit of the task reliability of the equipment in the given task time. However, in practical engineering application, because of the restrictions of development expense and development period, a large amount of task reliability tests cannot be carried out, and the objectivity and the authenticity of the task reliability verification result obtained under the condition of insufficient sample size are insufficient, so that the task reliability verification is carried out by combining a small amount of task reliability test data with working time and fault data during performance function verification, and data acquired by equipment during performance function verification are equipment-level data, so that how to fuse the equipment-level data to obtain second-level system data and further obtain first-level system data (equipment-level data) is a technical problem of the current equipment task reliability verification. In the unmanned ship field, the unmanned ship is used as new research equipment, is restricted by development expense and development period, and is greatly limited in development of the task reliability test. The unmanned ship has the advantages that the unmanned ship only relates to a propulsion and power system, a communication system and the like in the sailing and sailing stage in the execution of the unmanned ship task, the difficulty in checking the sailing task is small, available test data are sufficient, the unmanned ship sailing task comprises the checking of situation awareness capability, autonomous route planning capability, autonomous obstacle avoidance capability and the like, the practical test time is short, the available test data are very limited, and the difficulty is caused to the task reliability verification of the unmanned ship. On the other hand, the task section of the unmanned ship is usually composed of tasks in multiple stages, such as autonomous sailing, regional monitoring, collaborative operation and autonomous sailing, equipment in each stage has strong dependence (such as a navigation system and communication relay equipment), is influenced by offshore environments (storms and salt mist), has unknown equipment fault distribution, and is difficult to apply by the traditional task reliability verification method. The system components required by each stage of task in execution are different, and only the reliability of the task at a certain stage of equipment can be evaluated according to the test data of the equipment. For the related theoretical research of multi-stage task reliability evaluation at present, a plurality of scholars aim at the characteristics of unmanned and intelligent system structure level, complex logic composition and multi-stage task, the state dependence problem of shared equipment among stages is solved based on a Markov process, and the multi-stage task evaluation is realized by using methods such as Dynamic Fault Tree (DFT), bayesian network, generalized random Petri network (GSPN), monte Carlo simulation and the like. However, these theoretical methods have limitations in that, on one hand, the basic assumption that the evaluation by the current method is performed is that the fault distribution of the system component equipment is known, but in practical engineering application, the distribution of important component equipment of the offshore intelligent equipment is generally unknown, the above method is theoretically complete but is often not suitable for evaluating the task reliability of the current intelligent system, on the other hand, when the reliability of each stage task is calculated to be the same as the reliability of the multi-stage task, the reliability of each stage task is generally considered to be integrated to obtain the reliability of the multi-stage task, but the calculation method ignores the influence caused by simply integrating the ta