BACKGROUND Combined hepatocellular-cholangiocarcinoma (cHCC-CCA) is a rare and aggressive primary liver cancer characterized by the presence of both hepatocellular and cholangiocellular differentiation within the same tumor. Its diagnostic complexity and low incidence have resulted in a scarcity of well-established prognostic pathological parameters to guide clinical management. Given the rising global prevalence of metabolic dysfunction-associated steatotic liver disease and alcohol-related liver disease, and their close association with the occurrence of primary liver cancer, it is plausible that hepatic steatosis influences the prognosis of cHCC-CCA. AIM To investigate the specific clinicopathological role of background hepatic steatosis in patients with cHCC-CCA who underwent curative-intent liver resection. METHODS This multicenter study analyzed 310 patients with cHCC-CCA who underwent curative-intent hepatectomy between 2013 and 2017. Hepatic steatosis in liver tissue was graded pathologically: Grade 0-1 (< 33% hepatocytes) was defined as negligible hepatic steatosis (n = 283, 91.3%), and grade 2-3 (>= 33% hepatocytes) as severe hepatic steatosis (n = 27, 8.7%). Primary endpoints were recurrence-free survival (RFS), overall survival, early (<= 2 years) and late (> 2 years) RFS. RESULTS Patients with severe hepatic steatosis had a significantly prolonged RFS compared to those with negligible steatosis (P = 0.035), with 5-year RFS rates of 28.8% vs 15.7%. This benefit was primarily observed in early RFS (P = 0.024) and was more pronounced in non-cirrhotic patients. In contrast, no statistically significant difference was found in overall survival (P = 0.586) or late RFS (P = 0.931) between the two groups. Multivariate logistic regression analysis showed higher body mass index and lymphocyte count, and lower lactate dehydrogenase were independent predictors for the presence of severe steatosis. Furthermore, multivariate Cox regression analysis confirmed severe hepatic steatosis as an independent protective factor for RFS (hazard ratio = 0.614) and early RFS (hazard ratio = 0.577). CONCLUSION Severe hepatic steatosis is a protective factor associated with RFS and early RFS in patients with cHCC-CCA who underwent hepatectomy, which is especially significant in the absence of liver cirrhosis.

Background:Combined hepatocellular-cholangiocarcinoma (cHCC-CCA) is a rare primary liver cancer lacking personalized pathological parameters. Microscopic tumor focus (MTF), a microscopic intrahepatic metastatic pattern, remains poorly characterized, with unclear prognostic and therapeutic relevance in cHCC-CCA. In this study, the detailed significance of MTF in cHCC-CCA was investigated. Methods:A multicenter study included 310 cHCC-CCA patients who underwent curative-intent hepatectomy. MTF was further graded by its number/distribution (MTF 0: absence of MTF; MTF 1: 1-5 proximal MTFs; MTF 2: >5 proximal MTFs/≥1 distal MTFs) and subclassified by its histology (HCC-type MTF; CCA-type MTF; HCC + CCA-type MTF). Results:MTF was detected in 48.1% (149/310) of patients (MTF 1: 43.2%, MTF 2: 4.8%; HCC-type MTF: 17.7%; CCA-type MTF: 20.6%; HCC + CCA type MTF: 9.7%). MTF positivity correlated with poorer recurrence-free survival (RFS) (median time: 0.23 and 0.89 years), early RFS, and overall survival (OS) (median time: 1.53 and 2.95 years, all P<0.001), with MTF 2 and HCC + CCA-type MTF predicting the worst prognosis. For MTF-positive patients, surgical margin >0.1 cm improved RFS, early RFS, and OS, while surgical margin had no effect on MTF-negative patients. Age, tumor diameter, microvascular invasion and perineural invasion independently predicted MTF (area under the curve =0.704); MTF histology was an independent prognostic factor for RFS, early RFS, and OS (all Harrell's concordance index >0.7). Conclusions:This study standardized the definition of MTF and confirmed its prognostic value in cHCC-CCA. MTF stratification based on its number/distribution and histology enables more precise prognostic assessment. The presence of MTF could guide individualized selection of surgical margin.

BACKGROUND:The pathological diagnostic purpose of the tumor component proportion of combined hepatocellular-cholangiocarcinoma (CHC) remains a perplexing issue, hindering clinicians from characterizing its biological behavior and formulating intervention strategies. METHODS:A multicenter database of patients with CHC who underwent therapeutic hepatectomy was analyzed. Based on a > 70% hepatocellular carcinoma (HCC) or cholangiocarcinoma (CCA) component, the patients were stratified into three groups: CHCHCC-dominant, CHCbalanced, and CHCCCA-dominant. Survival outcomes were assessed using Kaplan-Meier analysis. Tumor markers and recurrence patterns were compared via Chi-square test, and prognostic factors were identified through Cox regression analysis. Postoperative adjuvant transarterial chemoembolization (PA-TACE) benefits were further evaluated. RESULTS:Among 19 641 patients with primary liver cancer, 306 with CHC were classified: 93 (30.4%) in the CHCHCC-dominant group, 101 (33.0%) in the CHCbalanced group, and 112 (36.6%) in the CHCCCA-dominant group. The best and the worst prognoses were observed in the CHCHCC-dominant group and CHCCCA-dominant group, respectively (median RFS time: CHCHCC-dominant group 1.02 years, CHCbalanced group 0.36 years, CHCCCA-dominant group 0.30 years; median OS time: CHCHCC-dominant group 3.16 years, CHCbalanced group 1.93 years, CHCCCA-dominant group 1.86 years). The highest ratios of elevated alpha-fetoprotein and carbohydrate antigen 19-9 were recorded in the CHCHCC-dominant group and the CHCCCA-dominant group, respectively. CHCHCC-dominant and CHCCCA-dominant patients showed the most frequent intrahepatic and extrahepatic recurrences, respectively. PA-TACE improved early RFS (< 2 years) only in CHCHCC-dominant patients. Multivariate analysis confirmed that tumor component proportion was an independent prognostic factor for all survival endpoints. CONCLUSIONS:Predominant HCC or CCA components (> 70% tumor composition) critically shape CHC biology and prognosis. Tumor component proportion emerges as a novel parameter for therapeutic decision-making. We proposed a comprehensive flow chart integrating pathological sampling, histological component evaluation, clinical data mapping, and postoperative intervention strategies for CHC.

BACKGROUND:The prognostic role of perineural invasion (PNI) in combined hepatocellular-cholangiocarcinoma (cHCC-CCA) remains undefined. This multicenter study assessed the incidence, patterns, and clinical significance of PNI in cHCC-CCA. METHODS:This study included 307 patients with cHCC-CCA undergoing hepatectomy with curative intent. Three independent pathologists conducted histopathological assessments of the frequency (single/multiple), involved tumor components (HCC/CCA/intermediate cell carcinoma), and spatial distribution (intratumoral/intratumoral + peritumoral/peritumoral) of PNI. The impact of postoperative adjuvant transarterial chemoembolization (PA-TACE) on the prognosis was further evaluated. Survival outcomes (recurrence-free survival [RFS], overall survival [OS], and early [≤2 years] and late [>2 years] RFS) were analyzed using Kaplan-Meier and Cox regression. Logistic regression identified the predictors of PNI. RESULTS:Among 307 patients (median age, 52.7 years; 87.6% male), the prevalence of PNI was 16.0% (49 of 307). PNI-positive patients had significantly worse median RFS (0.21 vs 0.54 years; hazard ratio [HR], 2.003; P <.001) and OS (1.18 vs 2.56 years; HR, 2.213; P <.001) than PNI-negative patients. Early RFS differed significantly (P <.001), but late RFS did not (P =.443). Subgroup analysis showed that the worst RFS and early RFS were noted in patients with HCC-component PNI (P =.025; P =.024) and purely intratumoral PNI (P =.025; P =.024). PA-TACE improved the early RFS only in PNI-negative patients (P =.031). Elevated alkaline phosphatase (odds ratio [OR], 1.005; P =.029), macrovascular invasion (OR, 2.873; P =.003), microvascular invasion (OR, 2.291; P =.033), and absent tumor capsule (OR, 2.539; P =.014) independently predicted PNI. Multivariable analysis confirmed PNI as an independent risk factor for RFS (HR, 1.700; P =.002), OS (HR, 1.760; P =.001), and early RFS (HR, 1.749; P =.001). CONCLUSION:PNI is a noteworthy pathological feature in cHCC-CCA, independently associated with early recurrence, inferior survival, and increased therapeutic challenge. Standardized pathological reporting of PNI status may improve prognostic stratification for patients with cHCC-CCA.

Background Combined hepatocellular-cholangiocarcinoma (cHCC-CCA) lacks standardized pathological diagnostic paradigm. The significance of tumor necrosis in cHCC-CCA remains undefined. This study aimed to decipher its role and develop a practical grading system. Methods A multicenter analysis was conducted on 307 cHCC-CCA patients. Tumor necrosis was pathologically graded as: TN1 (no necrosis, N = 35), TN2 (focal necrosis; maximum diameter of a single focus ≤0.55 cm/one low-power field, N = 129), and TN3 (extensive necrosis; >0.55 cm, N = 143). Associations between tumor necrosis and clinicopathological features, recurrence-free survival (RFS), overall survival (OS), early RFS (≤2 years), and late RFS (>2 years) were assessed. The impact of postoperative adjuvant transarterial chemoembolization (PA-TACE) was also evaluated. Results Higher necrosis grades (TN2/TN3) were significantly associated with aggressive tumor characteristics, including larger tumor size, vascular invasion, and lymph node metastasis. TN3 patients had the worst median RFS (0.28 years) and OS (1.36 years), compared to TN2 (RFS: 0.68 years; OS: 2.64 years) and TN1 (RFS: 1.46 years; OS: not reached). The grading system was an independent prognostic factor for RFS, OS, and early RFS in multivariate analysis. PA-TACE significantly improved early RFS in the TN2/TN3 groups and RFS in the TN3 group specifically, but not in the TN1 group. Conclusions Tumor necrosis in cHCC-CCA indicates aggressive tumor biology and poorer outcomes. The proposed three-tier grading system provides robust prognostic stratification. PA-TACE benefits patients with significant necrosis (TN2/TN3), particularly in preventing early recurrence. Standardized assessment of tumor necrosis is recommended to optimize risk stratification and guide adjuvant therapy decisions for cHCC-CCA patients.

Purpose:Tertiary lymphoid structure (TLS) has been well-established across multiple tumor types for predicting efficacy of immunotherapy and prognostic evaluation. However, its role in combined hepatocellular-cholangiocarcinoma (cHCC-CCA) remains unclear. Refinement of TLS pathological assessment could potentially optimize postoperative management in these patients. This study aimed to develop a practical histopathological grading system of intratumoral TLS to improve prognostic stratification of cHCC-CCA patients. Patients and Methods:A cohort of 310 cHCC-CCA patients undergoing hepatectomy with curative intent was analyzed. Three pathologists re-evaluated pathological slides to establish a four-tier TLS grading system: TLS 0 (absent), TLS 1 (immature TLS only), TLS 2a [single mature TLS (mTLS)], and TLS 2b (multiple mTLS). Associations with recurrence-free survival (RFS), overall survival (OS), early RFS (≤1 year), late RFS (>1 year), and recurrence patterns were assessed. Predictive factors for TLS were also investigated. Results:Patients were stratified into TLS 0 (29.4%), TLS 1 (51.6%), TLS 2a (6.8%), and TLS 2b (12.3%). Survival outcomes significantly correlated with TLS presence and maturation. Median RFS increased stepwise: 0.24 years (TLS 0), 0.49 years (TLS 1), 1.13 years (TLS 2a), and 1.16 years (TLS 2b) (P<0.001). Median OS also improved progressively: 1.32 years (TLS 0), 2.20 years (TLS 1), 3.24 years (TLS 2a), and 10.04 years (TLS 2b) (P<0.001). TLS presence was associated with increased extrahepatic recurrence. The TLS grading system emerged as an independent prognostic factor for RFS, OS, and early RFS. Smaller tumor diameter was the sole significant predictive factor for both TLS and mTLS. Conclusion:This novel TLS grading system effectively stratifies prognosis in cHCC-CCA, with increasing intratumoral mTLS indicating better outcomes. This practical method can be integrated into routine pathological reporting to aid clinical decision-making.

Current testing methods for autonomous driving systems primarily focus on simple traffic scenarios, generating test cases based on traffic accidents, while research on generating edge test cases for complex driving environments by traffic regulations is not adequately comprehensive. Therefore, we propose a method for scenario modeling and violation testing using an autonomous driving system based on traffic regulations named TraModeAVTest. Initially, TraModeAVTest constructs a Petri net model for complex scenarios based on the combination relationships of basic traffic regulation scenarios and verifies the consistency of the model’s design with traffic regulation requirements using formal methods, to provide a representation of traffic regulation scenario models for the violation testing of autonomous driving systems. Subsequently, based on the coverage criteria of the Petri net model, it utilizes a search strategy to generate model paths that represent traffic regulations, and employs a parameter combination method to generate test cases that cover the model paths, to test the violation behaviors of autonomous driving systems. Finally, simulation experiment results on the Baidu Apollo demonstrate that the test cases representing traffic regulations generated by TraModeAVTest can effectively identify the behaviors of autonomous vehicles violating traffic regulations, and TraModeAVTest can effectively improve the efficiency of generating different types of violation scenarios.

Recent advances in artificial intelligence technology and perception components have promoted the rapid development of autonomous vehicles. However, as safety‐critical software, autonomous driving systems often make wrong judgments, seriously threatening human and property safety. LiDAR is one of the most critical sensors in autonomous vehicles, capable of accurately perceiving the three‐dimensional information of the environment. Nevertheless, the high cost of manually collecting and labeling point cloud data leads to a dearth of testing methods for LiDAR‐based perception modules. To bridge the critical gap, we introduce MetaLiDAR, a novel automated metamorphic testing methodology for LiDAR‐based autonomous driving systems. First, we propose three object‐level metamorphic relations for the domain characteristics of autonomous driving systems. Next, we design three transformation modules so that MetaLiDAR can generate natural‐looking follow‐up point clouds. Finally, we define corresponding evaluation metrics based on metamorphic relations. MetaLiDAR automatically determines whether source and follow‐up test cases meet the metamorphic relations based on the evaluation metrics. Our empirical research on five state‐of‐the‐art LiDAR‐based object detection models shows that MetaLiDAR can not only generate natural‐looking test point clouds to detect 181,547 inconsistent behaviors of different models but also significantly enhance the robustness of models by retraining with synthetic point clouds.

Context: Deep neural networks (DNN) have been widely deployed in safety -critical domains, such as autonomous cars and healthcare, where error behaviors can lead to serious accidents, testing DNN is extremely important. Neuron coverage -guided fuzz testing (NCFT) has become an effective whitebox testing approach for testing DNN, which iteratively generates new test cases with the guidance of neuron coverage to explore different logics of DNN, and has found numerous defects. However, existing NCFT approaches ignore that the role of neurons is distinct for the final output of DNN. Given an input, only a fraction of neurons determines the final output of the DNN. These neurons hold the essential logic of the DNN. Objective: To ensure the quality of DNN and improve testing efficiency, NCFT should first cover neurons containing major logic of DNN. Method: In this paper, we propose the critical neurons that hold essential logic of DNN. In order to prioritize the detection of potential defects of critical neurons, we propose a fuzz testing framework, named CriticalFuzz, which mainly contains the energy -based test case generation and the critical neuron coverage criteria. The energy -based test case generation has the capability to produce test cases that are more likely to cover critical neurons and involves energy -based seed selection, power schedule, and seed mutation. The critical neuron coverage as a mechanism for providing feedback to guide the CriticalFuzz in prioritizing the coverage of critical neurons. To evaluate the significance of critical neurons and the performance of CriticalFuzz, we conducted experiments on popular DNNs and datasets. Results: The experiment results show that (1) the critical neurons have a 100% impact on the output of models, while the non -critical neurons have a lesser effect; (2) CriticalFuzz is effective in achieving 100% coverage of critical neurons and covering 10 classes of critical neurons, outperforming both DeepHunter and TensorFuzz. (3) CriticalFuzz exhibits exceptional error detection capabilities, successfully identifying thousands of errors across 10 diverse error classes within DNN. Conclusion: The critical neurons defined in this paper hold more significant logic of DNN than non -critical neurons. CriticalFuzz can preferentially cover critical neurons, thereby improving the efficiency of the NCFT process. Additionally, CriticalFuzz is capable of identifying a greater number of errors, thus enhancing the reliability and effectiveness of the NCFT.

With the rapid development of autonomous driving technology, autonomous vehicles have emerged as a prominent application of artificial intelligence. They play a vital role in addressing critical issues such as environmental pollution, energy consumption, traffic congestion, and traffic accidents, gradually progressing toward practical implementation. However, given that autonomous driving systems are complex software operating under uncertain conditions, they can be prone to making erroneous judgments in intricate external environments, which may lead to traffic accidents. Therefore, it is essential to conduct comprehensive testing of autonomous driving systems before deployment on public roads to ensure their safety and reliability. This paper discusses the key technologies involved in testing existing autonomous driving systems from two perspectives: simulation testing techniques and test case generation techniques.

The development of artificial intelligence and information communication technology has significantly propelled advancements in autonomous driving. The advent of autonomous driving has a profound impact on societal development and transportation methods. However, as intelligent systems, autonomous driving systems (ADSs) often make wrong judgements in specific scenarios, resulting in accidents. There is an urgent need for comprehensive testing and validation of ADSs. Metamorphic testing (MT) techniques have demonstrated effectiveness in testing ADSs. Nevertheless, existing testing methods primarily encompass relatively simple metamorphic relations (MRs) that only verify ADSs from a single perspective. To ensure the safety of ADSs, it is essential to consider the various elements of driving scenarios during the testing process. Therefore, this paper proposes MetaSem, a novel metamorphic testing method based on semantic information of autonomous driving scenes. Based on semantic information of the autonomous driving scenes and traffic regulations, we design 11 MRs targeting different scenario elements. Three transformation modules are developed to execute addition, deletion and replacement operations on various scene elements within the images. Finally, corresponding evaluation metrics are defined based on MRs. MetaSem automatically discovers inconsistent behaviours according to the evaluation metrics. Our empirical study on three advanced and popular autonomous driving models demonstrates that MetaSem not only efficiently generates visually natural and realistic scene images but also detects 11,787 inconsistent behaviours on three driving models. To test whether ADSs can accurately understand and respond to complex driving scenes, we proposed MetaSem, a novel metamorphic testing method based on semantic information of autonomous driving scenes. Based on real scene elements, we design MRs and conduct testing according to the elements' semantic information and corresponding traffic rules. image

随着人工智能技术的快速发展,自动驾驶系统应需而生,并逐渐成为人工智能技术的典型应用.自动驾驶系统集合人工智能、计算机视觉、雷达和定位系统等组件,是一个较为复杂的智能软件系统.在该文中,首先,对自动驾驶系统、自动驾驶场景和自动驾驶测试进行简要介绍.其次,调研数十篇国内外文献,对自动驾驶系统各个模块和组件的测试方法,以及测试用例生成方法的研究现状进行总结、阐述.最后,对自动驾驶这一领域未来面对的问题及挑战进行展望.

软件测试是各种软件投入使用之前的重要环节,由于软件测试内容比较多,如何在有限的时间内有效完成所有测试项目需要进行重点研究.在信息化、数字化技术飞速发展中,软件自动化测试可解决上述问题,还能够减轻软件维护工作压力.要想满足多种类型软件测试要求,提高相关测试程序的复用性,可以设计一种基于数据驱动的软件自动化测试框架,对软件产品、数据和系统进行全面测试,应保证软件自动化测试框架的自动化和独立性,方便维护与管理,同时也要能够根据不同项目特点对相关框架进行改进与拓展,快速完成软件项目测试.本文主要对基于数据驱动的软

在高等院校教育改革不断深入的背景下，大量高素质应用型人才被输送到社会之中，这就使得人才竞争日渐激烈，为了能够帮助更多毕业生顺利就业，在教学中教师应该积极寻求新的教学方法与教学理念，并能够将理论教学与实践教学相结合，从整体上提升学生的综合素养和专业能力。本文基于OBE理念之下，对C语言程序设计教学创新与实践进行几点分析总结。

With the development of artificial intelligence and other fields, the application range of automatic driving system is gradually expanding. Currently, most of the researches on automatic driving system are based on the scene of intersections, and there is a lack of researches based on the scene of high-speed sections.Because the highway is one of the special sections on the road where vehicles travel and the probability of accidents is highest, based on this, this paper proposes a test data generation method for automatic driving system in high-speed scenario. Petri net modeling is carried out on selected highway scenes, and adaptive genetic algorithm is applied to generate test data, which aims to make the test coverage of automatic driving system more comprehensive and more secure application.

自动驾驶汽车在缓解交通拥堵和消除交通事故方面发挥着重要作用.为了保证自动驾驶系统的安全性和可靠性,在自动驾驶汽车部署到公共道路之前,必须进行全面的测试.现有的测试场景数据大多来源于交通事故和交通违法场景,而且自动驾驶系统最基本的安全需求就是遵守交通法规,这充分体现了自动驾驶汽车遵守交通规则的重要性.然而,目前严重缺少针对交通法规构建的自动驾驶测试场景.因此,从交通法规出发,根据自动驾驶系统的安全需求,提出了交叉路口测试场景的Petri 网建模及形式化验证方法.首先,依据自动驾驶测试场景对交规进行分类,提取适合自动驾驶汽车的文本交规,并进行半形式化表征;其次,以覆盖道路交通安全法规以及测试场景功能测试规程为目标,融合交叉路口场景要素的交互行为,合理选择并组合测试场景要素,布设交叉路口测试场景;然后,基于交规的测试场景被建模为一个 Petri 网,其中,库所描述自动驾驶汽车的状态,变迁表示状态的触发条件,并选择时钟约束规范语言(CCSL)作为中间语义语言,将Petri网转换为一个可进行形式化验证的中间语义模型,提出了具体的转换方法;最后,通过Tina软件分析验证交规场景模型的活性、有界性和可达性,结果表明了所建模型的正确性,并基于SMT的分析工具MyCCSL来分析CCSL约束,采用LTL公式以形式化方法验证交规场景模型的一致性.

Autonomous vehicles play an important role in alleviating traffic congestion and eliminating traffic accidents. However, as safety-critical software, they must undergo testing before being deployed on public roads. Due to its scalability and repeatability, simulation testing has become a prominent research hotspots in autonomous driving testing techniques. In response to the issues of high similarity among generated scenarios and limited discovery of violation types in existing scenario testing methods, this paper proposed a method of generating autonomous driving safety violation scenarios based on multi-objective optimization (A V _ MOVS). The aim is to utilize a multi-objective genetic algorithm to guide the evolutionary search direction, thereby discovering a wider range of safety violation scenarios for autonomous driving systems and improving software testing efficiency. The simulation experiment results on the Baidu platform Apollo demonstrate that the efficiency of generating safety violation scenarios using the method proposed in this paper has been improved by 16.7% compared to other methods.

In recent years, deep neural networks (DNNs) have made great progress in people’s daily life since it becomes easier for data accessing and labeling. However, DNN has been proven to behave uncertainly, especially when facing small perturbations in their input data, which becomes a limitation for its application in self-driving and other safety-critical fields. Those human-made attacks like adversarial attacks would cause extremely serious consequences. In this work, we design and evaluate a safety testing method for DNNs based on mutation testing, and propose an adversarial training method based on testing results and joint optimization. First, we conduct an adversarial mutation on the test datasets and measure the performance of models in response to the adversarial samples by mutation scores. Next, we evaluate the validity of mutation scores as a quantitative indicator of safety by comparing DNN models and their updated versions. Finally, we construct a joint optimization problem with safety scores for adversarial training, thus improving the safety of the model as well as the generalizability of the defense capability.

With the rapid development of the information industry, intelligent software testing has become one of the hot research. This paper studies how to extract useful data from the original test set to test the modified model for the intelligent software. First, the initial test data are obtained according to the label classification of the original structured documents. Second, the non-dominated sorting genetic algorithm is used to prioritize the test data. Finally, the experimental results of intelligent software show that compared with the random method, the accuracy of the penalty_money prediction model is reduced by 0.04%~ 34.91 %, and the accuracy of the penalty_imprisonment prediction model is reduced by 0.08% ~ 32.63 %, indicating that the proposed method can find more defects of the penalty prediction model earlier, to reduce the cost of intelligent software testing.

Regression testing is the most costly testing method. The purpose of test data augmentation method is to make full use of the original test cases to augment the test data with higher path coverage and reduce the cost of regression testing. K-means algorithm is used to cluster the test data, and the proximity was measured according to Hamming distance, so that the test data assigned to the same cluster have higher path similarity. Evaluate the intra-cluster data according to the statement coverage information, select the better data to form the initial population, use the correlation between the test data crossing path and the target path to design the fitness function, and evolve to generate the augmented data. Experimental results show that compared with similar methods, the proposed method improves the path coverage by 2.19 ~ 6.62%, and the fault detection rate by 2.84 ~ 6.58%.