To address the issue of development requirement of trusted embedded systems， a trusted embedded hardware platform is designed， which is comprised of an full indigenous Root of Trust （RoT）， Loongson processors， Kunlun firmware and SylixOS domestic technology stack. On this basis， firstly， guided by an active defense philosophy， a dynamic runtime trusted verification mechanism is researched and designed for software code. Then， this mechanism performs real-time verification of software identity and code integrity both at program startup and execution， thereby preventing the execution of unauthorized programs or tampered code. Finally， a prototype of the trusted embedded platform is established and the proposed designs are validated. The experimental results demonstrate that the trustworthiness of equipment software can be effectively enhanced by using the proposed methodology.
An universal program parsing and test coverage analysis technology is introduced for multi-architecture embedded processors， including program parsing for compilers of different architectures and universal data storage structure， a kind of fast program information retrieval algorithm and a universal test coverage analysis architecture. By taking multiple embedded software onboard system of the new-generation launch vehicle as validation targets， the verification of each technical aspect is implemented. The validation results show that the proposed method is compatible with multiple compiler standards and processor architectures， and the provided universal coverage test framework can meet the requirements of mainstream embedded processors. The proposed method is extensible and can be extended to applications to the software that has new target processor architectures in the future.
Software factories can serve as key platforms for the efficient development of aerospace software. The dynamic interaction and real-time update of their frontend components have direct influence on development efficiency and resource utilization. Traditional fixed-time-interval refresh mechanisms are liable to cause resource waste or data latency issues， while existing semi-automatic adjustment methods struggle to adapt to complex scenarios in software factories， such as multiple component types， cross-network deployment and multi-environment deployment. An adaptive refresh algorithm is proposed for frontend components. The change values can be obtained by component differential comparison， and a CVI （Component volatility index）is established， which integrates short-term change trends with long-term average levels to quantify the volatility of business data， and scenario-specific adjustment strategies are designed. This algorithm can adapt to scenarios like multi-team collaboration and high-frequency component reuse， which effectively balances resource consumption and response efficiency and thus provides support for the efficient operation of software factories.
With the continuous applications expansion in the aerospace domain， the health management of complex equipment is oriented from scheduled maintenance towards data-driven predictive maintenance. A data-driven anomaly detection framework known as TS-ADF is proposed， which achieves effective identification of potential anomalies through the establishment of normal patterns， reconstructive analysis and feature fusion of multidimensional operational data. Specifically， preliminary screening is involved by using density peak clustering， deep features of time series are captured through LSTM-AE， and anomaly points are validated via time-frequency analysis and parameter variation analysis. Experimental results demonstrate the method's effectiveness in anomaly detection， which can serve intelligent health management and predictive maintenance of equipment.
Aiming at the problems of over-segmentation and text adhesion in natural scene text detection algorithms based on image segmentation， a natural scene text detection algorithm based on a cross-level attention mechanism is proposed. By designing a cross-level attention module， the network's focus on key features and contextual information in high-resolution feature maps is enhanced by applying the proposed algorithm that thereby improves the integration capability for fragmented text. Through the design of a feature decomposition and reorganization module， the fused features are decomposed into high-frequency and low-frequency components that enhance the network's ability of distinguishing text boundary regions. After integrating these two modules into the baseline model， performance tests are implemented on two mainstream datasets. The experimental results show those of current mainstream algorithms are all surpassed， and compared to the baseline model， the missed detection rate and false alarm rate are both declined.
In order to standardize usage of static test tools， combining the automation with platform-based characteristics of tools， a static testing multi-tool collaborative analysis framework supporting distributed deployment and parallel analysis is designed； A remote driving technology based on a distributed architecture is proposed， which establishes a test task queue mechanism and tool scheduling mechanism and thereby enables programmatic control of remotely driven tool analysis； A multi-source information fusion technique is developed to achieve normalization， which realize filtering and integration of analysis results from multiple tools. By applying these technologies to static testing across various aerospace domains， testing efficiency and quality are effectively improved.
Excessive axle temperature of vehicle can lead to a series of issues that affect driving safety and is an important index for evaluating the environmental adaptability of special vehicles. An equipment axle temperature prediction method based on a thermal equilibrium model is proposed. A differential equation describing the thermal variation of vehicle axles is firstly formulated. Parameters of the equation are calibrated by utilizing experimentally acquired data， and the equation structure is further optimized. Consequently， axle temperature predictions with a margin of error below 3°C are achieved， thereby the challenging problem is to forecast axle temperature fluctuations in special vehicles， which is solved under diverse environmental conditions.
To address issue of the inefficiency in collaborative design of launch vehicles flight sequences， a model-based systems engineering （MBSE）based methodology is proposed for modeling launch vehicles flight sequence and the collaborative design mechanism is investigated for such flight sequence within an MBSE framework. By focusing on a representative two-stage configuration and non-booster based launch vehicles， the organizational relationships are analyzed among distinct flight phases throughout the vehicle's life-cycle and a hierarchical architecture global scheme is presented for flight sequence modeling. For each hierarchy level， based on the structural composition of the sequence command skeleton， a meta-modeling approach is developed to establish elementary models of flight sequence command signals. On the basis of these elementary models， a comprehensive MBSE model based on the entire procedure of launch vehicles flight sequence is designed. Furthermore， oriented to the collaborative design strategy for flight sequence in the MBSE framework， the model distribution and merging mechanisms based cooperative design of flight time sequence is implemented. This work lays a foundation for the full-scale digital transformation of launch vehicles design processes.
To address the issue of low test coverage caused by implicit test scenarios such as missing boundary conditions and exception handling in software requirement documents， a code-enhanced requirement analysis method is proposed. The code-requirement associations are established through the proposed method which is based on semantic vector similarity and LLM verification， and five types of implicit test scenarios （boundary conditions， error handling， resource management， state transitions and performance stress） are extracted from function call chains to enhance requirement descriptions， and the enhanced requirements are decomposed into test function points and scenario-driven test cases are generated. The results of experiments on open-source projects show that compared with the baseline method by using LLM directly， the significant improvements are achieved by using CERA method in comprehensive test quality and test requirement coverage， which maintains higher API test accuracy. The effectiveness of three core components： scenario extraction， two-stage matching strategy and BERT-based rough screening is verified through ablation experiments. The good adaptability on both parsing libraries and embedded systems is demonstrated by the results of proposed method applied that is particularly suitable for third-party testing and acceptance testing scenarios.
An intelligent forecasting model based on ensemble learning， named iTransformer-XGBoost， is proposed in this paper to address the issue of precision limitations of traditional time series prediction models. Firstly， the Pearson correlation coefficient is employed to select the key features affecting time series data and construct an optimized input dataset. Then， the iTransformer model is utilized to capture long-term dependencies within the time series and generate preliminary prediction results. Meanwhile， the XGBoost algorithm is introduced to achieve nonlinear modeling of the time series data. Finally， a threshold-based combination strategy is applied to the prediction results fusion of iTransformer and XGBoost， thereby determining the integrated output and improving overall forecasting performance. The model is validated by using photovoltaic power-related data， and the experimental results demonstrate that the proposed approach achieves higher accuracy and stability in time series forecasting compared with traditional methods. Furthermore， it shows great potential for application in aerospace photovoltaic scenarios.
A specification and formal verification method is presented for dynamic failure modes in aerospace launch systems. The dynamic fault model designed through this approach enables real-time fault diagnosis systems in the aerospace field to possess diagnostic capabilities for complex and diverse fault characteristics. The dynamic fault models are represented by temporal facts， however， the complexity and abstract nature of temporal facts make them difficult to be validated through manual analysis or testing. By describing the semantics of temporal symptoms through temporal logic formulas， the correctness of the fault model is allowed to be verified through model checking methods. This verification process can be automated by using model checking tools. Moreover， the dynamic fault model does not require formal specifications， which allows domain experts to focus solely on domain-specific issues when constructing fault models. In aerospace engineering practice， the key properties of fault models through automated verification of model checking methods can provide domain experts with a reliable and verifiable mathematical approach for designing fault models.
To address the issue of guidance information extraction for strapdown phased array seekers with radome errors， a guidance information extraction method based on Kalman filtering is investigated. Firstly， a relative motion model is established in the body line-of-sight coordinate system， incorporating the radome error slope into the model， regarding the maneuvering targets， then the Singer maneuver model is employed， the radome error slope measured offline is treated as a known quantity， and observability analysis is implemented for the system of angle measurement observations undertaken， providing a theoretical basis for guidance information extraction in the presence of radome errors. Next， according to the isolation rising caused by gyroscope and beam control delays， a gyroscope delay compensation method is adopted to reduce the impact of airframe attitude disturbances. Finally， comparative simulations on isolation with and without gyroscope delay compensation for gyroscope and beam control delays are performed along with simulation validation in a high-altitude interception scenario against maneuvering targets. The simulation results demonstrate that the missile's isolation performance is improved after radome error compensation and a smaller miss distance is achieved in high-altitude interception of maneuvering targets.
