汽车嵌入式系统服务与数据融合中间件研究

Computing technology has brought great changes to automotive electronics. Under specific environment for heterogeneous computing structure, heterogeneous networks, heterogeneous data sources embedded automotive electronic systems, this project provides service-oriented data access and information communication for dynamic vehicle applications, and an automotive embedded middleware was implemented. For automotive embedded systems, we propose a service-oriented data integration middleware and software architecture, giving the flexible service delivery mechanism with characteristics of real-time, distributed and loosely coupled. We focus on embedded middleware services and virtualized resource scheduling algorithm, data fusion algorithm, as well as device-level service access calling for message exchanging format and realization based on OSGi. A formal definition method of semantic association and context-aware was proposed to build a scalable, configurable embedded service component. By solving the problems such as dynamic construction and adaptation mechanism of embedded system service, as well as multi-source and multi-modal data fusion, we implement the opening and management to device-level service capabilities and improve the efficiency of the data acquisition and data transmission to meet changing application needs. Finally, we build the test platform based on embedded middleware, implement the embedded service support environment, and software architecture, services and data fusion mechanism were applied to the smart car system.

计算技术给汽车电子带来了巨大的变化，本项目在汽车环境下，针对异构计算结构、异构网络、异构数据来源的嵌入式汽车电子系统，为多样化动态的车辆应用提供面向服务的数据接入与消息交换，实现汽车嵌入式中间件。针对汽车嵌入式系统，提出面向服务的数据融合中间件软件体系结构，给出具有实时性、分布式和松散耦合等特性的灵活服务提供机制。重点研究嵌入式中间件服务组合与虚拟化资源调度算法、数据融合算法、以及设备级服务访问调用的消息交换格式和基于OSGi的实现方法。提出语义关联和上下文感知的形式化定义方法，构建可扩展、可配置的嵌入式服务组件。通过解决嵌入式系统服务的动态构建与适配机理以及多源多模态数据融合等问题，实现对ECU服务能力的开放和管理，提高数据采集和数据传输的效率，为安全行车决策。最后，搭建基于嵌入式中间件测试平台，实现嵌入式服务支撑环境，并将所提出的软件体系架构、服务和数据融合机制应用到智能车系统中。

汽车电子系统是一类复杂的嵌入式和网络系统，迫切的需要一种跨平台、多传感网接入、多源数据融合的信息交换和控制平台。本项目提出车载环境下面向数据处理的嵌入式中间件软件体系架构和服务计算模型，从车载应用模式的角度出发，首先建立车载应用服务模型，形成汽车嵌入式服务组件，给出具有实时性、分布式和松散耦合等特性的灵活服务提供机制。其次，在车联网环境下，设计实现汽车嵌入式服务中间件。包括两部分内容：服务组合与虚拟化资源分配和多传感器数据交换与融合。在此基础上，设计了分布式服务目录选择算法、服务共享机制和服务发现协议，实现车辆间服务共享，提高服务发现成功率和服务质量。最后实现了基于OSGi的服务推荐中间件及基于上下文车联网位置隐私保护中间件，使用分布式OSGi中间件实现多源多模态数据融合、服务推荐和隐私保护，为行车决策服务，同时为用户推荐合理且具个性化的服务。. 本项目培养博士研究生1 名，硕士研究生8名，在国内外核心期刊发表论文20篇(SCI检索11篇，EI检索9篇)，获得专利授权7项，软件著作权3项。