table of Contents
1. Overview ... 2
2. Mission and goals ... 3
3. System Framework ... 4
4. The container apparatus (iNeuKernel) ... 4
The third party data import interface 9 ...
6. View Modeling (iNeuView) ... 10
7. Machine Learning (iNeuAI) ... 11
8. Call for demand ... 13
1 Overview
iNeuOS of Things cloud operating system (i: information and industry representatives, Neu: representatives of neurons and cells, OS: represents the operating system), which is our name from the time the idea. Published by the end of May " iNeuOS cloud operating system, .NET Core build the whole system " the netizen questioned: just copy the Windows 10 user interface, which can also be called the operating system. I can only say that there is a deviation understanding of things, did not do in the past explanation.
After a lapse of three months, we released version 2.0 iNeuOS information equipment container, view modeling and machine learning three major functional components. Container equipment mainly to solve a variety of multi-protocol communication mechanism device or sensor data access issues, support secondary development; view modeling is mainly to solve the problem that is the access device or sensor data is displayed, you can publish to the desktop applications, similar to the early ; thinking Plug and Play WindowsXP propaganda machine learning to train the model to solve the main device or sensor data access, derives from the knowledge and rules.
iNeuOS all use .NET Core development, can be deployed to the embedded device, thereby forming an edge gateway computing products, only we need to develop more device drivers running in the device container.
For now speaking, this is the general idea of our construction system. What is a good fellow human phase methods and ideas, you can always communicate with us.
Demo: enter iNeuOS system . (Recommended chrome browser)
Test Name: admin
Test password: admin
Note: When the last release, as more people log into your system, you feedback is slow, we rented a cloud server configuration does not too high, please understand! ! !
Login screen, as shown below:
Internal system interface, as shown:
2. Mission and objectives
Now everyone in the industry call big data and the Internet industry, but it can really fall much, but some enterprises to embark on information technology past the old slogan above the essence; now we are engaged in a cloud of millions of businesses, from hardware intensive point of view, can indeed reduce costs, but limited spatial location data and service changes, obviously reflects little effect or meaning; of course there are doing well, such as Ali of things (IOT) piece, systematic normalized to solve the problem of access devices, including rules and display views, iNeuOS seamless data docking Ali IOT.
For the industry concerned, and not all companies are willing to cloud, regardless of the form in the form of basic resources or access to things. For SMEs build a private cloud platform, facing the high cost of operation and maintenance and inadequate capacity situation. Based on the above, iNeuOS main mission is to solve the enterprise information lack of capacity, cost too high.
3. System Framework
iNeuOS now comprises three functional components: a container device (iNeuKernel), modeling view (iNeuView), machine learning (iNeuAI). A schematic framework, as follows:
FIG iNeuOS a schematic framework
4. The container apparatus (iNeuKernel)
Equipment container (iNeuKernel) is iNeuOS kernel operating equipment container, the frame is a physical device or sensor core build, mount a device driver can freely run in a container, the overall coordinating device driver (protocol), IO channel (COM and NET) , coordination mechanism between the operating mechanism (mode), so that seamless integration, operation and support of open secondary development. As shown below:
Service management, service instances in a plurality of device drivers may be run, as shown below:
Fig service management list
Plans to increase service instance
设备管理，一个设备驱动下可以采集多个数据点的信息，如下图：
图 设备管理列表
图 增加设备驱动
数据点管理，如下图：
图 设备驱动数据点列表
图 增加数据点
（1）iNeuKernel的特点
- 支持与实体设备之间多种模式交互，包括：轮询模式、自控模式、并发模式和单例模式。
- 支持协议驱动器，可以按规范开发标准协议和自定义协议。
- 支持发送数据缓存器，支持命令缓存重发和按优先级别发送。
- 支持协议过滤器，按规则筛选数据；并且可以承继接口，自定义过滤方式。
- 支持接收数据缓存器，可以缓存不符合过滤器的数据，和下次接收数据进行拼接。
- 支持按设备命令优先级别进行调度设备，保证有高级别命令的驱动及时发送。
- 支持一个设备驱动，同时支持串口和网络两种通讯方式，可以监视IO通道数据。
- 支持一个设备驱动，在网络通讯时可以支持TCP Server和TCP Client两种工作模式。
- 支持服务组件接口，例如：4-20mA输出、LED大屏显示、短信服务、以及多功能网关服务。
- 设备驱动与设备驱动，设备驱动与服务器（云端）可以实时双向交互，上传数据和指令下发。
- 支持创建多服务实例，完成不同业务的拆分。
- 支持WebSocket，进行数据采集与转发。
- 支持多设备通讯IO通道多路复用机制。
- 支持定时清理无效IO通道，并释放资源。
- 支持跨平台部署，可以运行在Linux和Windows系统。
（2）iNeuKernel解决的现实问题
物联网平台都会有自己的接入数据标准协议，但是每个平台的数据协议又不一样，所以并没有从根本上解决协议标准的问题。所以物联网、大数据、工业互联网的数据集成过程中仍然面临四个主要困难：设备多样性、协议多样性、通讯机制多样性、数据多样性。
5. 第三方数据导入接口
系统提供HTTP Post接口服务，方便导入其他系统的数据，以便用于视图建模（iNeuView）和机器学习（iNeuAI）功能组件服务。
导入配置数据接口形式，如下：
[{
"servername": "servername",
"devices": [{
"devicename": "mydevice",
"tags": [{
"tagname": "tagname"
}]
}]
}]
导入实时数据接口形式，如下：
[{
"timestamp": "2019-08-06T11:33:18.6983926+08:00",
"tagname": "tag_2_2",
"tagvalue": 2.2
}]
6. 视图建模（iNeuView）
视图建模（iNeuView）是iNeuOS视图建模功能组件，对设备容器（iNeuKernel）接入的设备或传感器数据进行视图建模组态、数据实时展示和预警等。如下图：
在原有图元基础上增加到3500多个SVG矢量图元，涉及到61个领域及分类。并且支持对编辑好的界面进行分享，可以对分享的链接和二维码集成到其他系统中。
视图建模（iNeuView）特点：
- 建模和组态的图元信息丰富，满足不同行业和领域实时监测的需求。
- B/S应用框架，部署、应用和发布方便，易于操作。
- 与设备容器（iNeuKernel）服务组件无缝对接，实现设备或传感器即插、即用、即显示。
- 对编辑的界面预览，查看实际应用效果。
- 对编辑好的界面应用进行分享，方便集成到其他系统中。
- 对编辑好的界面可以发布到iNeuOS桌面，类似Windows应用程序。
7. 机器学习（iNeuAI）
机器学习（iNeuAI）是iNeuOS的核心组件，对接入的数据进行数据样本标记、数据算法模型训练、工况智能预警。现在支持决策树和KMeans两种算法，从规划角度考虑，还要增加算法和完善产品体系。如下图：
机器学习（iNeuAI）从体量巨大、结构繁多的数据中挖掘出潜藏在数据中的规律和知识，使数据发挥最大化的价值。与传统软件开发不同之处，如下示意图：
应用效果，如下图：
8. 征集需求及合作
希望广大网友，针对领域应用提出具体的需求，在符合iNeuOS的基础上考虑集成进来。可以加入物联网&集成技术群：54256083，进行讨论。
我们不排斥任何形式的合作，技术方面和市场方面。当然，如果有实力的技术加入我们团队，那就更好了。
物联网&集成技术 QQ群：54256083
物联网&集成合作 QQ群：727664080