1 Introduction
2 Related works
2.1 IoT modeling schemes
2.2 IoT addressing schemes
3 nID physical object preliminaries
-
Space-time information can be jointly applied to determine ubiquitous objects. Thereinto, time information is always dynamically varied, and can be regarded as a unique parameter, and space information is also changed according to different positions. For instance, GPS [21] is a typical technology to detect the location information.
-
Biometric characteristic based identification systems mainly applies the unique biometric attributes for identification, such as finger print recognition, hand geometry, face recognition, and retina/iris recognition [22]. The biometric characteristics can provide high identification accuracy with increasing applications.
-
Other physical/chemical parameters mainly refer to non-unique attributes including environmental temperature /humidity /vibration /pressure, object material /color /size, and radar cross section (RCS). For instance, the sensing technologies (e.g., radar, and infrared) are applied to detect such parameters for identification.
-
Enrich the physical object modeling methods in the IoT: Both ID physical objects and nID physical objects exist in IoT. ID physical objects are addressable via the modeling schemes of IP addresses, EPC/UID code, and other identifiers. By developing nID based modeling schemes, nID physical objects will also become addressable like ID physical objects, therefore combined modeling and addressing schemes can be established in the IoT.
-
Realizing the things unity and fusion in the IoT: The heterogeneity of ubiquitous things impose great difficulty to their ID based modeling and addressing in IoT. The nID based modeling and addressing help to realize the unity and fusion of ubiquitous things through uniform machine understandable description.
-
Improving transmitting efficiency: As most nID sensing data, such as data in video form, require large communication resources for transmitting, it is necessary to transmit the decreased data after modeling so that transmission efficiency can be improved.
-
Reducing data storage: Since objects usually contain many elements and relations with other objects, storing their information directly will occupy large space. Assigning nID codes to them can reduce backend data storage.
4 Tree-Code modeling for physical objects
4.1 Data element abstraction for nID physical object modeling
-
Space-time information: Considering the physical objects’ boundary, each object in Physical-world has a unique location attribute at one moment in a coordinate system. In this paper, we use Location(\(t\)) to represent the space-time information for a physical object. If the location is accurate and the detection interval is short enough to reach space-time consistency, objects’ information at different time can be associated. In this case, their context information can also be associated. In some cases, an object can identified and addressed by only the Location(\(t\)).
-
Attribute: Besides the ID number and Space-time information, some kinds of sensors can detect object’s nature attributes, such as color, shape, weight, and material. For instance, in the radar eyes, the object’s attributes can be shown as RCS, glint, electromagnetic wave scattering sources, spatial power spectral, and Doppler. While a vision sensor can detect the object’s vision attributes, color, size and shape. In most cases, the attributes can be used for object’s category identification, not for unique. However, in some cases, the attributes can also be used to identify a thing uniquely. For instance, in the scenario of human recognition, when there is no ID based identification, attributes of the person such as fingerprint, face, iris, and even DNA can be used and translated into codes to accomplish the identification uniquely.
-
Behavior: The thing’s behavior data means not only its self action, but also the interaction between two objects or the thing and environment it exists in. Meanwhile, behaviors can also be applied for identification, for instance a person can be identified according to the sensed walking postures.
4.2 Tree-Code modeling with data elements
4.3 Tree-Code formal description
Algorithm |
---|
\(\langle \)
Tree-Code root
\(\rangle \)
|
\(\langle \)
Tree-Code namespace=“...”
\(\rangle \)
|
\(\langle \)
ID type=EPC
\(\rangle \)
VALUE
\(\langle \)
/ID
\(\rangle \)
|
\(\langle \)
Space-time information type=GPS
\(\rangle \)
Location(t)
|
\(\langle \)/Space-time information
\(\rangle \)
|
\(\langle \)
Attribure
\(\rangle \)
|
\(\langle \)
Color
\(\rangle \)
Red;
\(\langle \)/Color
\(\rangle \)
|
......
|
\(\langle \)/Attribute
\(\rangle \)
|
\(\langle \)
Behavior
\(\rangle \)
|
\(\langle \)
Velocity
\(\rangle \)
8m/s;
\(\langle \)/Velocity
\(\rangle \)
|
\(\langle \)
Hold
\(\rangle \)
Bag;
\(\langle \)
Hold
\(\rangle \)
|
......
|
\(\langle \)/Behavior
\(\rangle \)
|
\(\langle \)/Tree-Code
\(\rangle \)
|
5 Tree-Code based addressing for nID physical objects
5.1 Tree-Code based addressing scheme
-
Addressing services: Addressing services can be classified into ID-based addressing services and nID-based addressing services according to the type of input, which is part of the Tree-Code. For ID-based addressing services, the addressing process is similar to that of ONS. However, the addressing process of nID-based servers is different. The interested resource may be distributed in multiple servers. A good choice for nID physical object addressing is taking advantage of the Content delivery network (CDN) [25], in which resource are addressed and routed by some piece of the content. CDN is suitable for addressing physical objects modeled by Tree-Code due to the abstracted data elements in Tree-Code. In some other applications, nID attribute (e.g., GPS based location data) had been used in designing routing protocols, which can be regarded as a special case of physical object addressing [26].
-
Middleware: The middleware is a crucial function module for physical objects addressing which act as a conjunction that connects physical objects modeling, addressing services and distributed application platform. In the following, typical working modes of middleware are presented:1.Physical to Cyber working mode: This working mode indicates the process that a physical object enters the sensing range of some ubiquitous sensors, instantly, its ID type data and nID type data are gathered and encoded into Tree-Code with the modeling rules we have discussed. The Tree-Code sent to middleware for further analysis after being constructed. The middleware can choose ID or nID based addressing services automatically. Then, the ID or nID codes are respectively transmitted to ID based or nID based addressing services for final addressing. The result of addressing services returns to middleware and is delivered to distributed application platform for further intelligent process or decision making. Finally, the feedback from distributed application platform will be transmitted to the middleware.2.Cyber to Physical working mode: This working mode indicates that the physical objects addressing inquiry is launched from the Cyber-world (such as from distributed application platform), and the objective of the addressing process is to locate the physical objects. Similar to the Physical to Cyber working mode, the object that needs to been addressed is first modeled by Tree-Code, then the Tree-Code is send to the middleware which chooses addressing services automatically. The result from addressing services can be used by the middleware to locate a physical object based on which the data (e.g., controlling command) will be sent to it. ;3.Hybrid working mode: In some cases, the middleware works as the combination of the above working modes which is named as hybrid working mode here due to the fact that this working mode composes both the direction from Physical-world to Cyber-world and from Cyber-world back to Physical-world. Loop working mode is common in context aware and ubiquitous computing applications in IoT, in which physical objects together with the context information is modeled with Tree-Code. The middleware, addressing services, and the distributed application platform works together to return controlling command to the Physical-world.