Get Latest CSE Projects in your Email


A Semantic Situation Awareness Framework for Indoor Cyber-Physical Systems

ABSTRACT

Recently, the domain of cyber-physical systems (CPSs) has emerged as a successor to the traditional embedded systems and the wireless sensor networks. The relatively new cyber-physical domain offers tight integration of control, communication and computation components to develop advanced web based application in various heterogeneous domains such as health care, disaster management, automation and environment monitoring. The applications of indoor CPSs include remote patient monitoring, smart home, etc. with focus on situation awareness via event identification from context information.

The principal challenges associated with the development of situation awareness applications include uncertainty in contextual data, incomplete domain knowledge, interoperability between interconnected systems and effective utilization of spatial information. This dissertation addresses these challenges by providing a comprehensive situation awareness framework for event comprehension utilizing raw sensor data and spatial information. Semantic web based annotation and mapping techniques are used to provide interoperability. The framework contains contextual situation awareness and location awareness stages towards achieving effective event assessment.

The contextual situation awareness stage provides fuzzy abductive reasoning based architecture to transform raw physical sensor data to low-level fuzzy abstraction. These abstractions are used for event assessment with associated degree of certainty. The location awareness stage includes methodologies to hierarchically map indoor objects and define the object-event relationship in ontology, which is further exploited for event discrimination. This dissertation also presents a fusion based indoor positioning algorithm to provide accurate spatial information to assist location awareness. The algorithm uses extensive training of received signal strength (RSS) and time difference of arrival (TDoA) signals to estimate distance and position. The comprehensive framework is evaluated through an implementation of simulated indoor fire in a controlled environment.

SITUATION AWARENESS IN CYBER-PHYSICAL SYSTEMS

Figure 2.3: Examples of cyber-physical systems

Figure 2.3: Examples of cyber-physical systems

The CPSs are also being used for control of large infrastructure such as air traffic control, city traffic & congestion management and asset monitoring systems. In health care domain, CPSs are being quickly adopted for remote patient monitoring application and first responder systems. Surveillance and tracking applications in military domain using unmanned air vehicle can also be classified as a CPS application. Figure 2.3 shows few of these application domains for the cyber physical systems.

Figure 2.8: Semantic context abstractions

Figure 2.8: Semantic context abstractions

The concept of abstraction organizes context in the form of reoccurring patterns or set associated with an event. The raw sensor information from the environment can be translated in the form of low-level abstractions, widely known in semantic web domain as qualities. The context sources or associated events can be distributed in high-level abstractions, often known as entities. High temperature is a quality, which is associated with fire entity; similarly medium temperature abstraction can be explained from entity such as normal room condition.

CONTEXTUAL SITUATION AWARENESS VIA FUZZY ABDUCTIVE REASONING

Figure 3.3: Graphical representation of reasoning rules with crisp abstractions

Figure 3.3: Graphical representation of reasoning rules with crisp abstractions

Figure 3.3 represents the DKB containing io:quality abstraction with their ranges and their association with the io:entity. The DKB also includes the io:quality as a function of appropriate io:quality Type.

Figure 3.7: An event extraction framework from contextual data aided by ontologies

Figure 3.7: An event extraction framework from contextual data aided by ontologies

The SSW also provides support for modeling flexibilities for complex rules, interoperability via standards and autonomous and intelligent decision making. The dissertation utilizes the SSW assisted methodology to integrate semantic web with the proposed event identification framework as described in Figure 3.7.

Figure 3.10: The experimental setup containing two fire events and path of the mobile platform in the indoor environment

Figure 3.10: The experimental setup containing two fire events and path of the mobile platform in the indoor environment

The evaluation of the fuzzy abductive reasoning approach was performed on an indoor fire scenario consisting two distinct fires simulated at different locations with additional context sources in the environment. A platform mounted with an infrared temperature and a carbon dioxide sensor was used to obtain physical sensor data from the environment, originated from the context sources, with the goal of extracting the fire event.

AN ALGORITHM FOR ACCURATE INDOOR LOCALIZATION

Figure 4.2: Comparison of indoor object on 5 meter scale

Figure 4.2: Comparison of indoor object on 5 meter scale

Figure 4.2 displays comparison of different indoor object sizes with accuracy of the GPS. The dissertation proposes accuracy of the indoor positioning system to be in the range of 20-30 centimeters for efficient localization of the indoor objects.

Figure 4.7: Classical setup of an experimental indoor positioning system

Figure 4.7: Classical setup of an experimental indoor positioning system

Figure 4.7 shows a basic setup of an IPS having multiple beacons mote mounted on the ceiling and a listener mote on floor level. In the case of multiple beacon motes, the system can have compound numbers of trilateration combinations for beacon motes. This phenomenon can affect the efficiency of the system in either way by providing better average position or one faulty beacon can bring down the average value of the position estimation. In general application it is necessary to use more than three beacons for full room coverage.

OPTIMIZATION OF ENTITY IDENTIFICATION RESULTS USING SPATIAL INFORMATION

Figure 5.4: Relationship among POI individuals and structural individuals

Figure 5.4: Relationship among POI individuals and structural individuals

Figure 5.4 explains has POI and is Located In properties in detail. The BedRoom-1 is an individual of the BedRoom class while the BedRoom is a subclass of the Room class. BedRoom class has multiple individuals BedRoom-1 and BedRoom-2. Two different objects, a bed and a chair, are present in the Bedroom-1 and they are related to the BedRoom-1 with multiple has POI properties. These individuals, Bed-1 and Chair-1, are related with their respective parent class with has-individual property.

Figure 5.7: Raw spatial coordinates of the indoor objects in the experimental setup

Figure 5.7: Raw spatial coordinates of the indoor objects in the experimental setup

The evaluation of the proposed location based methodology for optimization of entity identification was performed using an experimental setup described in Figure 5.7. The indoor objects were semantically modeled in the indoor location ontology. The indoor location ontology also contained the associations between the indoor objects and the entities to be determined. The experimental setup included multiple indoor objects and fire entities were simulated at the Fireplace-1 and Chair-1 locations whereas the fire at the Chair-1 was only considered as the actual situation.

THE SITUATION AWARENESS FRAMEWORK AND APPLICATION CASES

Figure 6.3: Simulated indoor fire scenario

Figure 6.3: Simulated indoor fire scenario

The experimental setup had two isolated focus entities: a fire at a fireplace and a chair of fire. The fire entities were simulated using candles and were spatially isolated in the laboratory. The environment also contained context sources such as a room heater, people etc. effecting the contexts information generated from those fire. A mobile sensing platform, equipped with temperature and carbon dioxide sensors, was used to acquire physical context information from the entities. The mobile sensing platform also contained the reasoning mechanism for entity identification.

Figure 6.8: Location based entity discrimination

Figure 6.8: Location based entity discrimination

The indoor location ontology also contained the relationships between indoor objects and applicable entities for those objects. The fireplace, while in use, produces high temperature and carbon dioxide contexts. Hence, the entities such as a fire and the presence of a room heater cannot be considered as an actual fire situation in this case. Therefore, the fire and the presence of room heater were considered as not applicable entities at the fireplace and were modeled in the indoor location ontology.

Figure 6.10: Graph of entity detection rules for the subset of indoor patient monitoring system

Figure 6.10: Graph of entity detection rules for the subset of indoor patient monitoring system

Similarly, the high and low temperature at stove can be explained from the stove on or off entities. The reasoning rules displayed in Figure 6.10 provide situational context awareness from the physical sensory information obtained from the patient and the environment. For evaluation of efficient situation, modeling of the relationships between the indoor objects and the entities is necessary.

CONCLUSION AND FUTURE WORK

Conclusion

This dissertation introduced a framework to develop situation awareness applications in the cyber-physical domain. This work focused on entity identification task from the environmental context information effectively utilizing the spatial information. The framework was successfully deployed and evaluated for an indoor fire scenario simulated in a controlled laboratory environment. Earlier in this report, the undeveloped domain of cyber-physical system was introduced with its features, challenges and architecture. The dissertation focused on addressing the challenges associated with the cyber component of the CPSs. The dissertation also addressed the problem of situational awareness in the indoor CPSs in reference to related work. The challenges such as entity identification, interoperability, uncertainty-modeling and location awareness were handled via following contributions.

Future Work

The dissertation provided a step in the direction of modeling the object-entity relationship for situation awareness applications. In some cases, the entity observed at the indoor object has spatio-temporal implications with adjacent indoor objects. Assume a scenario where the physical context information explains HTHD entity at the treadmill. After a moment, the HTHD entity is also detected by the application at the adjacent chair. This phenomenon can be explicitly explained by one of the following cases: (a) patient is resting at the chair after a workout or (b) patient is observing the actual HTHD condition. According to the framework introduced in this dissertation, the HTHD will not be detected at the treadmill and will be detected at the chair from the physical context information observed from the body area sensors.

Source: Wright State University
Author: Pratikkumar Desai

Download Project

>> Latest IoT based Automation Projects for Engineering Students

For Free CSE Project Downloads:

Enter your email address:
( Its Free 100% )


Leave a Comment

Your email address will not be published. Required fields are marked *