16-01-2012, 06:15 PM
[attachment=16134]
DESIGN OF ACCELERATING TECHNIQUES FOR PERFORMANCE ISSUES IN RFID SYSTEMS
1. INTRODUCTION:
The present RFID systems have the ability to invigorate capture, processing and propulsion in several concerted and networked systems. These systems require low-profile, mobile and cost effective devices [1,2]. But on the other side, there exists heavy restraints for such systems which include, but are not limited to, the deficiency of large storage, the lack of powerful computational elements, a low bandwidth communication infrastructure and limited power resources. Apart from that, the performance objectives of RFID systems cannot be kept aloof from the environment with the real-world uncertainties in which these systems operate. As a result, the performance of RFID systems is immensely affected and these systems are bound not to deliver their best. In such a scenario, to tackle the challenges offered in the performance of RFID systems, there is an intense necessity to address the performance issues arising in RFID systems [3,5,15]. In this thesis, techniques for such performance issues in the existing RFID systems are designed and presented. These techniques further assists in moving towards the performance objectives of such systems which include low cost, data reliability, fast identification etc. The techniques proposed in the thesis are accelerating techniques that lead to improved performance of the RFID systems without the time and expense of upgrading hardware. An architecture accommodating these accelerating techniques has also been designed and proposed [12]. Also, an implementation of the RFID architecture so proposed as working RFID model at application layer has been performed.
2. OBJECTIVE AND SCOPE
The main objective of this research is to expose and address the performance issues associated with the existing RFID systems along with implementation of the a novel architecture for the current RFID systems at application layer. These issues are currently hindering the growth of RFID technology and preventing it from gaining everyday acceptance. From the results derived in this thesis, RFID can then be effectively applied to consumer applications as well as to enterprises while gaining momentum in acceptance. This research can help companies to perceive and implement RFID from bleeding edge technology to a leading edge technology.
Synopsis-1
3. DESCRIPTION OF THE RESEARCH WORK
3.1 RESEARCH PROBLEM
Across the globe, the major performance challenges faced by the RFID (Radio Frequency Identification) technology are uncertainty occurring in RFID Data, RFID Reader tag communication overhead and collisions. These issues resist the technology to perform upto the best of its capabilities. My research theme is to device scalable methods and algorithms of high performance which helps the technology excel and applicable across diverse domains.
More precisely, following are the closely knit dimensions of this research goal:
Research Goals:
G1. How to control the uncertainty appearing in the RFID Data?
G2. How to effectively utilize the time gap when an RFID reader is waiting for the RFID tag to respond to an earlier transmitted command after pre computation of one or more commands and what technique to apply in such a realistic scenario?
G3. What technique must be designed to address anti-collision in RFID Tags which not only support a large number of tags that overlap in time but also provides a unique encryption scheme?
G4. Design and proposal of an architecture which is capable of simulating the RFID scenario and accommodating at least one of the goals stated above.
G5. Implementation of the proposed application layer oriented novel architecture for RFID.
4.2 SOLUTION METHODOLOGIES AND RESULTS:
Synopsis-2
G1: How to control the uncertainty appearing in the RFID Data?
Despite of much research on RFID data management like proposal of smoothing algorithms etc, the issue of unreliable readings is not addressed from the point of fuzzy control methodology in the literature. This calls for a demand to address this issue for the improvement of the technology.
The proposed method:
This part of the work in particular proposes a fuzzy methodology for monitoring or controlling the amount of uncertainty occurring in the raw RFID data. This model is further built upon a set of user-supplied human language rules utilizing the local RFID system environment knowledge. In the proposed approach, a fuzzy model called RFID fuzzy inference system was designed and simulated which was fed up the input parameters noise (false positive reads) and detection rate (of the reader, which in our case was at most 200 tags per interrogation cycle). Based on these input parameters, inference rules and RFID specific membership functions were designed [23].
Results:
Figure 5, depicts one of the snapshots of the results of the simulations of the proposed fuzzy technique. This implementation has been done in Matlab version 7.0.1.It shows how rules were fired in the proposed fuzzy RFID uncertain readings simulation technique. It also visualizes the entire implication process of noise and detection rate influencing the RFID readings from beginning till the end.
Fig.5 Controlled amount of uncertainty in RFID data
Here, in Figure 5, corresponding to the simulation results, the line indices corresponding to the inputs noise and detection rate were moved horizontally to compute their effects on the output.
Synopsis-3
This is where the simulation results showed how the amount of uncertainty in the RFID readings (for example embedded with noise and detection rate) can be controlled and it can be used to handle uncertainty in RFID data for use in business applications to improve the performance of RFID systems.
G2: How to effectively utilize the time gap when an RFID reader is waiting for the RFID tag to respond to an earlier transmitted command (after pre computation of one or more commands and what technique to apply in such a realistic scenario)?
One of the significant issues still prevailing in the RFID systems and is considered as another rebellious issue is that of the large communication time in an RFID system. Primarily our work focuses on the time gap utilization when an RFID reader is waiting for the RFID tag to respond back. Although much of work has been done on RFID systems performance and RFID communication, but an accelerating technique like the one proposed in this work is found missing in the literature. The working behind the proposed genetic algorithm for pre-scheduling of commands in RFID reader systems is depicted in Figure 6. This algorithm makes use of the
process of natural selection and evolution. The proposed genetic algorithm for the RFID reader systems is designed to minimize the overall time required for pre-scheduling of all the available commands in RFID reader systems. Or, in other words, this genetic algorithm, aims to optimize the communication time in RFID reader systems.
DESIGN OF ACCELERATING TECHNIQUES FOR PERFORMANCE ISSUES IN RFID SYSTEMS
1. INTRODUCTION:
The present RFID systems have the ability to invigorate capture, processing and propulsion in several concerted and networked systems. These systems require low-profile, mobile and cost effective devices [1,2]. But on the other side, there exists heavy restraints for such systems which include, but are not limited to, the deficiency of large storage, the lack of powerful computational elements, a low bandwidth communication infrastructure and limited power resources. Apart from that, the performance objectives of RFID systems cannot be kept aloof from the environment with the real-world uncertainties in which these systems operate. As a result, the performance of RFID systems is immensely affected and these systems are bound not to deliver their best. In such a scenario, to tackle the challenges offered in the performance of RFID systems, there is an intense necessity to address the performance issues arising in RFID systems [3,5,15]. In this thesis, techniques for such performance issues in the existing RFID systems are designed and presented. These techniques further assists in moving towards the performance objectives of such systems which include low cost, data reliability, fast identification etc. The techniques proposed in the thesis are accelerating techniques that lead to improved performance of the RFID systems without the time and expense of upgrading hardware. An architecture accommodating these accelerating techniques has also been designed and proposed [12]. Also, an implementation of the RFID architecture so proposed as working RFID model at application layer has been performed.
2. OBJECTIVE AND SCOPE
The main objective of this research is to expose and address the performance issues associated with the existing RFID systems along with implementation of the a novel architecture for the current RFID systems at application layer. These issues are currently hindering the growth of RFID technology and preventing it from gaining everyday acceptance. From the results derived in this thesis, RFID can then be effectively applied to consumer applications as well as to enterprises while gaining momentum in acceptance. This research can help companies to perceive and implement RFID from bleeding edge technology to a leading edge technology.
Synopsis-1
3. DESCRIPTION OF THE RESEARCH WORK
3.1 RESEARCH PROBLEM
Across the globe, the major performance challenges faced by the RFID (Radio Frequency Identification) technology are uncertainty occurring in RFID Data, RFID Reader tag communication overhead and collisions. These issues resist the technology to perform upto the best of its capabilities. My research theme is to device scalable methods and algorithms of high performance which helps the technology excel and applicable across diverse domains.
More precisely, following are the closely knit dimensions of this research goal:
Research Goals:
G1. How to control the uncertainty appearing in the RFID Data?
G2. How to effectively utilize the time gap when an RFID reader is waiting for the RFID tag to respond to an earlier transmitted command after pre computation of one or more commands and what technique to apply in such a realistic scenario?
G3. What technique must be designed to address anti-collision in RFID Tags which not only support a large number of tags that overlap in time but also provides a unique encryption scheme?
G4. Design and proposal of an architecture which is capable of simulating the RFID scenario and accommodating at least one of the goals stated above.
G5. Implementation of the proposed application layer oriented novel architecture for RFID.
4.2 SOLUTION METHODOLOGIES AND RESULTS:
Synopsis-2
G1: How to control the uncertainty appearing in the RFID Data?
Despite of much research on RFID data management like proposal of smoothing algorithms etc, the issue of unreliable readings is not addressed from the point of fuzzy control methodology in the literature. This calls for a demand to address this issue for the improvement of the technology.
The proposed method:
This part of the work in particular proposes a fuzzy methodology for monitoring or controlling the amount of uncertainty occurring in the raw RFID data. This model is further built upon a set of user-supplied human language rules utilizing the local RFID system environment knowledge. In the proposed approach, a fuzzy model called RFID fuzzy inference system was designed and simulated which was fed up the input parameters noise (false positive reads) and detection rate (of the reader, which in our case was at most 200 tags per interrogation cycle). Based on these input parameters, inference rules and RFID specific membership functions were designed [23].
Results:
Figure 5, depicts one of the snapshots of the results of the simulations of the proposed fuzzy technique. This implementation has been done in Matlab version 7.0.1.It shows how rules were fired in the proposed fuzzy RFID uncertain readings simulation technique. It also visualizes the entire implication process of noise and detection rate influencing the RFID readings from beginning till the end.
Fig.5 Controlled amount of uncertainty in RFID data
Here, in Figure 5, corresponding to the simulation results, the line indices corresponding to the inputs noise and detection rate were moved horizontally to compute their effects on the output.
Synopsis-3
This is where the simulation results showed how the amount of uncertainty in the RFID readings (for example embedded with noise and detection rate) can be controlled and it can be used to handle uncertainty in RFID data for use in business applications to improve the performance of RFID systems.
G2: How to effectively utilize the time gap when an RFID reader is waiting for the RFID tag to respond to an earlier transmitted command (after pre computation of one or more commands and what technique to apply in such a realistic scenario)?
One of the significant issues still prevailing in the RFID systems and is considered as another rebellious issue is that of the large communication time in an RFID system. Primarily our work focuses on the time gap utilization when an RFID reader is waiting for the RFID tag to respond back. Although much of work has been done on RFID systems performance and RFID communication, but an accelerating technique like the one proposed in this work is found missing in the literature. The working behind the proposed genetic algorithm for pre-scheduling of commands in RFID reader systems is depicted in Figure 6. This algorithm makes use of the
process of natural selection and evolution. The proposed genetic algorithm for the RFID reader systems is designed to minimize the overall time required for pre-scheduling of all the available commands in RFID reader systems. Or, in other words, this genetic algorithm, aims to optimize the communication time in RFID reader systems.