RFID Protocol Design, Optimization, and Security for the Internet of Things

✍ Scribed by Alex X. Liu, Muhammad Shahzad, Xiulong Liu, Keqiu Li

• Publisher: The Institution of Engineering and Technology
• Year: 2017
• Tongue: English
• Leaves: 274
• Series: IET Control, Robotics and Sensors Series, 112
• Category: Library

✦ Synopsis

Radio-frequency identification (RFID) uses electromagnetic fields to automatically identify and track tags attached to objects. The tags contain electronically stored information. RFIDs have been widely used in countless applications such as object tracking, 3D positioning, indoor localization, supply chain management, automotive, inventory control, anti-theft, anti-counterfeit, and access control. The Internet of Things (IoT) promises a huge growth in RFID technology and usage.

This book covers the topic of RFID protocol design, optimization, and security. RFID systems allow for a much easier and error free inventory management and tracking, but the probabilistic nature of RFID protocols makes the design and optimization complex and challenging. Most existing commercial RFID systems are not well designed. In this book, the authors aim to demystify complicated RFID protocols and explain in depth the principles, techniques, and practices in designing and optimizing them.

✦ Table of Contents

Cover
Contents
List of figures
List of tables
Authors' biographies
1 RFID identification—design and optimization
1.1 Introduction
1.1.1 Background and problem statement
1.1.2 Summary and limitations of prior art
1.1.3 System model
1.1.4 Proposed approach
1.1.4.1 Population size estimation
1.1.4.2 Finding optimal level
1.1.4.3 Population size reestimation
1.1.4.4 Finding hopping destination
1.1.4.5 Population distribution conversion
1.2 Related work
1.2.1 Nondeterministic identification protocols
1.2.2 Deterministic identification protocols
1.2.3 Hybrid identification protocols
1.3 Optimal tree hopping
1.3.1 Average number of queries
1.3.2 Calculating optimal hopping level
1.3.3 Maximum number of queries
1.4 Minimizing identification time
1.5 Discussion
1.5.1 Virtual conversion of population distributions
1.5.2 Reliable tag identification
1.5.3 Continuous scanning
1.5.4 Multiple readers
1.6 Performance comparison
1.6.1 Reader side comparison
1.6.1.1 Normalized reader queries
1.6.1.2 Identification speed
1.6.2 Tag side comparison
1.6.2.1 Normalized tag responses
1.6.2.2 Tag response fairness
1.6.2.3 Normalized collisions
1.6.2.4 Normalized empty reads
1.7 Conclusion
2 RFID identification—fairness
2.1 Introduction
2.1.1 Motivation and problem statement
2.1.2 Limitations of prior art
2.1.3 Proposed approach
2.1.3.1 Identification process
2.1.3.2 Large frame size implementation
2.1.3.3 Multiple readers
2.1.4 Key novelty and contributions
2.2 Related work
2.2.1 Deterministic identification protocols
2.2.2 Nondeterministic identification protocols
2.2.3 Hybrid identification protocols
2.3 Optimal frame size
2.3.1 Jain's fairness index
2.3.2 Total identification time
2.3.3 Expected values of slots
2.3.4 Expected number of Aloha frames
2.3.5 Calculating optimal frame size
2.3.6 Large frame size implementation
2.4 Experimental results
2.4.1 Evaluation of FRIP
2.4.1.1 Fairness
2.4.1.2 Identification time
2.4.1.3 Effect of splitting tag population
2.4.1.4 Effect of error in population size estimate
2.4.2 Comparison with existing protocols
2.4.2.1 Fairness
2.4.2.2 Normalized tag transmissions
2.4.2.3 Normalized energy consumption
2.4.2.4 Normalized identification time
2.5 Conclusion
3 RFID estimation—design and optimization
3.1 Introduction
3.1.1 Motivation and problem statement
3.1.2 Proposed approach
3.1.3 Advantages of ART over prior art
3.2 Related work
3.3 ART—scheme overview
3.3.1 Communication protocol overview
3.3.2 Estimation scheme overview
3.3.3 Formal development: overview and assumptions
3.4 ART—estimation algorithm
3.5 ART—parameter tuning
3.5.1 Persistence probability p
3.5.2 Number of rounds n
3.5.3 Optimal frame size f
3.5.4 Obtaining population upper bound tm
3.6 ART—practical considerations
3.6.1 Unbounded tag population size
3.6.2 ART with multiple readers
3.7 ART—analysis
3.7.1 Independence of estimation time from tag population size
3.7.2 Computational complexity
3.7.3 Analytical comparison of estimators
3.8 Performance evaluation
3.8.1 Estimation time
3.8.2 Actual reliability
3.9 Conclusion
4 RFID estimation—impact of blocker tags
4.1 Introduction
4.1.1 Background and motivation
4.1.2 Problem statement
4.1.3 Limitations of prior art
4.1.4 Proposed approach
4.1.5 Challenges and proposed solutions
4.1.6 Novelty and advantage over prior art
4.2 REB protocol
4.2.1 System model
4.2.2 Protocol description
4.2.3 Functional estimator
4.2.4 Variance of estimator
4.2.5 Refined estimation with k frames
4.3 Parameter optimization
4.3.1 Minimizing time cost
4.3.1.1 Optimizing p
4.3.1.2 Optimizing f
4.3.2 Minimizing energy cost
4.3.2.1 Optimizing f
4.3.2.2 Optimizing p
4.3.3 Trade-off between time cost and energy cost
4.3.4 Dynamic parameter optimization
4.3.5 Avoiding premature termination
4.4 Performance evaluation
4.4.1 Verifying the convergence of f and p
4.4.2 Evaluating the actual reliability
4.4.3 Evaluating the time efficiency
4.4.3.1 Impact of tag cardinality
4.4.3.2 Impact of tag ratio
4.4.4 Evaluating the energy efficiency
4.4.4.1 Impact of tag cardinality
4.4.4.2 Impact of tag ratio
4.4.5 Performance with constraints on time/energy cost
4.5 Related work
4.6 Conclusion
5 RFID detection—missing tags
5.1 Introduction
5.1.1 Background and motivation
5.1.2 Summary and limitations of prior art
5.1.3 Problem statement and proposed approach
5.1.4 Technical challenges and solutions
5.1.5 Key novelty and advantages over prior art
5.2 Related work
5.2.1 Probabilistic protocols
5.2.2 Deterministic protocols
5.3 System model
5.3.1 Architecture
5.3.2 C1G2 compliance
5.3.3 Communication channel
5.3.4 Formal development assumption
5.4 Protocol for detection: RUND
5.5 Parameter optimization: RUND
5.5.1 Estimating number of unexpected tags
5.5.2 False-positive probability
5.5.3 Achieving required reliability
5.5.4 Minimizing execution time
5.5.5 Handling large frame sizes
5.5.6 Expected detection time
5.5.7 Estimating number of missing tags
5.6 Protocol for identification: RUNI
5.7 Parameter optimization: RUNI
5.7.1 Identifying all missing tags
5.7.2 Minimizing the execution time
5.8 Performance evaluation
5.8.1 Impact of number of missing tags on RUND
5.8.2 Impact of number of unexpected tags on RUND
5.8.3 Impact of number of missing tags on RUNI
5.8.4 Impact of number of unexpected tags on RUNI
5.8.5 Impact of deviation from threshold
5.8.6 Estimation accuracy
5.8.7 Comparison with tag ID collection protocol
5.9 Conclusions
6 RFID detection—unknown tags
6.1 Introduction
6.1.1 Background
6.1.2 Motivation and problem statement
6.1.3 Existing work and limitations
6.1.4 Main contributions
6.2 Related work
6.3 Preliminary
6.3.1 System model and assumption
6.3.2 Energy consumption model
6.3.3 Performance metrics
6.4 A sampling bloom filter-based unknown tag detection protocol
6.4.1 Overview of the sampling bloom filter
6.4.2 Protocol design of SBF-UDP
6.4.2.1 Constructing the sampling bloom filter
6.4.2.2 Verifying the tag identity
6.4.2.3 Announcing unknown identity
6.4.3 Investigating the detection accuracy
6.4.4 Analyzing the performance of SBF-UDP
6.4.4.1 Time cost
6.4.4.2 The minimum time cost
6.4.4.3 Energy cost
6.4.4.4 The minimum energy cost
6.4.4.5 Joint optimization of time and energy costs
6.5 Performance evaluation
6.5.1 Demonstrating the advantages of sampling bloom filter
6.5.2 Comparing with the prior related protocols
6.5.2.1 Execution time
6.5.2.2 Energy cost
6.5.3 The actual detection reliability
6.5.4 The impact of channel error
6.6 Conclusion
7 RFID queries—single category
7.1 Introduction
7.1.1 Background and motivation
7.1.2 Problem statement
7.1.3 Limitations of prior art
7.1.4 Proposed approach
7.1.5 Technical challenges and proposed solutions
7.1.6 Advantages over prior art
7.2 Related work
7.3 System model
7.3.1 Architecture
7.3.2 C1G2 compliance
7.3.3 Communication channel
7.3.4 Independence assumption
7.4 RFID tag search protocol
7.4.1 Protocol description
7.4.2 Estimating number of tags in set C
7.5 Parameter optimization
7.5.1 False positive probability
7.5.2 Confidence condition
7.5.3 Duration condition
7.5.4 Handling large frame sizes
7.6 Performance evaluation
7.6.1 Accuracy
7.6.1.1 Observed confidence interval vs. |A|
7.6.1.2 Observed confidence interval vs. |B|
7.6.1.3 Observed confidence interval vs. |C|
7.6.2 Execution time
7.7 Conclusion
8 RFID queries—multiple category
8.1 Introduction
8.1.1 Background and problem statement
8.1.2 Limitations of prior art
8.1.3 Proposed approach
8.1.3.1 Top-k query
8.1.3.2 Segmented perfect hashing
8.1.4 Technical challenges and solutions
8.1.5 Novelty and advantage over prior art
8.2 The basic protocol: TKQ
8.3 The supplementary protocol: SPH
8.3.1 Motivation and challenge
8.3.2 Case study
8.3.3 Detailed design of SPH
8.3.4 Parameter optimization
8.3.5 Discussion on some practical issues
8.3.5.1 Identification of category IDs
8.3.5.2 Deployment of multiple readers
8.4 Related work
8.5 Performance evaluation
8.5.1 Time efficiency
8.5.1.1 Impact of κ
8.5.1.2 Impact of ι
8.5.1.3 Impact of μ
8.5.1.4 Impact of σ
8.5.2 Reliability
8.5.3 Time efficiency vs. accuracy
8.6 Conclusion
9 RFID privacy and authentication protocols
9.1 Introduction
9.2 Premier RFID authentication and privacy protocols
9.2.1 Tag "killing" protocols
9.2.2 Cryptography protocols
9.3 RFID privacy devices
9.3.1 Faraday's cage
9.3.2 Active jamming device
9.3.3 Blocker tag
9.4 RFID protocols based on hash functions
9.4.1 Hash lock: the original hash function-based approach
9.4.2 Tree-based approaches
9.4.3 HashTree: a dynamic key-updating approach
9.5 Other RFID authentication and privacy protocols
9.5.1 Minimalist cryptography
9.5.2 RFIDGuard: an authentication and privacy protocol designed for passive RFID tags
9.6 Conclusion
References
Index
Back Cover