Programming RFID Tags Under

Programming RFID Tags Under Dynamic Environments Math, Science, Technology Center at Paul Laurence Dunbar High School Yuki Inoue Programming RFID Ta...
Author: Avice Rich
1 downloads 0 Views 4MB Size
Programming RFID Tags Under Dynamic Environments Math, Science, Technology Center at Paul Laurence Dunbar High School Yuki Inoue

Programming RFID Tags Under Dynamic Environments

Abstract Yuki Inoue, Dr. Johné M. Parker. University of Kentucky College of Engineering. 2012. RFID, or Radio Frequency Identification, is a newly developed object tracking system that uses radio waves, giving it a tremendous potential to change business. The aim of this project is to figure out which factors affect the RFID programming, so the RFID technology could be implemented onto a laser printer. Because of a laser printer’s unique mechanism, this requires tag programming to be done at a constant speed of at least 35 ppm (pages per minute, approximately 16.3 cm/sec), giving rise to some unique variables not tested in previous RFID-related studies such as the tag velocity and tag orientation. In addition, variables such as air gap, reader input power, tag type, tag uniqueness, and offset distance were tested under both static and dynamic conditions. The results were then analyzed to determine the effect of each factor on the error rate. The air gap, reader power, and velocity affected the error rate uniformly among the tags, with increasing air gap increasing the error rate, increasing reader power decreasing the error rate, and increasing velocity increasing the error rate. The orientation and the offset distance affected tags differently. The results from the static testing suggest that the effect of orientation is strongly tied with the shape of the tags. As for the offset distance, local minima were observed in regions other than the 0mm point. These findings would be of great help in the future creation and commercialization of a laser printer with a robust RFID option.

Programming RFID Tags Under Dynamic Environment

TABLE OF CONTENTS CHAPTER 1: INTRODUCTION

5

1.1 PURPOSE 1.2 PROBLEM

5 5

CHAPTER 2: BACKGROUNDS

6

2.1 BARCODES 2.2 HOW RFID COMMUNICATION WORKS 2.2.1 READERS 2.2.2 TAGS 2.2.3 ELECTRONIC PRODUCT CODE (EPC) 2.2.4 TWO DIFFERENT MEANS OF COMMUNICATION CHAPTER 3: EXPERIMENTAL SET UP

6 8 8 8 9 10 12

3.1 INDEPENDENT VARIABLES AND HYPOTHESIS 3.1.1 TAG ORIENTATION 3.1.2 AIR GAP 3.1.3 READER INPUT POWER 3.1.4 TAG TYPE 3.1.5 TAG UNIQUENESS 3.1.6 OFFSET DISTANCE 3.1.7 TAG VELOCITY 3.2 FIXTURE DESIGN 3.3 SOFTWARE DEVELOPMENT 3.3.1 MDRIVE MOTOR 3.3.2 READER SYSTEM 3.3.2.1 Tag Reading 3.3.2.2 Tag Writing 3.3.2.3 Blocks and Addresses 3.3.3 READ COMMAND VS. WRITE COMMAND 3.3.4 IMPLEMENTATION ON C# 3.3.4.1 SerialPort 3.3.4.2 STPv3 3.4 STATIC PROFILING 3.4.1 STATIC TESTING 3.4.2 OFFSET TESTING 3.5 DYNAMIC PROFILING CHAPTER 4: RESULTS

12 12 13 13 13 14 14 15 15 16 16 19 20 20 20 21 23 23 23 25 25 26 27 29

4.1 STATIC TESTING 4.1.1 ORIENTATION RESULTS 4.1.2 AIR GAP RESULTS

29 29 32

1

Programming RFID Tags Under Dynamic Environment

4.1.3 READER INPUT POWER RESULTS 4.1.4 TAG UNIQUENESS RESULTS 4.1.5 NEXT STEP 4.2 OFFSET TESTING 4.2.1 BUTTON 4.2.2 DOGBONE 4.2.3 AD-805 4.2.4 NEXT STEP 4.3 DYNAMIC TESTING 4.3.1 BUTTON 4.3.2 DOGBONE 4.3.3 AD-805 CHAPTER 5: ANALYSIS

32 33 34 34 34 35 37 38 38 39 40 42 44

5.1 AIR GAP ANALYSIS 5.2 READER INPUT POWER ANALYSIS 5.3 TAG UNIQUENESS ANALYSIS 5.4 TAG TYPE ANALYSIS 5.5 ORIENTATION ANALYSIS 5.6 OFFSET DISTANCE ANALYSIS 5.7 VELOCITY ANALYSIS

44 46 48 49 50 52 55

CHAPTER 6: CONCLUSION

57

ACKNOWLEDGEMENTS

59

REFERENCES

60

2

Programming RFID Tags Under Dynamic Environment

LIST OF TABLES AND FIGURES TABLES Table 1: List of Frequency Categories

8

Table 2: Advantages/ Disadvantages of Tag Types

9

Table 3: EPC Classes

10

Table 4: List of Major Commands Available in MCode

18

Table 5: Static Testing Factors

26

Table 6: Offset Testing Factors

27

Table 7: Dynamic Testing Factors

28

FIGURES Figure 1: Barcode Components

6

Figure 2: EPC Memory

10

Figure 3: Tag Orientation Disk

12

Figure 4: Tags offset at their maximum offset distances

15

Figure 5: Fixture Design

15

Figure 6: MCode Programming

17

Figure 7: Components of a Reader (from left to right): Antenna, Radio, MUX Board

20

Figure 8: Blocks and Addresses

21

Figure 9: Read Command vs. Write Command

22

Figure 10: Chosen Tags (from left): Button, AD-805, Dogbone

27

Figure 11: Button Tag Error Rate

30

Figure 12: Dogbone Tag Error Rate

31

Figure 13: AD-805 Tag Error Rate

32

Figure 14: Uniqueness Studies

33

Figure 15: Offset Button Distance Comparison

35

3

Programming RFID Tags Under Dynamic Environment

Figure 16: Offset Button Power Comparison

35

Figure 17: Offset Dogbone 90 Degrees Air Gap Comparison

36

Figure 18: Offset Dogbone Orientation Comparison

37

Figure 19: Offset AD-805 Orientation Comparison

38

Figure 20: Dynamic Button Power, Velocity, and Air Gap Comparisons

40

Figure 21: Dynamic Dogbone Orientation Comparison

40

Figure 22: Dynamic Dogbone 90 Degrees Power, Velocity, and Air Gap Studies

42

Figure 23: Dynamic AD-805 Orientation Study

43

Figure 24: Theoretical Representation of the Gaussian Surface

45

Figure 25: Static Testing Air Gap Study

45

Figure 26: Static Testing Power Study

47

Figure 27: Interaction Between Power and Orientation

48

Figure 28: Static Testing Tag Uniqueness Study

49

Figure 29: Static Testing Dogbone Orientation Study

50

Figure 30: Tag Shape v. Error Rate

51

Figure 31: AD-805: Orientation v. Offset Distance

52

Figure 32: Velocity and Orientation

52

Figure 33: Angle of Attack

54

Figure 34: Offset Distance Graphs

55

Figure 35: Blind Spot For AD-805

55

Figure 36: Velocity Studies

56

4

Programming RFID Tags Under Dynamic Environment

Chapter 1: Introduction 1.1 Purpose Radio Frequency Identification (RFID) technology allows the users and the industry to operate business more efficiently. As the price of RFID tags fell to a couple of cents in the recent years, large retailers such as Wal-Mart started to see the cost effectiveness and the applicability of the RFID technology. Since 2004, Wal-Mart and the Department of Defense started using RFID tags to track items in their organizations. This forced manufacturers to adapt to RFID technology as well, increasing the demand of the technology. Because information stored in tags is invisible, it is hard for human operators to distinguish one tag from another without the aid of readers. Therefore, it becomes crucial to embed tags inside of papers and print information onto the paper so people can distinguish one tag from another. This is easily done if there is a printer with an RFID option. But following problems arise when a laser printer tries to program a tag while printing: 1. Because a laser printer cannot halt while printing a page, the motor inside of a laser printer cannot stop. The result is dynamic programming of the tags. 2. The tags must be able to withstand intense heat and pressure that the paper undergoes when being printed. Problem #2 has already been solved- current tags can withstand intense heat and pressure. That leaves problem 1. Therefore, the aim of this project is to solve those problems. The purpose of this project is to study how tags behave differently when handled dynamically, and also to identify the factor or a set of factors that affect the tag handling success rate.

1.2 Problem How do tags behave under dynamic environment, and how do different factors affect the RFID programming error rate?

5

Programming RFID Tags Under Dynamic Environment

Chapter 2: Backgrounds 2.1 Barcodes Product identification systems were first suggested in 1932 in the form of a punch card system by Wallace Flint at Harvard Business School. The precursor to the modern 1D barcode was invented in 1948 by Drexel graduate students Bernard Silver and Norman Woodland as a way to automatically read product data during checkout at a local supermarket, improving productivity and speeding up the customer process (Wikipedia, 2011). After several systems designs, an iteration inspired by Morse Code was introduced using patterns of lines. These patterns are recognized optically, with the use of speciallymade light bulbs. The resulting design was called “Classifying Apparatus and Method,” patent #2612994, issued on October 7, 1952.

Country Code

Item Identifier

Manufacturer ID

Check digit

Figure 1: Barcode Components

In 1973, the grocery industry set standards for product identification, leading to the Universal Product Code (UPC), a 12-digit number unique to each product. In 1976, the European Article Numbering (EAN) barcode system, derived from the UPC system, was launched on an international scale by having a 13digit number code–the extra number used to identify the country. These 12 or 13 digits serial codes indicate the location of the product information in the database. Note that therefore barcodes do not contain information such as prices and where they were made. They merely point to the place where that information can be found. To read a barcode, one needs to break it up into four different components and understand what those components mean. The first number (or two or more numbers) on a barcode is (are) the country code. As the name implies, the country code indicates which country the product was

6

Programming RFID Tags Under Dynamic Environment

made in, such as 0 for America. The next two five-digit sequences are the manufacturer ID and the item identifier, indicating which manufacturer made the product, and which product it is. The very last digit of a barcode, the check digit, allows the scanning device to check whether it actually obtained the correct data. The confirmation is done through following the following steps: 1. Add every number on the barcode, located at odd distances away from the check digit, and multiply the result by 3 (ex. (0+8+6+4+2+0)*3=60). 2. Add every number on the barcode, located at even distances away from the check digit (ex. 9+7+5+3+1=25). 3. Add the results obtained from steps 1 and 2 (60+25=85). 4. mod 10 the number obtained in step 3, and subtract that number from 10 (85 mod10=5, 10-5=5). This number should match the check digit of a barcode. Ever since barcodes have infiltrated the transportation industry, they have been the universal system of item identification. However, as the need for faster, more improved and secure identification practices developed, wireless tracking, such as the RFID technology, became popular. Just like the barcodes, RFID tags allow the manufacturers to track products quickly and efficiently. But because RFID communication utilizes radio waves instead of visible waves like in barcodes, RFID tags can be read and programmed without readers visually capturing the surface of the tags. This allows for a fast and efficient product tracking. Also, unlike the barcodes, which only allow the manufacturer to put one unalterable numerical code for each item, RFID tags allow the users to alter their information whenever necessary. This means that the information on RFID tags could be updated while they are being shipped, making it possible for consumers to check information such as a product’s shipment record.

7

Programming RFID Tags Under Dynamic Environment

2.2 How RFID Communication Works 2.2.1 Readers As mentioned before, an RFID system uses radio waves to communicate with tags. The signal is sent by a reader, sometimes called the interrogator, and received by the antenna structure on tags. A tag activates, or “couples,” if and only if the frequency and the power of the received signal are strong enough. The ability of a reader to communicate with a tag is largely dependent on four factors: the design of the antenna, the frequency and the power used to send the signal, and environmental conditions. An example of “environmental conditions” include types of materials used in equipments used to test RFID systems, since metal surfaces are known to reflect radio waves. For common RFID communications, the following four frequency categories are used most often: Name

Frequency Ranges

Low Frequency (LF)

125-134 KHz

High Frequency (HF)

13.553-13.567 MHz

Ultra-High Frequency (UHF)

400-1,000 MHz

Microwave

2.45GHz

Table 1: List of Frequency Categories Therefore it is very important for the reader to be at a specific distance away from tags, sending out signals that have the right frequency and power. 2.2.2 Tags There are three different types of RFID tags: passive, active and semi-passive. Characteristics of passive and active tags are summarized in the following table: Topics

Passive Tags

Active Tags

Cost

Lesser

More

Lifetime

Longer

Shorter

8

Programming RFID Tags Under Dynamic Environment

(theoretically unlimited)

(as long as batteries last)

Data Transmission Rate

Lower

Higher

Activation Power

Higher

Lower

Power Source

External

Internal

(from captured radio waves)

(battery)

Weight

Lighter

Heavier

Range

Lesser

Greater

Table 2: Advantages/ Disadvantages of Tag Types Semi-passive tags have characteristics that are somewhere between active tags and passive tags- they use batteries to run the IC, but communicates by drawing power from the radio wave. As shown in the table, active tags and semi-passive tags contain batteries, making them very expensive. Therefore, they are usually used to track expensive objects. Therefore, passive tags, costing only around a couple of cents, are the most likely the tags that replace the barcodes. For that reason, only the passive tags were tested in this experiment. Once the signal is captured by a tag’s antenna, passive tags try to activate their Integrated Circuit Chip by using the electricity generated from the received radio wave. The activation of the tag, or the “coupling,” is successful if and only if the strength of the radio wave signal is strong enough. Therefore, it is very important that tags are read by specific types of readers and read at a certain distance with a certain frequency. Upon the activation of the tag, it sends the data back to the reader to be interpreted. If the tag is read/write, then a tag may receive information from the reader to alter its contents. In the experiment, all of the tags are Electronic Product Code (EPC, explained later) Tag Class 1 Generation 2 tags, which means that they all have read/write capabilities. 2.2.3 Electronic Product Code (EPC) As mentioned before, barcodes can store much information in them by dividing their 12 digits memory into 4 different components (country code, manufacturer ID, item identifier and check digit). Just like the

9

Programming RFID Tags Under Dynamic Environment

barcodes, RFID tags can store a wealth of information by applying an organization scheme called the Electronic Product Type (EPC), dividing their 96-bits memory into different components. The first 8 bits of EPC is called the Header, and it specifies the version and the generation of the EPC. The next 28 bits are called the EPC Manager, which identifies the organization that is responsible for maintaining the tag. The next 24 bits are called the Object Class, and it specifies the type of the tag (a list of classes shown beneath the example EPC). The last 36 bits (Serial Number) uniquely identifies the item. An example EPC with its components labeled is shown below. Note that most EPC’s are represented using hexadecimal digits (1 hexadecimal digit = 4 binary digits). 0

1

2

3

4

5

6

7

8

9

A

B

C

D

E

F

0

1

2

3

4

Header

EPC Manager

Object Class

Serial Number

(8 bits)

(28 bits)

(24 bits)

(36 bits)

5

6

7

Figure 2: EPC Memory

EPC Tag Class

Type of Tag

Descriptions

Class 0

Passive

Read Only

Class 1

Passive

Write Once

Class 2

Passive

Read/Write, extended memory

Class 3

Semi-Passive

Includes battery source

Class 4

Active

Active transmitter & sensors

Class 5

Reader

Can communicate with other tags

Table 3: EPC Classes 2.2.4 Two Different Means of Communication Coupling between tags and antennas are accomplished thorough either Near Field Communication (NFC) or Far Field Communication (FFC). When a reader attempts to communicate with tags, it generates a magnetic field at its antenna. As the field propagates, the electric field develops by induction. At a

10

Programming RFID Tags Under Dynamic Environment

specific distance between the reader and a tag, there is a zone in which the electromagnetic field separates from the reader antenna to make an electromagnetic wave, which can be solved by the following equation:

d

 2

where d is distance (meters) from the reader and λ is the wavelength of the signal (in meters). The wavelength could be calculated by:

 f  c

where λ = wavelength of frequency band (meters), c = the speed of light 3 x 108 m/s, and f = frequency at 

which system operates (Hz, sec-1). When the distance is smaller than this number, communication is considered near field. When the distance is greater than this number, it is considered far field. Because the main agent of the communication differs from NFC (magnetic waves) to FFC (electric wave), the behaviors of tags significantly differ for the two communications. Therefore, tags and antennas are specifically made for either NFC or FFC.



d 

 2

Near Field

d 

 2

Far Field

According to the SkyeModule M9 Reference Guide, the antenna used in the experiment should emit electromagnetic waves that are 915MHz in frequency. Using equation 2, this means that the electromagnetic wave used in the experiment is roughly 0.3279m or 32.79cm. Using the first equation, the line at which the communication pattern switches from NFC to FFC is approximately 0.0522m or 5.22cm. Therefore, any tag-antenna communications done under 5cm is considered to be NFC, while anything higher than that is considered to be FFC.

11

Programming RFID Tags Under Dynamic Environment

Chapter 3: Experimental Set Up 3.1 Independent Variables and Hypothesis The following are the factors that are tested in the experiment: tag orientations, Air gaps (air gaps), type of tags, reader power input, and the velocity of the motor. 3.1.1 Tag Orientation Orientation is one of requirements for tag specification. However, this is done about the axis parallel to the plane of the antenna. However, in the case with tag-embedded paper going through a printer, an orientation test along an axis perpendicular to the plane of the antenna is much more important, for there is a possibility that the papers are inserted crookedly. Although manufacturers claim that rotation along that axis does not significantly affect the overall performance of tags, their tests are also considered “indicative, performance in real-life applications may vary.” Since they use prototype hardware built specifically for their tags and because they do never disclose the full design of their experiments, testing the effect of orientation is needed. Tag orientation was tested by putting tags on a circular piece of paper with eight lines, each offset counterclockwise at 45 degrees. Since tags do not have a definite center, the integrated circuit of each tag was treated as the center. Therefore I hypothesize that the variable orientations of the tags should not affect their readability.

Figure 3: Tag Orientation Disk

12

Programming RFID Tags Under Dynamic Environment

3.1.2 Air gap Air gap is the distance between a tag and an antenna. As mentioned in the background section, air gap determines if a tag is in the NFC or the FFC. Since most of the prepared tags are for use in FFC, it is expected that the tags would operate under both conditions. However, it is also expected that the behavior of the tags (ex. how they react to different variables) are very different under the two conditions. For NFC tags (i.e. Button and Paperclip), it is expected that they do not operate under FFC condition at all. Therefore, it is expected that tags would behave very differently when they are read closer than 5cm and further than 5cm. Since further distance directly means that less of the emitted electromagnetic waves will reach the tag, it is less likely for the tags to be able to extract the required power from the electromagnetic wave. Therefore, it is expected that a greater air gap would increase the error rate. 3.1.3 Reader Input Power For the particular model of the reader used in the project (SkyModule M9), the power goes from 10dBm to 27dBm. There are multiple studies done on the relationship between reader power and tag readability, and they all conclude that they are directly proportional: bigger the power, higher the readability. Therefore it is expected that higher the power, higher the readability. The reason for studying this variable, however, is because lower power input is better for economic and political reasons. Economic, for lower power implies wasting less energy, and thus less electricity, and political, because each country has different regulations on how strong these antennas can be. Therefore, important to study the behavior of input power so maximum readability under minimum power could be realized. Since larger power input should result in more energized electromagnetic waves, it should be easier for the tags to activate. Therefore, greater reader input power should result in a lower error rate. 3.1.4 Tag Type There were 17 different tags being tested in this experiment. Of the 17, two are deemed near field tags (i.e. Button and Paperclip), one both near field and far field (i.e. Satellite), and the rest are far field. Previous

13

Programming RFID Tags Under Dynamic Environment

studies have revealed that far field tags operate under both NFC and FFC, while near field tags only work under NFC. Therefore it is expected that far field tags work better than near field tags. 3.1.5 Tag Uniqueness Because it is impossible to make the exact same thing, it is expected that the tags of the same tag type behave differently even if they are under the same condition. I expect the difference in performance between tags to be significant, since tag companies usually do not have a standard percentage of times that a tag must successfully interact with the antenna. This is because usual implementations of RFID tags do not have a limit on how many times the user can scan RFID tags. Therefore it is unnecessary to standardize success percentages among the tags. Therefore, it is highly likely that the error rates for the tags are different among the tags of same population. And if this were the case, testing multiple samples of each population of tags would be needed. 3.1.6 Offset Distance Offset distance is only used in the Offset Testing, and it is the distance from the tower, parallel to the track of the linear slide (refer to figure 4). This is used to examine the behavior of the tags when they are not directly under the antenna. Since moving tags are under the tower for only a fraction of a second, offsetting distance from the tower will provide more accurate information about how tags behave when they are moving. Since larger offset distance equates to larger distance between the tag and the reader, it is expected that the readability plummets. However, the severity of the fall as compared to the air gap is uncertain, for there has not been any study done on the matter. However, since the antenna matrix is facing down, it is expected that the offset distance, which is a horizontal displacement, should affect the error rate more significantly than air gap, which is a vertical displacement.

14

Programming RFID Tags Under Dynamic Environment

Figure 4: Tags offset at their maximum offset distances 3.1.7 Tag Velocity Tag velocity is only used in the Dynamic Testing. As the name implies, this is the speed at which the tag moves along the linear slide. Because objects moving at a higher velocity should destabilize the communication between readers and tags, it is expected that higher tag velocities result in lower readability.

3.2 Fixture Design

Figure 5: Fixture Design Since RFID communication is strongly affected by its surroundings, the testing fixture (see above picture) cannot contain open surfaces that are either metallic or liquid. Therefore, the fixture was made of hard

15

Programming RFID Tags Under Dynamic Environment

rubber (the tower), and Lexan (the base). Because the transition from the NFC to the FFC occurs at different distances for every tag, it is important to have a flexible distance options. As the picture shows, the motor slide underneath the tag platform allows the tags to be horizontally offset, while the tower/clip allows the air gap to be vertically manipulated.

3.3 Software Development Because the only software available at the University of Kentucky was programs hard-coded into the hardware (i.e. the motor and the antenna radio), developing a software that can orchestrate the hardware using sophisticated computer language was desperately needed. Therefore, first half of the project period was spent developing software. 3.3.1 MDrive Motor MDrive stepper motor by Schneider Electronics was used in this experiment. MDrive stepper motors have their own programming language called MCode. MCode is a low-level language that can control the motor using two different ways. The first mode is the command-line mode, where the user can type a line of command (ex. MA 1000), and the motor moves accordingly. The other mode is the programming mode, in which the user can create a program to run the motor. By using the programming mode, the user can command the motor to perform complex movements. An example MCode program and a table of commands are shown in Figure 6:

16

Programming RFID Tags Under Dynamic Environment

Figure 6: MCode Programming Command Names

Functions

SL

Slew Rate Move at 2000 µsteps/sec

SL=2000

Move Absolute

MA

Move to position 2000 µsteps

MA=2000

Move Relative

MR

Move 2000 µsteps away from the current position

MR=2000 Label

LB

Declaration of Label Y

LB Y P

Position

17

Programming RFID Tags Under Dynamic Environment

Let the current position be 0

P=0

Call Subroutines

CL

Go to Subroutine Y

CL Y Print

PR

Print the Current Position

PR P Hold

H H 2000

Hold for 2000 milliseconds

H

Hold until the previous command has been executed Branch (repeat the program)

BR

Repeat Label Y

BR Y

Numerical Variables

R1-R4

Set variable R1to be 3

R1=3

Variable

VA

Declaration of user variable X1

VA X1

Running Current

RC

Set RC value to be 50%

RC 50 E

Exit Program

PG

Program

IP

Initial Parameters

PS

Pause

RS

Resume

RT

Return to the Main Program

FD

Reset MDrive

Table 4: List of Major Commands Available in MCode The first three commands on the list (Slew Rate, Move Absolute, and Move Relative) are the most important, since they are the commands that move the slide. The term “µsteps” contained in the description of those three commands is a unit of distance that MDrive motor units use. The relationship

18

Programming RFID Tags Under Dynamic Environment

between µsteps and inches is in = 400*µsteps*(µsteps setting), and for centimeters, it is cm = 157.48*µsteps*(µsteps setting). µsteps setting was kept at 16 throughout the experiment. Using these commands, the user can fully control the movements of the motor, everything from the velocity of the motor to the distance it travels. 3.3.2 Reader System The reader system includes the following: an antenna, a radio, and an MUX Board. The SkyeModule M9 radio by SkyeTek Inc. was used in this experiment (shown right) as a radio, and it controls the different reading/writing settings, such as the reader input power and the frequency of the radio waves, which are sent to the antenna. Then the MUX Board decides which antenna (out of the 4 antennas) the signal is sent to. Finally, the antenna emits radio waves, according to the settings signaled by the radio. As one may notice, there are four loops on the antenna. Each loop is an antenna that is fully functional by themselves. The reason for having four antennas is so that one can program multiple tags at once (in this case, four tags at once). Because this project did not study the multiple tag programming, only one tag was utilized at all times (the second one from the right). Another important fact about the antenna is that it is very inefficient. Because of its design, after the radio waves propagate for a short distance (this distance is dependent on the input power), its strength dies dramatically. In other words, the antenna creates a very strong electromagnetic field near the antenna, but zero to no field after a certain distance. This was so that the antenna does not rewrite the completed tags on top of the printer, waiting for someone to pick them up. Also, there are different laws that regulate how strong the radio waves can be from a certain distances away from the antenna. By creating an antenna that inefficiently drops the strength of the radio waves, both of those problems were solved.

19

Programming RFID Tags Under Dynamic Environment

Figure 7: Components of a Reader (from left to right): Antenna, Radio, MUX Board 3.3.2.1 Tag Reading The antenna starts reading tags in the field when the command READ_TAG is executed. If the reader is unable to find any tags to read from, the flag READ_TAG_DATA_FAIL is returned. Otherwise the flag READ_TAG_DATA_PASS is returned, and the reader obtains the information about the tag. The location on the tag at which the data is collected and the size of the data collected are determined by the address and the block of the reader (explained later). 3.3.2.2 Tag Writing The antenna starts writing tags in the field when the command WRITE_TAG is executed. If the reader is unable to find any tags to write on, the flag WRITE_TAG_DATA_FAIL is returned. Otherwise the flag WRITE_TAG_DATA_PASS is returned, and the reader writes user-specified data onto the tag. The location on the tag at which the data is written is determined by the address of the reader (explained later). 3.3.2.3 Blocks and Addresses As mentioned in the EPC section of this paper, a tag’s information is stored in hexadecimal digits. In the SkyeModule series, four hexadecimal digits of data, or 16 binary digits of data are called blocks, and by altering these values, the user can control things like how much information is obtained when READ_TAG command is executed. So setting the block value to be 1 means that 1 block of data, or 4 hexadecimal digits of data, will be obtained from a tag. Another variable important in the handling of a tag is the address. Address, as one may guess, specifies where data is stored. An address is the index of where the data is on a tag. One address has four hexadecimal digits of data stored- therefore setting the address as 0 will make the reader to read/write from the 1st digit of the tag, and setting the address as 1

20

Programming RFID Tags Under Dynamic Environment

would mean that the reader will read/write from the 5 th digit of the information and so on. By specifying blocks and the addresses of the reader, a user can specify how much he/she wants to read/write, from where on the tag. One warning regarding blocks and tag writing is that the block value and the size of the data being written on the tag must match (ex. When writing 0x49EF on a tag, block size must be 1). Otherwise, an error message INVALID_DATA_LEN response will be returned by the reader. The following picture visually describes the relationship between blocks and addresses.

Figure 8: Blocks and Addresses For example, the reader will return “A3D22341” when READ_TAG is executed under the conditions address=3, block=2. Also, the reader will write “1234” at address 2, when WRITE_TAG is executed under the conditions address=2, input data = “1234.” 3.3.3 Read Command vs. Write Command Read commands take about half as long (around 20ms) as write commands (around 42ms). Since the number of tests conducted in the experiment is going to be very large, it would be beneficial if read commands could be used to assess tag writability. Therefore, a preliminary test was launched to see the performance difference for write commands and read commands. The procedure used in this preliminary test is the exact same as the one used in static testing (refer to section 3.4). The tags tested were Button tag, a near field tag, and the Dogbone tag, a far field tag. The result (Figure 9) revealed that the error rate graph for read commands and write commands are almost identical to each other, indicating that the read command could be used to assess writability. Therefore read commands will be used to assess tag writability instead write commands for static and offset tests in order to quicken the testing processes.

21

Programming RFID Tags Under Dynamic Environment

Figure 9: Read Command vs. Write Command

22

Programming RFID Tags Under Dynamic Environment

3.3.4 Implementation on C# The next step was to implement MCode and antenna commands in more sophisticated programming language, so the motor and the antenna movement could be orchestrated from central software. C# was used for the implementation, because there was a C# program given to the lab by SkyeTek Inc. that could control both the motor and the antenna. Two variables, “SerialPort” for representing the motor, and “STPv3” for representing antenna, are birthed at the beginning of the software. Using these variables, the software can communicate with the hardware directly. The connection between the hard drive and the hardware are established through COM Ports. 3.3.4.1 SerialPort Serial Port class in C# is designed to foster any communications between two hardware. The connection between the two is established through COM Ports, which is done by the code SP.PortName = “Name of the Port.” Once the connection to the motor is established, the software can execute any commands to the motor by running SerialPort.WriteLine(“Command String”). This allows the C# to execute one command at a time. Unfortunately, there is no way to execute an MCode program from C#. Also, since MDrive can only execute one line of code at a time and command execution takes some time, if any new commands are issued by the C# software while MDrive is still executing the previous command, the new command will override the previous one. Therefore the C# software has to wait until the MDrive is done with the command execution. This wait time is around 75 milliseconds for most commands, except for FD, or Factory Reset, taking around a second. 3.3.4.2 STPv3 STPv3, standing for SkyeTek Protocol Version 3, is a group of non-editable codes developed by the SkyeTek. Member classes of STPv3 have “STPv3” as their prefixes, and they act as middle-men between the C# software and the reader hardware. These classes, as well as their functions are as follows: STPv3Request allows the programmer to change the reader settings.

23

Programming RFID Tags Under Dynamic Environment

.Data is a property value in byte[]. This value is the data that will be written on the tag when WRITE_TAG command is issued. This value will change to the value on the tag when READ_TAG command is issued. .Address changes the address value. .Block changes the block value. .Command allows the programmer to execute different commands. The names of the commands must be exactly the same as how it appears in the STPv3Commands (ex. WRITE_TAG). Syntax: requestTag.Command = STPv3Commands.SELECT_TAG; STPv3Commands is a list of commands that can be executed by the reader. These commands are to be used with the .Command function of the STPv3Request objects. SELECT_TAG commands the reader to select a tag. If the tag selection is successful, SELECT_TAG_PASS is returned. If not successful, SELECT_TAG_FAIL is returned. READ_TAG commands the reader to read a tag. If successful, READ_TAG_DATA_PASS is returned. If else, READ_TAG_DATA_FAIL is returned. WRITE_TAG commands the reader to write a tag. If successful, WRITE_TAG_DATA_PASS is returned. If else, WRITE_TAG_DATA_FAIL is returned. *For other commands and their corresponding respond codes, refer to SkyeTekProtocol_V3_Messages.pdf STPv3ResponseCode lists out response codes that could be returned as a result of the executed command. Syntax: response.ResponseCode == STPv3ResponseCode.SELECT_TAG_PASS *For more specific code details, refer to the attached program documents.

24

Programming RFID Tags Under Dynamic Environment

3.4 Static Profiling Five different factors are being tested in this project: tag orientation, air gap, tag type, reader power, and the velocity. In those five different variables, eight angles (45 degrees increments), six air gaps (5mm increments from 5mm), 17 tag types, 5 reader powers (4.25 dBm increments from 10 dBm), and 4 different velocities (5cm/sec increments from 5cm/sec) were tested. Since testing for every possible combination is time-consuming and will result in a data set too huge to be analyzed effectively, it is crucial to somehow eliminate some combinations without actually testing them. This is why static testing, essentially a simplified version of offset testing, is conducted first. Using the data collected from running static tests for all combinations, the most promising combinations of variables will be selected. Also, note that the static and the offset testing test readability, not writability. This is because read commands take much less time to be executed, and the preliminary testing shows that they are very similar to each other. The only test that actually measures the writability of tags is the dynamic testing. 3.4.1 Static Testing Static Test tests how tag type, air gap, reader power, and tag orientation affect the tag handling. Since this test takes the shortest amount of time to run, it is used to eliminate variable combinations that behaves similarly or does not work at all. The Static Test collects 100 data points for every variable combination. The procedure for this test is as follows: 1. Raise reader Antennas to 5mm (+/- 1 mm) from the tag platform and securely clamp. 2. Align Tag on platform circle with “T” facing the fixture Tower to start -Tag tested 100 times for every power. 3. Rotate the tag and test all 8 orientations. 4. Repeat process for all six heights. 5. Repeat process for three tags of the same tag type. 6. Repeat process for 16 different tag types. Factors tested for this test were:

25

Programming RFID Tags Under Dynamic Environment

Table 5: Static Testing Factors 3.4.2 Offset Testing Although the Static Testing allows us to understand the basic behaviors of tags directly beneath the tower, it does not tell us how they behave anywhere else. And because dynamic tags stay under the tower for fraction of a second, the data collected from the Static Test is simply not enough. The Offset Testing solves this problem by conducting the Static Testing at different offset distances away from the tower. Because this results in the multiplication the number of tests for each variable combination, the Offset Testing is more time-consuming than the Static Testing. Using the result from the Static Testing, three tags (UPM “Button,” UPM “Dogbone,” and Avery “AD-805”, refer to the Figure 10) were chosen for this test (refer to the analysis section for more detailed discussion). The tags started from 20cm away from the tower to 10cm away from the tower (opposite side), at 1 cm increments. The procedure for this test is as follows: 1. Raise reader Antennas to 5mm from the tag platform and securely clamp. 2. Align Tag on platform circle with “T” facing the fixture Tower to start. -Tag tested 100 times for three different power levels 3. Repeat process for different orientations 4. Repeat process for heights 5-30 mm 5. Repeat process for all three tags (Button, Dogbone, and AD-805)

26

Programming RFID Tags Under Dynamic Environment

Factors tested for the offset testing are:

Table 6: Offset Testing Factors

Figure 10: Chosen Tags (from left): Button, AD-805, Dogbone

3.5 Dynamic Profiling Dynamic Testing introduces a new variable to the experiment- the velocity of the tag. It examines how increasing the speed at which the tag is being read affect the writability of tags. This test emulates what goes inside of a printer but using the MDrive stepper motor as the motor movement in a printer. Since read/write commands execution takes some time, the number of data points collected per one end-to-end motion of the linear slide is limited. Therefore the end-to-end motion is repeated until 500 data points are collected for each variable combination. Also, write command has to be used in this experiment, since from preliminary tests, it is observed that write and read commands behave differently when tags are being handled dynamically. The procedure for this test is as follows: 1. Raise reader Antennas to 10mm (+/- 1 mm) from the tag platform and securely clamp.

27

Programming RFID Tags Under Dynamic Environment

2. Align Tag on platform circle with “T” facing the fixture Tower to start. -Tag tested 500 times for three different power levels. 3. Repeat process for orientations chosen for each tags. 4. Repeat process for heights 15m and 25 mm. 5. Repeat process for all three tags (Button, AD-805, Dogbone). Factors tested for the dynamic testing are:

Table 7: Dynamic Testing Factors

28

Programming RFID Tags Under Dynamic Environment

Chapter 4: Results 4.1 Static Testing The purpose of this testing was to get the general idea of how different variables interact and affect the overall readability of a tag. From looking at the graphs, it is very obvious (conducting any statistical testing was impossible, for there were too many factors that were manipulated) that orientation, air gap distance, and reader input power affected the readability of the tags. 4.1.1 Orientation Results Contrary to the claims held by tag companies, the static test revealed that the orientation of the tag affected the readability tags. Although this was not always the case, in the most severe cases the change in the orientation of a tag caused it to vary the readability over 60%. In general, the orientation increased the readability the most when the largest surface area of the antenna was under the antenna. In general, tags’ reactions to the orientation were affected by the general shape of the tag. According to how tags behaved under varying orientations, three tag groups were identified: the Button Group, the Dogbone Group, and the AD-805 Group. The Button Group includes rotationally symmetrical tags such as the Button and Paperclip tags. The characteristic of this tag group is that the orientation did not affect their readability. For example, the maximum readability variation for the Button tag was less than 10% (Figure 11). Because the tag is rotationally symmetric and the general relationship between tag and the antenna did not change much from the rotations, it makes sense that the change in the orientation of the tag did not affect the interactions between the emitted electromagnetic waves and the tags.

29

Programming RFID Tags Under Dynamic Environment

Button Tag Average Error Rate 1 315

0

0.8

45

0.6 0.4 0.2

270

90

0

225

135 180

Figure 11: Button Tag Error Rate The second group, the Dogbone Group, includes tags that are more elongated, such as the Dogbone, Hammer, AD-223, and Webx tags. For tags in this category, the tag orientation sharply affected their error rates. For example, depending on the orientation of the tag, the error rates of the Dogbone tags varied anywhere from around 80% for 0 and 180 degrees to around 20% for 90 and 270 degrees (Figure 12). These variations in values are significant, especially when compared to the variation observed in the Button Group.

30

Programming RFID Tags Under Dynamic Environment

Dogbone Average Error Rate 1 315

0

0.8

45

0.6 0.4 0.2 270

90

0

225

135 180

Figure 12: Dogbone Tag Error Rate The last group, the AD-805 Group, is a group of tags that are neither too elongated nor too circular. These include tags such as AD-805 and Satellite tags. As one may speculate from the tags’ neutral shapes, the characteristic of this tag group is that orientation only somewhat affects the readability. Therefore, the graphs of the tags in this category exhibits symmetry, although they are not as polarized as that of Dogbone Group. In general, the variation in the error rate is somewhere between 20 to 40%. For example, the graph of AD-805 (Figure 13) exhibits a symmetry line 45-225 degrees, although the difference in the maximum and the minimum error rates are much less severe than it was for Dogbone (about 20%).

31

Programming RFID Tags Under Dynamic Environment

AD-805 Average Error Rate 1 315

0

0.8

45

0.6 0.4 0.2 270

90

0

225

135 180

Figure 13: AD-805 Tag Error Rate 4.1.2 Air Gap Results Unlike the orientation which affected the error rates differently for different tags, the air gap affected the error rate of the tags much more straightforward. In short, an increase in the distance between the tag and the antenna also increased the error rates, which matches with the hypothesis. However, this linear relationship between the error rate and the air gap did not hold true when the tags were 5mm away from the antenna. In general, the error rates are significantly higher when the tags are at 5mm away from the antenna. This indicates that 5mm setting is just too close for the RFID programming to take place. 4.1.3 Reader Input Power Results As it was the case for the air gaps, the reader input power also established a linear relationship between the error rates. As predicted the hypothesis, higher power resulted in a lower error rate. One interesting observation is that the 22.8dBm setting and 27dBm setting did not make any error rate difference. This indicates that the tested antenna is not constructed to make use of input powers above 22.8dBm.

32

Programming RFID Tags Under Dynamic Environment

4.1.4 Tag Uniqueness Results Three tags from each tag population was chosen for the test, since there is a possibility that each tag is unique enough to vary the error rates even within the same tag type. As it turned out, the error rates of the tags of the same tag types do not vary significantly (Figure 14). Most variations in the error rates amounted to less than 10%, which is not very significant. Although this was on contrary with the hypothesis, it is nevertheless good news, since this result indicates that any one tag could represent the whole population. Therefore, testing three different tags of the same tag type is not needed for the offset and the dynamic tests.

1 0.8

Dogbone Uniqueness Study

Error Rate

0.8

0.6

0.6

0.4

0.4 0.2

0.2

0 1

2 Tag Number

1 Error Rate

Error Rate

1

Button Uniqueness Study

3

0 1

2 Tag Number

AD-805 Uniqueness Study

0.8 0.6 0.4 0.2 0 1

2 Tag Number

Figure 14: Uniqueness Studies

33

3

3

Programming RFID Tags Under Dynamic Environment

4.1.5 Next Step The sole reason of the static testing was to identify what factors affected the error rates, so unaffecting factors could be eliminated from the later tests. From the information gathered on the tag orientations, three tags were selected to be tested in the offset testing- Button, Dogbone, and the AD-805 tags. Also, the Button tag will only be tested for 0 degrees, the Dogbone will be tested for 0 and 90 degrees, and the AD-805 will be tested for five orientations (0, 45, 225, 270, and 315 degrees). These decisions were made according to how they behaved under different orientations. As for the air gap, all 6 settings were kept. Finally, only one tag from each tag type will be tested in the offset testing, since the static test revealed that there is little to no variations among the tags of the same tag population. For hypothesis, it is expected that the tendencies discovered in the static testing (larger air gaps resulting in higher error rates, larger input powers resulting in lower error rates) to hold true in the offset testing as well. As for the offset distance, it is expected to affect the tags similar to how an air gap does.

4.2 Offset Testing Overall, the hypotheses were correct for this test. The air gap and the reader power continued to affect the error rates linearly, and the offset distance, the new variable, affected the error rates linearly. 4.2.1 Button Button tag continued to be the worst performing tag of the three, and it was especially vulnerable to offset changes. As the graph indicates (Figure 15), the only air gap that succeeded in making any communication between the tag and the antenna other than right beneath the tower was the 5mm. This was to be expected, since Button tags are near field tags, and thus only operates when the air gap is very small.

34

Programming RFID Tags Under Dynamic Environment

Offset Button Distance Study Error Rate

1 0.8

5(+/-)mm

0.6

10(+/-)mm

0.4

15(+/-)mm

0.2

20(+/-)mm 25(+/-)mm

0 -30

-20

-10

0

10

20

30

30(+/-)mm

Distance from the Tower (mm) Figure 15: Offset Button Distance Comparison Although it is hard to see because of the low success rate, the effect of reader power is also just as expected: as the power increased, error rate decreased (Figure 16).

Error Rate

Offset Button Power Study 1

10.0dBm

0.8

14.3dBm

0.6

18.5dBm

0.4

22.8dBm

0.2

27.0dBm

0 -30

-20

-10

0

10

20

30

Distance from the Tower (mm) Figure 16: Offset Button Power Comparison 4.2.2 Dogbone The trend for the input power was as expected, but an increase in the air gap did not increase the error rate as it has done in almost all of the previous tests (Figure 17). In fact, the air gap trend is completely

35

Programming RFID Tags Under Dynamic Environment

reversed- from -130 to -40mm, two low air gap settings (10 and 15 mm) have error rates that are about 20% more than the three higher air gap settings.

Offset Dogbone 90 Deg Air Gap Comparison Error Rate

1 0.8 0.6 0.4 0.2 0 -200

-150

-100

-50

0

50

Offset Distance (mm) 5(+/-)mm

10(+/-)mm

15(+/-)mm

20(+/-)mm

25(+/-)mm

30(+/-)mm

Figure 17: Offset Dogbone 90 Degrees Air Gap Comparison Just as it did in the static testing, the 90 degrees orientation performed significantly better than the 0 degrees orientation (Figure 18). An interesting observation about the graph in Figure 18 is that the offset distance did not affect the error rate linearly. Although the error rate decreases as the tag get nearer to the tower (when the offset distance is 0mm), there are two local minima, one around -120mm and another around 30mm, where the error rates are lower than its surroundings. The bar graph somewhat resembles a sinusoidal graph. This may be due of the fact that the electromagnetic waves emitted by the antenna propagate through the air in a sinusoidal path.

36

100

Programming RFID Tags Under Dynamic Environment

Offset Dogbone Orientation Comparison 1

Error Rate

0.8 0.6 0.4

0° 90°

0.2 0 -200

-150

-100

-50

0

50

100

Offset Distance from the Tower (mm) Figure 18: Offset Dogbone Orientation Comparison 4.2.3 AD-805 The tag trend for the air gap and the input power were as expected for the AD-805 tag. An interesting observation about the graph in Figure 19 is that there seems to be a “blind spot” for this tag. For four orientations out of five, error rates for the offset distance -10mm is significantly higher than they are for the rest of the offset distances. For all orientations except for 45 degrees, the error rate increases by about 30% that of its surroundings, which is very odd. But except for that -10mm oddity in error rate, the rest of the graph is what was expected- error rate decreasing as the tag gets closer to the tower.

The series of the chart below the bar graph (Figure 18) should make it easier to observe how the orientation affected the error rate. They are 360°radar graphs for offset distances between -30 to 30mm. As one can observe from the charts, there is not a convincing pattern that could be analyzed from themalthough the error rates of the orientations are similar in some parts, there are some orientations (ex. when the offset distance was -10mm) where the orientations affect the error rate seemingly randomly.

37

Programming RFID Tags Under Dynamic Environment

Figure 19: Offset AD-805 Orientation Comparison 4.2.4 Next Step After conducting the offset test, it became clear that the tendencies for the power input and the air gap are straightforward, linear trends. Therefore couple air gap and power input settings were eliminated in the dynamic testing. These eliminations include 5, 20, and 30mm air gap, 10 and 27dBm power settings. 5mm air gap was eliminated because far field tags behave unpredictably in NFC anyways, and 27dBm was eliminated because it usually only performed as well as 22.8dBm. The orientations were kept the same, since there were no significant performance overlaps between tag orientations.

4.3 Dynamic Testing This test introduced the final independent variable: velocity. The hypothesis for velocity was that increasing velocity destabilizes the RFID communication, increasing the error rate. In all cases, this hypothesis was correct. Increasing velocities did result in increasing error rate. Also, tendencies that were established in earlier tests (i.e. the power and the air gap settings) were conserved in the test for the most part as well, indicating that the velocity affected error rate independently.

38

Programming RFID Tags Under Dynamic Environment

4.3.1 Button Since Button tag was not performing well under static conditions, it was not surprising to find out that it continued to badly perform under dynamic conditions. As expected, the only condition that favored Button tag communication was when the tag was right underneath the tower (Figure 20). Referring to the graphs in Figure 20, it is clear that the power and the air gap tendencies were conserved even under different velocities.

Dynamic Button Power Study Error Rate

1 0.8 0.6

14.3dBm

0.4

18.5dBm

0.2

22.8dBm

0 -30

-20

-10

0

10

20

30

Distance from the Tower (mm)

Dynamic Button Velocity Study Error Rate

1 0.8 0.6

5cm/sec

0.4

10cm/sec

0.2

15cm/sec

0 -30

-20

-10

0

10

Distance from the Tower (mm)

39

20

30

20cm/sec

Programming RFID Tags Under Dynamic Environment

Dynamic Button Air Gap Study Error Rate

1 0.8 0.6

10(+/-)mm

0.4

15(+/-)mm

0.2

25(+/-)mm

0 -30

-20

-10

0

10

20

30

Distance from the Tower (mm) Figure 20: Dynamic Button Power, Velocity, and Air Gap Comparisons 4.3.2 Dogbone Just like it was in the static and the offset testing, the 0°tag orientation performed worse than 90°tag orientation did (Figure 21). Also, the bar graph of the error rate in different offset distances away from the tower had multiple minima and maxima- just like it did for the offset testing.

Dynamic Dogbone Orientation Comparison

Error Rate

1 0.8 0.6 0°

0.4

90°

0.2 0 -200

-150

-100

-50

0

50

100

Offset Distance from the Tower (mm) Figure 21: Dynamic Dogbone Orientation Comparison As for the three independent variables in this test (i.e. air gap, reader power, and the tag velocity), the air gap and the reader power trends were conserved. Also, the tendency for the air gap was reversed (increasing air gap decreased the error rate), just like it was for offset testing (Figure 22). Although the

40

Programming RFID Tags Under Dynamic Environment

reversal of the tendency was expected from the offset testing, the reversal was more articulated than it was for the offset testing. Instead of smaller air gaps having only 20% more error rates, there are numerous occasions in which the smallest air gap setting (10mm) have error rates that double the error rates for the largest air gap setting (25mm).

Dynamic Dogbone 90 Deg Power Comparison 1

Error Rate

0.8 0.6 0.4 0.2 0 -200

-150

-100 -50 0 Offset Distance from the Tower (mm)

50

14.3dBm

18.5dBm

100

22.8dBm

Dynamic Dogbone 90 Deg Velocity Comparison 1 Error Rate

0.8 0.6 0.4 0.2 0 -200

-150

-100

-50

0

Offset Distance from the Tower (mm) 5cm/sec

10cm/sec

15cm/sec

41

20cm/sec

50

100

Programming RFID Tags Under Dynamic Environment

Dynamic Dogbone 90 Deg Air Gap Comparison 1

Error Rate

0.8 0.6 0.4 0.2 0 -200

-150

-100

-50

0

50

Offset Distance from the Tower (mm) 10(+/-)mm

15(+/-)mm

25(+/-)mm

Figure 22: Dynamic Dogbone 90 Degrees Power, Velocity, and Air Gap Studies 4.3.3 AD-805 The power input and the air gap tendencies were conserved for AD-805, and also the faster velocities increased the error rates. This tag was affected by velocity more than it did Dogbone. This may be because the AD-805 needs more induced emf to activate, making it harder to overcome the destabilization caused by the tag movement. The blind spot of -10mm could be observed in this test as well, except it is much harder to observe due to the overall increase in the error rate. The shape of the bar graphs was parabolic like it was in offset testing, with the absolute minimum somewhere around 10mm (Figure 23). Just as in the offset testing, there seem to be no pattern that governs the relationship between orientation and the tag performance.

42

100

Programming RFID Tags Under Dynamic Environment

Dynamic AD-805 Orientation Study 1

Error Rate

0.8 0.6 0.4 0.2 0 -100

-50

0

50

Distance from the Tower (mm) 0°

45°

225°

270°

315°

Figure 23: Dynamic AD-805 Orientation Study

43

100

Programming RFID Tags Under Dynamic Environment

Chapter 5: Analysis 5.1 Air Gap Analysis Unlike the orientation which affected the error rates differently for different tags, the relationship between the error rate and the air gap was much more straightforward. In short, an increase in the distance between the tag and the antenna also increased the error rates (refer to Figure 25), which matches with the hypothesis. This was expected, since as electromagnetic waves propagate through the air, they scatter, allowing less of them to come in contact with the tags. This could be explained by applying the fact that

dB  B dS . Since altering air gap does not affect anything about the antenna, it is reasonable to assume that the changing of the air gap does not affect the change in magnetic flux d B . But since the Gaussian



Surface increases as the air gap increases, the surface area dS increases as the air gap increases, resulting in the magnetic field received by the tag decreasing. According  to one of Maxwell’s equation (i.e. Faraday-Maxwell equation: V  

 d B ), a decrease in the magnetic flux would result in a decrease in the dt

induced voltage V. Since it would be harder for the tags to activate at under lower voltages, it is obvious

 result in higher error rates. why greater air gaps

44

Programming RFID Tags Under Dynamic Environment

Figure 24: Theoretical Representation of the Gaussian Surface

Figure 25: Static Testing Air Gap Study

45

Programming RFID Tags Under Dynamic Environment

5.2 Reader Input Power Analysis As it was the case for air gaps, the reader input power also established a linear relationship between the error rates. As predicted in the hypothesis, a higher power resulted in a lower error rate. This is nothing to be surprised of. Since higher power implies that the antennas emit more energy per unit of time, the average magnetic field received by the tags increase as well. According to the equation of the magnetic flux and the Faraday-Maxwell equation, an increase in the magnetic field results in more electromotive force being induced, making it easier for the tags to be activated. Therefore, it is reasonable that an increase in the radio power decreases error rates. This relationship between the radio power and the error rate is well known, and it is by no means a significant discovery. However, the observation of this relationship does confirm that the results obtained from this project and the previous researches do match up.

One interesting fact regarding input power is that the 22.8dBm setting and the 27dBm setting did not make any performance difference. Not only did they not make any difference, 22.8dBm setting often outperformed 27dBm setting. This indicates that the antenna was not designed to be operated at such a high input power. Therefore, it is highly likely that if the electromagnetic wave is measured in front of the antenna, the strengths of the electromagnetic waves are similar for 22.8 and 27dBm power settings. Finally, changing the input power conserved the general shape of the tag orientation graph (refer to Figure 27), indicating that the orientation and the input power affect the tag performance independently.

46

Programming RFID Tags Under Dynamic Environment

Figure 26: Static Testing Power Study

47

Programming RFID Tags Under Dynamic Environment

Figure 27: Interaction Between Power and Orientation

5.3 Tag Uniqueness Analysis Three tags from each tag populations were chosen for the test, since there is a possibility that each tag is unique enough to vary the error rates within the same tag. As it turned out, the variation of the performances from the same tag types was not significant. Although this was on contrary with the hypothesis, it was nevertheless good news, since this result indicates that any one tag could represent the whole population. Therefore, testing three different tags of the same tag type is not necessary for the future tests.

48

Programming RFID Tags Under Dynamic Environment

Figure 28: Static Testing Tag Uniqueness Study

5.4 Tag Type Analysis As expected in the hypothesis, near field tag (i.e. Button tags and Paperclip) did not work when air gap was too big. Also, far field tags worked even when the air gap was very small. However, it was observed that the far field tags behaved very differently when the air gap was very small (around 5mm). This was especially evident for the relationship between the orientation and their performances. This is attributed to the fact that far field tags are optimized to work with the electric field, not the magnetic field. However, the results suggest that even if there is a size confinement inside of a printer such that only near field communication could be established, far field tags as well as the near field tags could be programmed.

49

Programming RFID Tags Under Dynamic Environment

5.5 Orientation Analysis Although RFID tag companies claim that the orientation along the area vector of the tags do not affect their readability, the results showed otherwise. In fact, tag orientation affects error rate as much as, if not more than, variables such as power levels and air gaps. As in a case with Dogbone tags, choosing the wrong orientation could result in an error rate that is triple the optimal orientation (Figure 29, while the worst orientation, 0 degrees, maintain an error rate of around 83%, more than triple that of the best orientation, 90 degrees). This could seriously rewrite the tag specification sheet published by the Alien Technology, since it claims that the preferred orientation for the Dogbone tags are 0 degrees and 180 degrees, two of the worst performing orientations as discovered in the experiment.

Another things discovered in the experiment is that less circular the shape of a tag is, more likely it is for them to be more reactionary to the orientation. As can be seen from the Figure 30, Button tag, the most circular tag, has consistent error rates from one orientation to another, while Dogbone, a very elongated tag, is affected by the orientation very much. This trend was observed for all of the tags tested in the static testing, as well as offset and dynamic tests.

Dogbone Average Error Rate 1 315

0

0.8

45

0.6 0.4 0.2

270

90

0

225

135 180

Figure 29: Static Testing Dogbone Orientation Study

50

Programming RFID Tags Under Dynamic Environment

Figure 30: Tag Shape v. Error Rate Also, the worst performing orientations for the elongated tags were all when the elongated sides of the tags were perpendicular to the tag matrix. This is probably due to the antenna design. As mentioned before, the antenna used in this experiment is very unique in that its electromagnetic field drops significantly after a certain distance. Since there are little to no researches done on antennas this inefficient, the creation of this antenna was very experimental, and it is highly likely that this antenna has many problems and is not the optimal design. Therefore, it is possible that the intentional creation of inefficient tags, which involved the focusing of the radio waves so much that the waves get dramatically reduced to nothingness after a certain distance, caused the antenna to be vulnerable to the changes in orientations of the tags.

Although orientation affected the error rate independently from all of the other variables (refer to Figure 33: 0°performs much better than 90°, even after the introduction of velocity), it was affected by the offset distance very much (Figure 31, the original shape of the radar graph was not conserved when the tags were offset from the original location).

51

Programming RFID Tags Under Dynamic Environment

Figure 31: AD-805: Orientation v. Offset Distance

Figure 32: Velocity and Orientation

5.6 Offset Distance Analysis The hypothesis expected the offset distance to affect the error rate the same way the air gap did- an increase in offset distance would increase the error rate, since distancing tags away from the tag enlarges the Gaussian surface, resulting in less magnetic field going through the tag. Although error rate did increase as offset distance was increased, the error rate did not decrease uniformly. There were some “bumps” in the average error rate vs. offset distance graph. After analyzing what caused this, we realized that low air gap distances had an abnormally high error rates than expected, enlarging the average error rate. When air gap comparison graph (refer to Figure 34) was constructed, we surprisingly found out that for most offset distance settings, lower air gaps yielded much higher error rates than that of higher air

52

Programming RFID Tags Under Dynamic Environment

gaps. This was especially evident in dynamic testing, when it was observed that 10mm air gap performed much worse than 15mm air gap, contrary to the trends observed in the static testing. This reversal of the air gap trend occurred since offset and air gap distances together determine an important factor- an angle at which the electromagnetic wave comes in contact with the tag (Figure 33). This angle of attack affects the amount of magnetic flux that goes through a tag, since the equation of the magnetic flux is dB  B dS , where dS  Acos (A is the area of the tag, and  is the angle of attack).

 was not taken in account of in the static testing, since it is always 0 degrees for the tag is directly under the antenna. However, as thetags move away from under the antenna,   value changes, and so does d B . 



Since cosine is a decreasing function from 0 to π/2, with its value determined by the values of air gap and the offset distance according to the equation  

  Air Gap    tan 1 , an increase in air Offset Distance 2



gap decreases error rate. Therefore, when the air gap is increased under condition where tags are not

 in the angle of attack opposes the change in the size of directly underneath the antenna, a change Gaussian surface. Since the result indicates that an increase in air gap reduces the error rate, it is concluded that the change in angle of attack affects the error rate more than the change in the size of the Gaussian surface does (Figure 34). However, this relationship only holds true for tags such as Dogbone, which could establish communication at far distances. Tags such as Button and AD-805 cease to communicate before a larger air gap becomes beneficial. Another observed phenomenon regarding the offset distance is the existence of “blind spot”- a region along the linear slide where the error rates are abnormally higher than the regions surrounding them (Figure 35). The exact location of the blind spot varied from one tag to another. The blind spot of the AD805 was located at -10mm from the tower, while it was -40mm for the Dogbone. In both cases, the error rates for that specific region spiked up. This implies that the antenna is faulty in that it creates a region of low electromagnetic fields in middle of high electromagnetic fields. Since this project did not study the behavior of the antenna, this would have to be confirmed by others.

53

Programming RFID Tags Under Dynamic Environment

Figure 33: Angle of Attack

54

Programming RFID Tags Under Dynamic Environment

Figure 34: Offset Distance Graphs

Figure 35: Blind Spot For AD-805

5.7 Velocity Analysis As expected in the hypothesis, velocity did affect the tag performance negatively. This is because the tag movement destabilizes the RFID communication. It is also important to note that velocity affected error rate independently from the other variables, as could be seen from the Figure 36.

55

Programming RFID Tags Under Dynamic Environment

Figure 36: Velocity Studies

56

Programming RFID Tags Under Dynamic Environment

Chapter 6: Conclusion The purpose of this project was to determine how do different factors affect the RFID programming success rate. From conducting the three tests, the following were observed:

1. Increase in reader power decreases the error rate 2. Increase in air gap generally increases the error rate, unless a tag is strong enough to be activated at smaller power input. In that case, greater air gap could result in lower error rate, especially at regions offset from the antenna. 3. Tag uniqueness is almost non-existent. 4. Tag orientation does affect the error rate, and the effect of the tag orientation becomes more contrasting as the shape of the tag becomes more elongated. Rotationally similar shapes, such as a circle and a square, are less affected by the tag orientation. 5. In general, an increase in offset distance increases the error rate. However, there are “blind spots,” or regions where antenna-tag communications are very hard to be established. 6. Increasing velocity causes the error rate to increase, under any conditions. 7. It is possible to program far field tags through near field communication, but not vice versa.

Another objective of this experiment was to learn how tags behave under dynamic environment. From conducting dynamic testing, it was very clear that increasing velocity caused error rates to increase as well. However, the velocity affected some tags more than the others. This indicates that some tags are better at coping against the destabilization caused by the tag movement.

Choosing the three tags (Button, Dogbone, AD-805) out of 16 tags proved very helpful, because the three exhibited very different characteristics throughout the tests, without having to test every single tag, which would have been time consuming. Button tag, although its error rate was often 1, allowed an analysis on

57

Programming RFID Tags Under Dynamic Environment

how near field tags behave under different conditions. From testing the Button tag, it is now very clear that programming near field tags in dynamic condition is not a good idea. Dogbone tag allowed an analysis on how long-range far field tags behave under different conditions. From conducting tests on it, it allowed me to realize that if the error rate does not simply decrease around directly under that antenna, but it decreases in a sinusoidal fashion around the antenna. Also, the fact that the increasing air gap could reduce error rates would have been overlooked if it were not for testing on Dogbone tags. Finally, AD805 allowed an analysis of tags that are somewhere between the Button and the Dogbone tags- tags that are most likely be used in laser printers in the near future. From looking at AD-805, I learned that oddities such as blind spots could exist in each tag.

If I were to revisit the project, I would fix one thing about the project. Because half of the tag orientations were eliminated for AD-805, it became very unclear as to whether it conserved symmetry in offset and dynamic tests. So it would have been better if all of the tag orientations were tested for offset and possibly dynamic tests.

Although this may be beyond the scope of this paper, it is my hope that the data gathered in this paper could be used to create a laser printer that can program multiple tags under dynamic environment in the future.

58

Programming RFID Tags Under Dynamic Environment

ACKNOWLEDGEMENTS I would like to show my appreciation to my mentor, Dr. Johne Parker, the graduate student Kassy Lum, and Lexmark representative Donnie Profitt. I also would like to thank the Math, Science, Technology Center at Paul Laurence Dunbar High School, and Mrs. Smith for providing me with this special opportunity. Finally, I would like to thank my parents and friends for always supporting me.

59

Programming RFID Tags Under Dynamic Environment

REFERENCES D’Mello, S. D., Mathews, E., McCauley, L., & Markham, J. (2008). Impact of position and orientation of RFID tags on real time asset tracking in a supply chain. Journal of Theoretical and Applied Electronic Commerce Research, 3(1), 1-12. doi: ISSN 0718-1876 EPC Global. EPCglobal Tag Data Standards Version 1.3. s.l. : EPC Global Inc., 2006. Georgetown University Law Center and The Constitution Project. Webcast--Video Surveillance: Legal and Tecnological Challenges. Georgetown University Law. [Online] March 23, 2004. http://www.law.georgetown.edu/webcast/eventDetail.cfm?eventID=33. Landt, J. (2005). The history of rfid. IEEE Spectrum, 8-11. RFID Journal. Mojix Wins First-Ever "Best in Show" Award at RFID Journal LIVE! 2008. RFID Journal. [Online] April 23, 2008. [Cited: November 25, 2011.] Wyld, D. C. (2005). The right frequency for government. Hammond: IBM Center for the Business of Government. 2006. New SkyeTek Protocol V3 Messages (pp. 1-25). Westminster: SkyeTek Inc. 2008. In SkyeModule M9 Reference Guide Version 080527 (pp. 1-132). Westminster: SkyeTek Inc. 2006. In SkyePlus MXH and MXU Multiplexer Reference Guide (pp. 1-45). Westminster: SkyeTek Inc. 2008. In MCode Programming and Software Reference (pp. 1-120). Marlborough: Intelligent Motion Systems, Inc. 2008. In MDrive 17 & 23 Plus Motion Control Integrated Motor and Driver (pp. 1-154). Marlborough: Intelligent Motion Systems, Inc.

60

Suggest Documents