RF Locator


Lost Item Locator 

 

Project Summary

     The proposed project is a wireless locating device that locates transponders attached to frequently lost objects of all sizes.  The locating device will interface with a cell phone, which will be running an application that displays signal strength/separation distance between the transponder and reader as well as allows the user to select an item from a stored library of items that the reader must search for. 

 

 

Introduction

 

Keys, cards, glasses, remotes, and many other household objects are misplaced on an almost daily basis, resulting in the waste of many precious minutes or hours in the search process.  Currently, a low cost and convenient solution to this problem is lacking in the market. 

 

The Lost Item Locator Project aims to solve this common and so far inadequately addressed problem.  Existing solutions to this problem typically have large transponders, limiting the items to which they could be attached.  Additionally, existing locators are prone to misplacement themselves, as they would be used only a few times a week.  The way in which our project will solve the former concern is by minimizing the transponder size.  A tag with minimal thickness and width will expand the list of locatable items, introducing individual cards and eye glasses to this list.  The way in which the latter issue that exists with existing solutions will be addressed by our device is the fact that our locator will interface with a cell phone while permanently residing within the phone's case.  As cell phones typically physically stay with an individual for most of the day, the risk of misplacing the phone, and thus locating device, is greatly reduced.  These, among others discussed later in the project description, are two important ways in which our device will be differentiated from other similar devices in the market. 

 

 

Background
     
     
Below are links to various sites we have visited for research purposes:
     Websites on Antenna Theory and Design

http://www.antenna-theory.com/basics/main.html   
     Provides an overview of basic antenna theory
http://www.tscm.com/radiapat.pdf
     Provides descriptions and radiation patters of various antenna designs
http://www.ece.msstate.edu/~donohoe/ece4990notes5.pdf
     Theory of loop antennas
http://www.angelfire.com/mb/amandx/dipole.html
     Provides dimensions of antennas operating at various frequencies
http://www.hamuniverse.com/dipivcal.html
     Information about dipole, inverted v, and ground plane antennas
http://www.mapability.com/ei8ic/rhombic/text.php
     Rhombic antenna design
http://www.w8ji.com/rhombic_antennas.htm
     Rhombic, v-beam, and inverted v antenna design
     Books on Antenna Theory and Design

Chen, Ning, and Michael Y.W. Chia. Broadband Planar Antennas: Design and Applications. West Sussex: John Wiley and Sons Ltd., 2006.

 

Elliott, Robert S. Antenna Theory and Design. Englewood Cliffs: Prentice-Hall, Inc., 1981.

 

Jasik, Johnson and. Antenna Engineering Handbook. New York: McGraw-Hill, 1961.

 

     Ultrathin Lithium Batteries

http://www.powerstream.com/thin-lithium-ion.htm 
     Websites on RFM transceivers
http://jeelabs.net/projects/cafe/wiki/RF12  
     Provides a library of sample code to use with the RFM12B transceiver
http://www.hoperf.com/upfile/RFM12B.pdf 
     This is the datasheet for the RFM12B
http://www.sparkfun.com/datasheets/Wireless/General/RFM22B.pdf
     This is the datasheet for the RFM22B-S2
     There are many devices that attempt to solve the above discussed problem, but they all seem to have significant limitations.  We will attempt to construct a device that will lack the drawbacks of the currently existing solutions.  Below are several existing item locating devices.
 
     Click 'n Dig Key Finder 
     Click 'n Dig Key Finder Wireless RF Item Locator Remote Control Pet Wallet Keyfinder. Free Extra batteries. The Click 'n Dig Key finder consists of a radio signal transmitter and four tags that can be hooked onto key chains.  The transmitter sends out a unique radio frequency signal for each tag, and if the tag detects its corresponding signal, it outputs a beeping sound.  The range of this device is claimed to be up to 80 meters.  This device has three obvious limitations.  First, the target list of items of this device is extremely narrow.  The receivers are designed to be hooked onto key chains, so the only items that can be located by this device are keys and other objects that have something like a key ring.  A second drawback is the relatively large size of the receivers, which would add bulk to smaller items like keys and wallets, restricting the storing options of those items.  The third drawback is that the device requires the user to rely on his/her hearing to locate the lost item, as the the receiver merely outputs a beeping noise when searched by the locator.  If the lost item is buried under cushions or in another room, the beeping sound may be muffled to the point of being inaudible. 
 
EZ-FIND! 25 Item Wireless Electronic Locator
 The EZ-FIND wireless locator is a more technologically advanced solution to misplaced items.  This device allows the user to program into it a library of up to 25 items it can search for and offers a digital display screen.  A drawback with this device is the sizes of both the locator and tag.  The large sizes make storage more difficult as well as limit the type of items that can be used with this device.  Furthermore, this device lacks the ability to guide effectively when searching for a misplaced item, as the tags are only capable of outputting a sound. 
 
Essentially, all existing item locators, though they may offer different features of convenience and levels of aesthetic appeal, are the same, having similar drawbacks and lacking important features.  But the single most uniting drawback of currently existing solutions is the fact that the locating devices are just as prone to misplacement as the household items it is meant to find. 
 
  
Intellectual Property
Below are listed links to several patents for devices similar to the project we are working on.

http://www.google.com/patents?id=hXyGAAAAEBAJ&printsec=abstract&zoom=4#v=onepage&q&f=false

 

Similar Features

                Consists of wireless tags that are to be attached onto personal items.

                Features a signal transmitter/control that allows an individual to select which tag he/she is interested in finding

Dissimilarities

                The tag outputs an audio signal when in range of the transmitter, which one listens for in order to locate the tag.

                Reader does not interface with a cellular phone.

---------------------------------------------------------------------------------------------------------------------

 

http://www.google.com/patents?id=__vWAAAAEBAJ&printsec=abstract&zoom=4#v=onepage&q&f=false

 

Similarities

     The device described in this patent sends an activating signal to a tag.

     The signal from the tag is then analyzed, alerting the user if the correct tag has been detected.

     The item locator is programmable by the user.

     The wireless signal of the locator could be initiated by a cell phone.

 

Dissimilarities

     The reader outputs an audio signal, vibration, or light to alert the user that the tag was found.  

 

---------------------------------------------------------------------------------------------------------------------

 

http://www.google.com/patents?id=EzSwAAAAEBAJ&printsec=abstract&zoom=4&source=gbs_overview_r&cad=0#v=onepage&q&f=false

 

Similarities

     The patent above gives a vague, conceptual description of a device that can identify and locate numerous tags using electromagnetic radiation.  

     The tags that the device locates can be attached to various large and small objects.

 

Dissimilarities 

     The device described would have voice recognition capabilities.  

     The tag would have to be waterproof as well as resistive/impervious to shock, oil, dirt, gas, chemicals, and high pressure.

     Device is 

 

Proposed Solution 

 

Our device will be the solution to many of the limitations faced by the existing devices in the market.  To immensely reduce the likelihood of misplacing the locator itself, our locator will essentially become part of one's cell phone.  Because cell phones are typically kept on one's person for the majority of the day, they are rarely misplaced.  Therefore, embedding the locating device into a cell phone's case will protect it from misplacement. 

 

To address the problem of effective guidance during the searching process, we will also create an application for either the Android or Apple platforms that will provide the user with a plethora of important features as well as equip the locator with a directional antenna.  The application will have a realtime display of signal strength that may also include an estimate of the distance between the locator and tag.  This will allow the user to continually adjust where they are searching for their lost items by letting them know if they are getting closer or farther from the item.  The directional antenna will provide further guidance by letting the user know which direction in 3D space the tag is located from the phone.  These two features of the device will allow the user to quickly and efficiently find their items whether they are in the same or separate room. 

 

An additional advantage of our device will be the small size and method of attachment of the tags onto the items of interest.  Their small size and adhesive property will allow the user to easily attach them onto credit cards, glasses, and various other thin, small items.  This avoids the problem of having to find a place on the item onto which to hook the tag, which is a feature of most item locators on the market. 

 

 

 

Requirements

 

Reader

      -Will interface with a cell phone, which will utilize an application that can search for and identify different transponders and give a real time display of the signal strength of the transponder

     -Will guide the individual to within half a meter to the item that is being searched in less than two minutes, given a room sized searching area.

 

Transponder

     -The size of the transponder must be at most 2 in. by 2 in., with a thickness of no more than a half of an inch.

     -Will be detected at a minimal distance of 15 feet from the receiver. 

     -The battery of the transponder must last for at least 100 hours of normal operation, or else be rechargeable. 

 

First Prototype

 

 

     Parts list:

 

          Jeenode v5 kits

               -RFM12B radio transceivers

               -ATmega328p processor

          Surface mount jee boards

          USB Bub

          RFM22b-s2

          Arduino pro

          FTDI cable

 

     Prototype and Design Process Description

 

     The project started out with the purchasing of several Jeenode v5 kits and USB bub boards.  The components of these kits included an RFM12B radio transceiver, an atmega328p microcontroller, and various other circuit elements.  The pieces were soldered to the jeenode printed circuit boards, resulting in two assembled jeenodes that were preprogrammed to transmit and receive wireless data.  After the jeenodes were assembled, the USB bub boards were constructed.  These bub boards were necessary to interface the jeenodes with computers.  Wireless communication between the the two jeenodes was established.  

 

 

 

 

 

     As can be seen in the above pictures, these assemblies are fairly large.  Not only are they too long to be utilized as transponders, they are too thick.  A first step to size reduction was to replace the original jeenode v5's with surface mount jee boards.  Though these new boards did not reduce the length of the device, they are significantly flatter.  

 

 

 

     The next step in the project was to acquire signal strength information from one of the transmitters.  The team, therefore, went about finding a way to extract analog signal strength information from the RFM12B radio transceiver.  As there was no arssi (analog received signal strength indicator) pin on the transceiver, we soldered a wire on one of the leads of the capacitor circled in the figure below, as was instructed in the website from which the picture was acquired, and measured the AC voltage on the capacitor with respect to the jeenode's ground pin.  The observed AC voltage was proportional to the amplitude of the received signal.

 

ARSSI Connection                                                              Taken from http://blog.strobotics.com.au/2008/06/17/rfm12-tutorial-part2/

 

     We encountered two problems with the above method of extracting signal strength data.  First, the voltage on the capacitor changed very little for a very large separation distance between the two jeenodes.  Even with amplified multimeter leads, the voltage drop for a 20 meter change in separation distance was about 10 mV.  The second, and most troublesome issue with this method was the immense instability of the signal.  At any particular separation distance, the signal fluctuated within an interval of 20 or 30 mV, which made it that much harder to observe a response to changing separation distance.  This method of analyzing signal strength was eventually abandoned.  

 

The second attempt to acquire signal strength information involved utilization of the drssi (digital received signal strength indicator) pin of the RFM12B.  The drssi pin is initially turned off (takes on value 0).  When the signal strength exceeds a programmed threshold value, the pin turns on (takes on value 1).  The value of the drssi is read on the computer every time a packet of data a jeenode receives from the other jeenode.  In order to continuously monitor the signal strength, code was written to step up the threshold value whenever the threshold is exceeded.  Unfortunately, because the RFM12B has a range of about 200 meters and there are only five values the threshold can take on, the drssi pin has very poor resolution.  Therefore, using the drssi of the RFM12B to analyze signal strength is impractical, as one of the project requirements is good signal strength resolution in the 3-5 meter range.  

 

     To resolve this issue, RFM22b-s2 radio transceivers, which have, among many other useful features, rssi register, have been purchased.  This rssi register will simplify the signal strength extraction process.  The RFM22B also has a feature for adjusting the output power which means the radius of interrogation for the device will be adjustable. 

 

 

Second Prototype

 

Working with the RF22B-S2

The first steps in developing the the second prototype were assembling the RFM22B-S2 transceiver, microcontroller, and usb-bub and carry out tests to ensure that there was wireless communication between two transceivers. The  code that was used in the tests is linked here. The RFM22B is controlled by writing configuration settings in its registers through an SPI interface. The linked code tests the capability to write to registers and the ability to receive data. Images of the setup are shown below.

     Above (from left to right) are shown the usb-bub board, arduino pro board containing an atmega 328 microcontroller, and RFM22B-S2 transceiver.

 

The RFM12 was replaced by the RFM22 because the RFM22 offered us a wider range of functions and capabilities that were absent in the 12.  A very useful feature of the 22 is it's ability to provide signal strength information with very high resolution.  During testing, we witnessed the fact that the signal strength indicator is sensitive to changes of a few inches in the distance between two transceivers.  Furthermore, one can control the power that the transceiver outputs to the antenna.  This gives the user the ability to locate items that are both far away and nearby.  Outputting full power to the antenna allows the user to detect transponders that are up to 200 meters away.  When the transponder is near the reader, the user can decrease the output power in order to attain high resolution signal strength information and avoid saturation of the signal strength readout. 

 

Once the assembly of two RFM22's was complete, with a non-directional 17cm straight wire antenna attached to each of them, we started testing.  We hid the transmitting RFM22 in a ~ 25x15 ft^2 room and had one of the group members search for it, solely relying on the signal strength information outputted to his computer by the receiving RFM22.  In less than a minute, he was guided to within half a meter to the transmitter (with maximal power to the antenna), at which point the orientation of the wire antenna became important.  For more accurate guidance, a directional receiving antenna and a non-directional transmitting antenna would need to be designed.  For the receiving end, two other antenna designs were considered and tested: a rectangular loop antenna and a VEE antenna.  For the transmitting (which will become the tag), a circular loop antenna was considered.  These antennas are shown below with their radiation patterns.   





 

Antennas were designed for an operation frequency of 915Mhz.  For optimal efficiency, the legs of the VEE antenna were calculated to be 7.5cm each. The length of the rectangular shape of the antenna was calculated to be 11.2cm and the width 5.6cm.  Extensive testing revealed that the signal strength response to changes in orientation of the rectangular shaped antenna was more predictable than that of the VEE antenna.  As this antenna will eventually be part of the tag reader, the dimensions of the antenna were compatible with those of a standard smartphone. 

 

     Testing

Correlation of Signal Strength with Orientation

Several procedures were taken to test the antennas' directionality as well as the signal strength feature of the RFM22B-S2.  To test the directionality of the antennas, we constructed a circle having marks every 45 decrees, with 0 degrees pointing straight towards the transmitter.  The antenna-transceiver assembly was then placed on the circle and rotated every 45 degrees.  For each rotation, we measured the signal strength.

 

 

                                                                                          

 

We realized from this test that at long distances (greater than a few meters) from the transmitter, changes in the antenna's orientation does not give a very predictable signal strength response.  However, antenna orientation seemed to be important at near distances. 

 

We decided to run another orientation test, this time recording all our signal strength data.  At six different separation distances, signal strength values were obtained for eight different antenna orientations - loops facing each other, the receiver rotated 45 degrees, 90 degrees, etc.  We expected to see a peak signal strength response when the loops face each other, as the below illustration demonstrates.

 

 

 

          Below is the actual signal strength response to antenna orientation.  The signal strength doesn't seem to have at all a predictable response to antenna orientation.

 

 

 

 

 

 

 

Correlation of Signal Strength with Distance 

While there is a correlation between the signal strength and the separation between the transceivers, it has been difficult to characterize because the experiment environment affects the observed radiation pattern as the results linked here will indicate. Below are shown plots of the total signal strength of 50 consecutive values against the separation distance. For the first set of results, the experiment was conducted indoors and for the second set in open space.

 



As the graphs conclusively demonstrate, noise in the signal strength output is primarily a result of indoor reflection.  Therefore, in order for the device to function reliably indoors, the high frequency noise will need to be filtered out of the signal strength data.  

 

Several filtering techniques were implemented.  One of them was a fairly simple filtering method.  Rather than allowing the serial monitor to display each signal strength value, 30-50 signal strength values are collected, either averaged or summed, and then displayed.  Simply by not displaying the high frequency jumps and just displaying the mean gives a clearer representation of the received signal strength.  The other filtering technique operated on a similar concept.  An array of signal strength values are collected and the averages between two consecutive values are calculated.  It then puts out a visual representation of the data that is "smoother" than the original, unfiltered data.  Below are shown the effects of this filtering method.

 

 

                   Testing the filtering code on a theoretical string of signal strength values riddled with noise

 

 

 

     The filtering code seems to work well theoretically.  The code was then implemented on an actual array of signal strength values.  Below is shown a plot of the unfiltered (dark blue) and filtered (light blue) signal strengths.  As can be seen on the graph, sharp peaks (high frequency amplitude jumps) are smoothed out and a more constant array of signal strength values is outputted by the filter. 

 

 

 

Problems in signal strength readout:

 

We discovered that 36 points of data were being taken each time we thought only 1 was. Not only does this kill the potential battery life and increase power consumption enormously, but it prevented data collection during those 35 remaining points. The large discontinuities for small samples of RF packets showed us there must be something wrong. The key observation was that the discontinuities always occurred after 36 data points (or some multiple of 36).





Minimizing tag size:

 

To minimize the tag size, we decided to look at the eagle files for the arduino pro and see what circuit components were not necessary for the operation of our tag.  The result of this was a significant decrease in size of our transmitter (about 1.5 x 1.5 in^2).  The circuit design was sent to PCB Pool to be printed.  To minimize thickness, the new boards were designed to be surface mount.  The circuit schematic is shown below.

 

 

 

 

Tools

 

Weekly Design Report

https://spreadsheets1.google.com/a/brown.edu/pub?hl=en&hl=en&key=0AvM6JwNF2XvrdGlsMWVfS3g1aE9NZnZxVHFldmgtcXc&output=html

 

Gantt Chart

https://spreadsheets2.google.com/a/brown.edu/pub?hl=en&hl=en&key=0AvJjWpOX-c8tdC1nVVd0eTN6RENWRTUwR09TY3RjNUE&output=html