This post will contain the results of all experiment done with the metal trolley.
Cardboard test results
Plastic test results
Glass test results
Investigating how easily tags can be read outside of the trolley
Investigating the benefits of RFID tags oriented in two axes instead of just one
Measuring the field strength around the trolley
The project I am working on at the moment involves testing how well RFID tags can be read when they are placed in a metal trolley. There are several factors that influence whether a tag can be read or not. The most important ones for this test are orientation, the material of the product they are attached to and where the antenna is mounted on the trolley. Due to the number of different materials that I am going to be testing with RFID tags there will be a lot of data and graphs.
To make sense of the graphs and data, some reference pictures are required. These outline how the height and position of a tag are referred to.
The tag is tested at three different heights:
Above each graph it might say something like ‘tag could not be read at position 6’. The picture below would be used to know what that refers to.
On each graph, every line refers to a position in the trolley. To know which line corresponds to a certain position, use the picture below.
In each page there will be several antenna positions, and for each antenna position there are six sets of tests done. These test are as follows, for each height, there are two tests done with the tags, vertically polarised ( standing up in the trolley) and horizontally polarised (lying down in the trolley). Three heights with two tag orientations per height means six different tests per antenna position.
When I refer to vertical and horizontal orientations this is what I mean.
Vertical orientation ( tag is facing the handle bar of the trolley ) :
Horizontal orientation:
Material test : Cardboard
How well can carboard based products be read with an RFID reader when placed in a metal trolley?
This is the box used:
Reference Pictures:
Results:
Antenna position : handle mounted
Height 0:
Tag vertical:
Tag horizontal:
Height 1 :
Tag vertical : (tag not read at position 6)
Tag horizontal:
Height 2 :
Tag vertical:
Tag horizontal:
Antenna position : mounted inside trolley
(Inside the the trolley)
Height 0 :
Tag horizontal (could not read tag at position 3):
Tag vertical:
Height 1 :
Tag vertical:
Tag horizontal:
Height 2 :
Tag vertical:
Tag horizontal:
Antenna position : mounted horizontally on front of trolley
Height 0 :
Tag vertical:
Tag horizontal:
Height 1 :
Tag vertical:
Tag horizontal:
Height 2 :
Tag vertical:
Tag horizontal:
Material test : Plastic
This is the plastic bottle used:
Reference Pictures:
Results:
Antenna position : handle mounted
Height 0:
Tag vertical:
Tag horizontal (tag not read at position 7):
Height 1 :
Tag vertical :
Tag horizontal ( tags not read at positions 3 and 7):
Height 2 :
Tag vertical:
Tag horizontal:
Antenna position : mounted inside trolley
(Inside the the trolley)
Height 0 :
Tag vertical:
Tag horizontal:
Height 1 :
Tag vertical:
Tag horizontal:
Height 2 :
Tag vertical:
Tag horizontal:
Antenna position : mounted horizontally on front of trolley
Height 0 :
Tag vertical:
Tag horizontal:
Height 1 :
Tag vertical:
Tag horizontal:
Height 2 :
Tag vertical:
Tag horizontal:
Material test : Glass
This post is a set of results of a particular experiment, the post available here explains what is going on.
This is the glass jar used in this test:
Reference Pictures:
Results:
Antenna position : handle mounted
Height 0:
Tag vertical:
Tag horizontal:
Height 1 :
Tag vertical (tag not read at positions 3 and 7):
Tag horizontal:
Height 2 :
Tag vertical:
Tag horizontal:
Antenna position : mounted inside trolley
(Inside the the trolley)
Height 0 :
Tag vertical (tag not read at position 2):
Tag horizontal:
Height 1 :
Tag vertical:
Tag horizontal:
Height 2 :
Tag vertical (tag not read at position 1 and 6:
Tag horizontal:
Antenna position : mounted horizontally on front of trolley
Height 0 :
Tag vertical:
Tag horizontal:
Height 1 :
Tag vertical:
Tag horizontal:
Height 2 :
Tag vertical:
Tag horizontal:
Investigating how easily tags can be read outside of the trolley
There is a problem with the RFID trolley idea involving accidentally reading tags from other trolleys passing by and reading products that are still on the shelves. We only want products to be detected when they are placed in the trolley. In this experiment I will test each of the three antenna positions with transmit power levels starting at 30dB to see how far away a tag can be read outside of the trolley. I will decrease the power until the tags can only be read inside the trolley.
The experiment setup:
The trolley remains in the same place for the entire experiment. The table is moved away from the trolley in steps of 50 cm. Wherever the antenna is placed on the antenna, there are only three directions that need to be tested. In front, to the left and to the right of the antenna. There is very little energy transmitted behind the antenna. The table is 0.9m high. The tag was in the vertical position for the duration of the experiment.
Results:
Antenna mounted on front of trolley like this:
30dB (1 Watt transmit power):
20dB (100 mW transmit power):
10dB (10mW transmit power):
Antenna was then mounted inside the trolley like so:
30dB (1 Watt transmit power):
20dB (100 mW transmit power):
Now the antenna is mounted like this:
30dB (1 Watt transmit power):
20dB (100 mW transmit power):
10dB (10mW transmit power):
Conclusions:
When the antenna is mounted outside the trolley the transmit power needs to be reduced to around 10dB to achieve acceptable leakage levels. If the antenna is mounted inside the trolley it can transmit at 20dB and leak very little energy outside the trolley. Inside mounting seems to be the best option in terms of how well the transmit energy is confined to the trolley.
Investigating the benefits of RFID tags oriented in two axes instead of just one.
This is a tag orientated in one axis:
This is a tag oriented in two axes: 
Passive RFID are very hard to read if they are in close proximity to metal. In the case of the trolley, if a product had a tag on only one side and this side happened to be touching the metal of the trolley then the tag could not be read. However if the tag is orintated in two axes at once, then no matter how the product is placed in the trolley there will always be a portion of the tag that is not touching metal. For an example of what I mean, look at this image: 
The question is, by how much does this improve the success rate of reading all products in the trolley? To find out I wrote a Python script to work out the average read success rate of each method. Firstly I fill up the trolley with all of the products, then the script is run and the sucess rate is found. This is repeated two more times to get an average sucess rate. This was done once with the tags on only one axis and once with the tags spanning two axes. Here are the results:
Tags in one axis only: 
Tags in two axes:
The results are surprisingly similar. It might have been by chance that none of the straight tags were touching metal. I was expecting the bent tags to have a greater success rate. I will retry the experiment and see if I get the same results.
Measuring the field strength around the trolley
Using a broadband field meter, I measured the electromagnetic field strength surrounding the trolley. This information is necessary when designing a product that emits radio waves and operates in close proximity to humans. Here are the three positions that I tested:
Results (Transmit power: 500mW):
Position A : 12.36 V/m
Position B: 2.12 V/m
Position C: 4.19 V/m
