jtiernan2008
asked on
obtaining sample figures for a DTOA/Multilateration system
I have a major project on locating system using RFID tag and four readers. However, the devices advised for me to purchase by supervisor is not capable of nanosecond resolution which is the requirement of this system. Basically the system is based on differential time of arrival. For example if the device is in the middle of the 4 readers the the DTOA will be 0 between all the readers however if I move the tag closer to one of the readers. The displacement of 1ft will be 1 ns as light travels ~1ft per ns. We cannot not work with ms either as 1ms = displacement of ~300m which is not useable. There is also the problem of jitter in the ms which thus introduces inaccuracies which trows the system out by metres.
So... as a result I have worked on the java software which can communicate with the readers via RS-232 -> USB (interface with the reader adding more jitter that cannot be resolved) and it stores the port, the beacon name of the rfid tag, and a crude time stamp that is placed with a Java system.nanosecond function call which returns a nanosecond time however does not guarantee any accuaracy.
My supervisor agrees that this is a good part of the project done but and has welcomed the idea that we have 2 databases. one database that is populated by the data from the readers to prove that the system communicates. And another database that can be populated by 'made up' but accuratly correct figures that software can use for the multilateration (intersection of hyperbolas).
What is the best way to make up these figures?
What would be the x,y,z co-ordinates of the readers, and how do I plot the hyperbolas.
Your advise would be really appreciated.
Thanks in advance.
So... as a result I have worked on the java software which can communicate with the readers via RS-232 -> USB (interface with the reader adding more jitter that cannot be resolved) and it stores the port, the beacon name of the rfid tag, and a crude time stamp that is placed with a Java system.nanosecond function call which returns a nanosecond time however does not guarantee any accuaracy.
My supervisor agrees that this is a good part of the project done but and has welcomed the idea that we have 2 databases. one database that is populated by the data from the readers to prove that the system communicates. And another database that can be populated by 'made up' but accuratly correct figures that software can use for the multilateration (intersection of hyperbolas).
What is the best way to make up these figures?
What would be the x,y,z co-ordinates of the readers, and how do I plot the hyperbolas.
Your advise would be really appreciated.
Thanks in advance.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
=>The side of the square should be less than the range of TX/RX pairs by sqrt(2).
Does this mean the squareroot of a figure less than the range of the TX? Why is this?
=> If the TX can move in 3D, you might opt for three bottom corners of a cube, and one on a top corner.
Why do you choose this layout?
=> Can I please confirm that this is calculated as follows
TOA = d/s
=> To simulate av1 GHz clock, round up to the next N ns.
What is the relationship between the speed of the clock and speed of RF?
Does this mean the squareroot of a figure less than the range of the TX? Why is this?
=> If the TX can move in 3D, you might opt for three bottom corners of a cube, and one on a top corner.
Why do you choose this layout?
=> Can I please confirm that this is calculated as follows
TOA = d/s
=> To simulate av1 GHz clock, round up to the next N ns.
What is the relationship between the speed of the clock and speed of RF?
If the RX locations are too far apart, some of them may not get any signal at all from
the TX.
No matter where you put three RX, they will define a plane.
Maybe it's okay if the 4th is also on a corner of the square. There may be ambiguity
about above and below ground, but you should be able to resolve this be default.
Time = Distance/Rate
The speed of the clock has nothing to do with the speed of light or RF.
But it does determine your time/spatial resolution.
the TX.
No matter where you put three RX, they will define a plane.
Maybe it's okay if the 4th is also on a corner of the square. There may be ambiguity
about above and below ground, but you should be able to resolve this be default.
Time = Distance/Rate
The speed of the clock has nothing to do with the speed of light or RF.
But it does determine your time/spatial resolution.
ASKER
=>The speed of the clock has nothing to do with the speed of light or RF.
But it does determine your time/spatial resolution.
=> To simulate av1 GHz clock, round up to the next N ns.
f=Hz per sec so Is this (1 second) divided by 1 000 000 000 = 1 ns.
I understand now :)
Does this mean that the resolution of 1 Ghz clock is 1 ms?
But it does determine your time/spatial resolution.
=> To simulate av1 GHz clock, round up to the next N ns.
f=Hz per sec so Is this (1 second) divided by 1 000 000 000 = 1 ns.
I understand now :)
Does this mean that the resolution of 1 Ghz clock is 1 ms?
ASKER
sorry ignore above..
I understand all now 1GHz = 1 ns resolution
thanks for your help
I understand all now 1GHz = 1 ns resolution
thanks for your help
ASKER