From: Athanasios A. <th...@at...> - 2005-07-14 09:28:08
|
Hello I was just thinking about the ape location system and there are some parameters that will probably make your life difficult. 1. You probably have a serious multipath environment which will make triangulation a pain (Not only triangulation, every kind of transmision) 2. The space is confined which means that small errors in the triangulation will introduce large errors in the position. 3. The problem is not about tracking ONE ape...it is about tracking N apes. So triangulation using a bluetooth device (non modified that is) is out of the question as you can not tell which transmiter is the one you are trying to triangulate (please correct my ignorance on this, if any) 4. Having the colars waking up and transmiting introduces the problem of colisions which gets more significant as the number of colars increases. 5. If you fill the place with a network of underground computers, the cost is going to be high for both installation AND maintainance.....which is ok if your budget allows it. 6. Puting a gumstix on the colar of the ape seems attractive to your application, so lets take it from there. Last week there was a discussion about navigation systems in the mailing list...The more i look into it the more i think it is feasible to have a small board of suitable accelerometers transmiting their signals to a gumstix which will be working out the position of the ape. Each individual gumstix can then be given a unique ID and transmit its ID and location using bluetooth or any other transmision method (there are a lot of things that can be done for robust transmision recepetion but that is not the point here) to a base station that will be keeping a list of ape IDs and positions. Something like JANE:45:33:67 MARK:46:32:89. The initial position can be set relatively to a known steady point and the initial calibration of position and unique ID can be done over a simple serial connection. When you need to change the battery you would have to recalibrate the device which would be required at some point anyway so as to "reset" integration errors in the position detection system. To address the problem of low power bluetooth modules there is a number of approaches that can be taken. For example, multiple antennas (instead of multiple receivers) scattered around the place. If the space is confined like a cage with bars on the outer perimeter it is worth the trouble of examining if the cage can actually be used as a loop antenna to a bluetooth receiver...For this, a small circuit is needed to adapt the impedance of the cage to the impedance of the receiver...Make sure that the two ends of the cage do not meet otherwise the antenna is short circuited...The whole area becomes the antenna then..... The movement of the ape is "robust", meaning that the ape will never move so slow as to trick the accelerometer and not record its movement. (Some accelerometers provide DC output proportional to the gravity field they can detect) Ofcourse, a 3 axis accelerometer along with a 3 axis gyroscope are needed to determine the 3D vector of movement and speed. An additional benefit of this would be a pattern recognition application that can detect the "mood" of the apes. If (and infact which) an ape is violent (his vector would be jumping up and down, back and forth, rolling, jagging, moving violently), or if an ape stays on one place all the time...if an ape has hurt its leg and has to stop periodically to get a rest and so on....Additional information about this can be obtained by additional accelerometers on the hands or legs of the apes so that their motion can be tracked.... This application does not aim in removing the human factor duties. I hope you find this useful. Best Regards. Athanasios Anastasiou Signal Processing and Multimedia Communications Laboratory University of Plymouth - UK. http://www.tech.plym.ac.uk/spmc/ |