Second test satellite is in orbit. I will not further comment on that. There’s already enough people talking (and silently laughing) about Galileo. Someone said it is a loser, the commentators (obviously Europeans) said it is not, and other said it is susceptible to military radar. Anyway, congratulations for the launch of Giove-B.
I saw the trailing text news of the recent Galileo deal in one of the German TV stations a few days ago, but just found the news in BBC today. Probably they need some time for a little mock up between the real event and the actual press release 
Finally, after the endless struggle among the companies involved in the consortium, resulting in the death of their Public Private Partnership (PPP) funding scheme, the project continues. Fully covered by European citizen’s tax. Taking spare money from agriculture funds. Until today, there is only one test satellite (Giove-A) in orbit, out of 30 planned, launched end 2005 (note: one of the purposes of its launch was to secure the frequency allocation for Galileo given by the ITU which would have been otherwise canceled should there be no satellite in orbit by 2006. And by the way, Giove-A was built by Surrey Satellite Technologies, SSTL, which is not a member of the original consortium)
Why do the European insist on building their own satellite navigation system anyway? With the emergence of Beidou from China, plan from Russia to re-strengthen their GLONASS, and the dawn of GPS III, competing for the “piece of the cake” of satellite navigation market, it is really questionable, whether the system is really worth the price. Accuracy will not be the factor that will determine the market dominance of any of the systems, as they will have approximately the same positioning accuracies.
Will Galileo be independent from GPS? This was another argument trying to justify Galileo. Since GPS is owned by US military, it can be shut down at their will, and considering that GPS is also used for aviation navigation, it may raise safety-of-life issue. Galileo will be a civilian system. However, to have a system with complete global coverage having ultimate accuracy, it is necessary to have GPS and Galileo combined, in a sense that a receiver should be able to process navigation signals from both systems. In other words, Galileo is not really independent, is it?
Just stumbled on this flowchart from this page … lol (no comment)
,----------------- Sparse and Infrequent -----------------,
| Observations |
\/ \/
Observational Incorrect Interpretation
Errors of Observations
| |
| |
| |
'--------------> Theoretical <--------------------'
Misunderstanding
|
|
|
Management <--, \/
Directives | Oversimplified <----------------,
| | Models |
'---------' | |
| |
\/ |
Computer Models <------- Controversy -----------> Further Refinement of
| | Unimportant Details
| |
\/ \/
Code Unrealistic
Errors Assumptions
| |
\/ \/
Crude Diagnostic
Tools -------> Confusion ------------> Further Misunderstanding
| |
| |
\/ |
Coincidental Agreement <-------------'
between Theory and
Observations
|
|
\/
Publication
For anyone who’s STILL using LaTeX, keep this in your mind:
LABEL MUST BE PUT AFTER CAPTION, OTHERWISE YOUR CROSS-REFERENCE NUMBER WILL BE WRONG!!
Read also this FAQ.
is one interesting phenomenon, where the “total” is different from (usually less than) the “sum of the individuals”. For instance, if in average there are 100 people entering a coffee shop in one day, it doesn’t mean that 100 counters are needed, since the chances are low that those 100 people came at the same time. Even if they did, most people could still accept some delay in the service.
It has a beneficial effect on economy, since we know that we can safe some money when buying a commodity to be used by many people. Examples ranging from shower gel to satellite transponder capacity.
It poses some problems, though, when a decision is to be made, i.e., in the design process. For instance, how many bottles of shower gel should be bought to be used by 4 people in the house so that we don’t need to buy more in 2 months, how much transponder bandwidth do we need to provide in-flight entertainment services in European region, etc. Simple algebraic add-subtract-multiply-divide rules do not apply here. We need to make some statistical assumptions regarding the processes involved. And sometimes this can be quite complicated.
Sandro Scalise was asking me if I would like to review a paper to be submitted into a journal, to which I am also submitting…
Of course I would say yes!!!
I am currently sitting in a hotel room in Salzburg, Austria, finishing and rehearsing a presentation for tomorrow, which is going to be my first “appearance” in an international conference.
But for God’s sake tomorrow is the first day of Ramadhan! And my session will be at 4 o’clock in the afternoon. That means tomorrow’s going to be tough. Not to mention that I only rely on INDOMIE for sahur; did not have enough time today to search for kebap shop nearby. Let’s hope that I will not pass out during the session; neither by exhaustion nor by sleepiness ;p
Ramadhan Mubarak people!!
I have just finished editing the paper for the conference in Rome, and notifying other co-authors. Suddenly I feel like I have no deadline now, and I thought writing a blog (again) is not at all a bad idea…
I haven’t been so enthusiastic in following social activities in München these days. No pimpong, no soccer, not so much involved in PM3, no Café München. Not like I don’t like the people or the activities, it’s just I have so many questions in mind, and so many things to learn, that I find it difficult to share the focus. That’s why I decided to write this blog entry, just in case someone would care to share his mind and simply talk about these stuffs.
Question no.1:
We’re questioning how much gain can be obtained in terms of savings in the required satellite capacity, if we use priority-queuing scheduler, compared to other “ordinary”
scheduling mechanisms, say, round-robin (RR). I have some idea on how to derive analytically the queuing delay in each queue for the RR, and later with small effort try to simulate it in Omnet++. It’s just a matter of allocating the time. Pushing the result to Eurocontrol or to the consortium of the EC project who funded this work is another, less-theoretical, matter.
Question no.2:
In another EC research project, we are trying to identify the issues involved in a system where we have network nodes having the capability of changing the communication links, whenever more than one link is available. From here several questions arise:
- Transport layer issue: TCP with its variants are currently the most prominent transport layer protocol in the Internet. Yet it screws when faced with high bandwidth-delay-product links such as the geostationary (GEO) satellite. There exist some solutions to optimize TCP performance for GEO satellite link. One solution is to split the network into 2 parts, the one with and without satellite link, and use the optimum transport protocol in each segment.
- The above solution results in other problem. First of all it implies that we have to violate the end-to-end principle of TCP. OK, you said, what’s the problem with that? well since we’re now breaking the connection at the transport layer, we are faced with the fact that we can not apply security and mobility solutions in the lower (IP) layer, namely IPSec and Mobile IP (or its variants). This is currently a hot discussion topic of the colleagues in DLR. I am not directly involved. But nonetheless I still need to understand all problems and all existing solutions in order to be able to judge whatever the conclusion of this discussion will be. That means hard work. I need to understand all rationale behind solutions made in IP security, IP mobility, transport layer design and improvement, etc.
Question no.3:
I am going to SatNex summer school in Pisa from the 27th to the 31st of August. There there will be some lectures that either I have received in my undergraduate/master but I have forgotten completely, or are completely new to me. Since being blank in the class is the last thing I’ve ever wanted, I need to review/learn a lot of things. I’ve broken down the topics into several modules, to each of which I commit myself to read as much as possible in one week time. It’s been running for several weeks now. Some topics are: detection+estimation+modulation theory, receiver synchronization, time-series analysis, and Markov decision process. That means, unfortunately, no afternoon-sight-seeing this month, sorry.
Question no.4:
We are doing some tests with our GSM software. A couple of days ago I have managed to install the software in a ruggedized PC, and make some phone calls and SMS. The used back haul link was still our good-old-friend the Internet, using VPN to our provider. Since we want to use BGAN satellite link, some questions arise:
- How is the PDP context in the BGAN terminal managed?
- Is GPRS supported?
- How to combine GSM with WiFi traffic without jeopardizing the voice QoS? (the colleague in DLR is pushing a lot on this)
- Will there be a room for implementation of positioning algorithms in the network? It’s something the student which I supervised has been working on all this time.
Well, that’s the main problems inside my mind these days. I still have some list of questions, like how security problem is solved in Mobile IPv6 route optimization, what is the “ultimate meaning” of the phrase “capacity dimensioning”, or how access, queuing, and routing fit together within the phrase.
That’s a lot of philosophical questions.
Anyway, have a nice weekend everyone!!
It’s an almost obvious statement which says:
“If there are N pigeonholes and N+1 or more pigeons, then at least one pigeonhole is occupied by more than one pigeon”
Easy, right? OK, now excercise: What is the minimum number of socks that should be taken from a shelve containing red, blue, and white socks, to ensure that you get a pair of socks with the same color?
The answer is four. In this case, there are 3 pigeonholes (one for each color), and to be sure of getting two socks (pigeons) with the same color (in the same pigeonhole), all we have to do is to grab 4 (= 3 + 1) socks from the shelve.
OK, bonus problem: Find the minimum number of elements that one needs to take from the set S = {1,2,3,…9} to be sure that two of the numbers add up to 10.
Source: Schaum’s Outline of Theory and Problems of Discrete Mathematics, 2nd edition.
–edit: it is also known as Dirichlet’s box principle–
In medical, the term “Zebra” is used as a metaphor for an obscure and unlikely diagnosis from ordinary symptoms. It’s like when your doctor diagnoses you as having a disease caused by a mosquito-bite from New Guinea, while in fact you’re just having a flu. The term comes from a saying which goes “When you hear hoofbeats, think horses, not zebras”. (For those who are interested, it’s just another form of Occam’s razor)
In my case, I had just run a Matlab simulation for global aircraft traffic, and I got a high peak number of aircrafts, higher than continental Europe, somewhere near the Antarctic. Then I said to myself, “Félicitations, mon ami, you’ve just had a zebra!” ;p
Recent comments