Down Time Down South
I have a little downtime before boarding here in Roanoke, as is usual. Carl was nice enough to drive me to the airport and we got here right on schedule at about quarter to 8. I gave myself an hour before my flight was scheduled to depart JUST IN CASE, though in all my experience with the Roanoke airport the process of checking in and getting through security has never taken more than 15 minutes. Plus, getting here early isn't that big of a problem since they have free wireless internet :).
Why don't bigger airports do that? Sure, Atlanta has wireless but it's through T-mobil or some other stupid company that charges you $10 for 1 hour or something.
This past week was pretty darn hectic. I actually was pretty good about getting into work at or before 9 every day (except the days I swam from 8-9:30 of course). I still went home for lunch and on occasion took a bit longer than I really intended and ended up having to stay at the lab until 9 on a few days.
The draft of my paper is at a pretty good stage I think. Stilwell said it's well on its way to being a good qualifier paper. It's going to need a little more to be accepted to the IROS conference. The work I've done now and show in the paper doesn't really concretely proof that our solution works well, it just says "this is what we did and it seemed to work in practice". Just to bring everyone up to speed, for those not familiar with my current research it involves cooperative localization which is the process of a robot figuring out where it is with the aid of information shared with other robots in the area. A particularly interesting challenge we are faced with is that we don't have a very good ground truth to compare our algorithms against. If we were dealing with land vehicles we could use GPS to determine how close our estimate was to the true position. Once the vehicles dive we have absolutely no way of knowing what path they are following, we just know the point the dived at and the point the surfaced at. We can use these two points, along with information about prop speed and fin deflection to make a good guess of where the vehicle was at any point in time during the dive but it's still just a guess. Recently we collected some real world data and performed two types of algorithms on it and both results were pretty close to our "best guess", but what is that really saying? Anything? We just have three guesses, one is a bit better than the others and we're using it as a reference.
Anyway, that's kind of the problem, in order to show that our algorithm really works well we'd like to come up with an expression that some how compares it to what the vehicle is REALLY doing.
I've just been rambling on a lot about that, I'm sorry. I have to do take my qualifier at the end of April so I'm going to be thinking about this stuff a whole lot over the next couple months.
The food at the little "fresh bakery" counter here smells really good, or maybe it's the egg on a croissant sandwich that this guy next to me is eating. I kind of want one but of course everything is overpriced.
I a few short hours I will be in Roatan, Honduras with the crew from Holyoke Underwater ready for a week of diving in tropical water. It's gonna be pretty damn nice. I will be taking many pictures of course and may even post a few online while I'm down there, I'm told there's free wireless internet at the resort.
I hope it's the type of place where I can get there, throw my bags in the room and jump in the ocean. I miss salt water.
Why don't bigger airports do that? Sure, Atlanta has wireless but it's through T-mobil or some other stupid company that charges you $10 for 1 hour or something.
This past week was pretty darn hectic. I actually was pretty good about getting into work at or before 9 every day (except the days I swam from 8-9:30 of course). I still went home for lunch and on occasion took a bit longer than I really intended and ended up having to stay at the lab until 9 on a few days.
The draft of my paper is at a pretty good stage I think. Stilwell said it's well on its way to being a good qualifier paper. It's going to need a little more to be accepted to the IROS conference. The work I've done now and show in the paper doesn't really concretely proof that our solution works well, it just says "this is what we did and it seemed to work in practice". Just to bring everyone up to speed, for those not familiar with my current research it involves cooperative localization which is the process of a robot figuring out where it is with the aid of information shared with other robots in the area. A particularly interesting challenge we are faced with is that we don't have a very good ground truth to compare our algorithms against. If we were dealing with land vehicles we could use GPS to determine how close our estimate was to the true position. Once the vehicles dive we have absolutely no way of knowing what path they are following, we just know the point the dived at and the point the surfaced at. We can use these two points, along with information about prop speed and fin deflection to make a good guess of where the vehicle was at any point in time during the dive but it's still just a guess. Recently we collected some real world data and performed two types of algorithms on it and both results were pretty close to our "best guess", but what is that really saying? Anything? We just have three guesses, one is a bit better than the others and we're using it as a reference.
Anyway, that's kind of the problem, in order to show that our algorithm really works well we'd like to come up with an expression that some how compares it to what the vehicle is REALLY doing.
I've just been rambling on a lot about that, I'm sorry. I have to do take my qualifier at the end of April so I'm going to be thinking about this stuff a whole lot over the next couple months.
The food at the little "fresh bakery" counter here smells really good, or maybe it's the egg on a croissant sandwich that this guy next to me is eating. I kind of want one but of course everything is overpriced.
I a few short hours I will be in Roatan, Honduras with the crew from Holyoke Underwater ready for a week of diving in tropical water. It's gonna be pretty damn nice. I will be taking many pictures of course and may even post a few online while I'm down there, I'm told there's free wireless internet at the resort.
I hope it's the type of place where I can get there, throw my bags in the room and jump in the ocean. I miss salt water.