This week, I worked on temperature testing and parallel testing on the electronic load. I did some research on the oscillation issue I was having with the electronic load and simulated the circuit in LTspice. I am still reading up on control system, stability and compensation resistor and capacitors. I still need to figure out how to receive the pH sensor data through the UART. Other than that, everything is good! I will be heading out on Rachael Carson tomorrow.

Week Five:

I was able to access rov ctd data on Monday and make a preliminary plot comparing the data to cruise ctd cast data.  On Tuesday, I looked at the calCOFI data of Line 67 which is off of Monterey Bay.  I was able to confirm the U-shape in oxygen at depth that others have found in data from S. Cal and from denitrification and other analyses.  This suggests that decadal variability is a fairly important component in fluctuations of oxygen on this coast.  On Wednesday, I tried to get a spatial regression plot going in one day of coding.  I currently have a headache.  I attended the weekly CeNCOOS meeting and met with Francisco Chavez; a "half point" check-in for the internship.  On Thursday, I will try to get my regression plot to work, and I will look at rockfish survey data, and I will tie up loose ends looking at the rovctd data.  (Or, I will complete at least 2 out of those 3 items).

I am at an exciting point in my stay here because I am transitioning away from data collection and moving towards data analysis including comparisons with model data, and model data analysis, which is "on schedule," but I am tired!

This week has come as a bit of a shock. With half of the internship over and lots to do left on my project it seems like it has come too soon. Luckily I am putting the first sensor on C1 today so i will be able to begin collecting data. Sensor number 2 will hopefully be in the water by the end of week 6. I am interested to see how soldering goes one handed, but i am sure it will make for an interesting product.

This week I've been working with command line to set up GitHub (an open source, project development service), Stoqs (a geospatial database software), and Django (a web framework software). The only problem during the installation of the software was creating a ssh (secure shell), which is crucial for encryption. The solution was to downgrade VirtualBox (a virtual machine, or an emulator of different operating systems) and Vagrant (software that creates development environments, often used with VirtualBox).

Last week I mentioned that I ran the files I'm working with locally, because reading it over the web wasn't working. Well it turns out that it wasn't working, because there are null bytes in the file when being read into IPython. Here's an example:

-In # = input line #
-Out # is what's being outputted

In 1: data = urllib2.urlopen("http://argo.jcommops.org/FTPRoot/Argo/Status/argo_all.txt").read()
In 2: data = data.split()
In 3: data0 #The 0 signifies to fetch only the first row
Out 3: '\xff\xfeW\x00M\x00O\x00,\x00T\x00E\x00L\x00E\x00C\x00O\x00M\x00,\x00T\x00T\x00Y\x00P\x00E\x00,\x00M\x00Y\x00_\x00I\x00D\x00,\x00S\x00E\x00R\x00I\x00A\x00L\x00_\x00N\x00O\x00,\x00D\x00A\x00T\x00E\x000\x00,\x00N\x00O\x00T\x00I\x00F\x00_\x00D\x00A\x00T\x00E\x00,\x00M\x00O\x00N\x00T\x00H\x000\x00,\x00Y\x00E\x00A\x00R\x000\x00,\x00L\x00A\x00T\x000\x00,\x00L\x00O\x00N\x000\x00,\x00S\x00H\x00I\x00P\x00,\x00C\x00R\x00U\x00I\x00S\x00E\x00,\x00D\x00A\x00T\x00E\x00_\x00,\x00L\x00A\x00T\x00_\x00,\x00L\x00O\x00N\x00_\x00,\x00M\x00O\x00D\x00E\x00L\x00,\x00F\x00U\x00L\x00L\x00_\x00N\x00A\x00M\x00E\x00,\x00E\x00M\x00A\x00I\x00L\x00,\x00P\x00R\x00O\x00G\x00R\x00A\x00M\x00,\x00C\x00O\x00U\x00N\x00T\x00R\x00Y\x00,\x00A\x00G\x00E\x00,\x00A\x00G\x00E\x00_\x00W\x00E\x00I\x00G\x00H\x00T\x00,\x00G\x00T\x00S\x00,\x00P\x00R\x00O\x00F\x00I\x00L\x00E\x00S\x00,\x00D\x00M\x00P\x00R\x00O\x00F\x00I\x00L\x00E\x00S\x00,\x00G\x00T\x00S\x00P\x00R\x00E\x00S\x00S\x00,\x00G\x00R\x00E\x00Y\x00L\x00I\x00S\x00T\x00,\x00C\x00Y\x00C\x00L\x00E\x00_\x00T\x00I\x00M\x00E\x00,\x00D\x00R\x00I\x00F\x00T\x00_\x00P\x00R\x00E\x00S\x00S\x00,\x00P\x00R\x00O\x00F\x00I\x00L\x00E\x00_\x00P\x00R\x00E\x00S\x00S\x00,\x00O\x00X\x00Y\x00G\x00E\x00N\x00,\x00F\x00L\x00U\x00O\x00,\x00T\x00U\x00R\x00B\x00I\x00D\x00I\x00T\x00Y\x00,\x00R\x00A\x00F\x00O\x00S\x00,\x00C\x00D\x00O\x00M\x00,\x00B\x00E\x00A\x00M\x00,\x00B\x00A\x00C\x00K\x00S\x00C\x00A\x00T\x00T\x00,\x00D\x00O\x00W\x00N\x00W\x00E\x00L\x00L\x00,\x00S\x00S\x00T\x00,\x00S\x00S\x00S\x00,\x00P\x00A\x00L\x00,\x00O\x00T\x00H\x00E\x00R\x00,\x00I\x00C\x00E\x00_\x00D\x00E\x00T\x00E\x00C\x00T\x00,\x00A\x00C\x00T\x00I\x00V\x00E\x00,\x00B\x00E\x00A\x00C\x00H\x00E\x00D\x00,\x00R\x00E\x00T\x00I\x00R\x00E\x00D\x00'

Yeah all that \x00 and what not in between the letters and numbers is hex code (which Mike thinks it's because it was written in C and this is how IPython is interpreting it)

Here's the other example:

In 1: data = urllib2.urlopen("http://argo.jcommops.org/FTPRoot/Argo/Status/argo_all.txt").read()
In 2: data = data.split()
In 3: print data0 #The 0 signifies to fetch only the first row
Out 3:
??WMO,TELECOM,TTYPE,MY_ID,SERIAL_NO,DATE0,NOTIF_DATE,MONTH0,YEAR0,LAT0,LON0,SHIP,CRUISE,DATE_,LAT_,LON_,MODEL,FULL_NAME,EMAIL,PROGRAM,COUNTRY,AGE,AGE_WEIGHT,GTS,PROFILES,DMPROFILES,GTSPRESS,GREYLIST,CYCLE_TIME,DRIFT_PRESS,PROFILE_PRESS,OXYGEN,FLUO,TURBIDITY,RAFOS,CDOM,BEAM,BACKSCATT,DOWNWELL,SST,SSS,PAL,OTHER,ICE_DETECT,ACTIVE,BEACHED,RETIRED

There is actually an extra blank line that's being outputted after the last line of words. At first I thought it was nothing I should worry about, but Mike was telling me it was another null byte line of characters. Oh and the symbols in front of "WMO" is anothe example of IPython trying to output special characters not known in Python. So yeah that's my life right now, but there is a light at the end pof the tunnel. I can have the program take out the backslashes by using the .replace function. The only problem is that I can't do it all at one time, only row by row so I have to make a loop that will go row by row and do it. Should be fun!

MARY SILVER!!!!!!! Also, Kris and I nearly finished prepping the eDNA samples for sequencing. We should be able to send them off to Standford monday or tuesday next week. Wiley, Jose and Madison joined me with team Chavez for a CTD cruise. Rienecker gave them a brief intro to the CTD and ROV.

This week I got 5 additional sample splits from Charlie Paull's lab and ran them with the Raman pressure booster system. I also have started modifying the Matlab code that I will need to process my data. Thankfully,

our lab has started code that's been used to take difference spectra and integrate peaks from other types of samples. It will probably take a long time to figure out the best method to process my spectra, since so many of the peaks

that I need to look at are overlapping. I look forward to getting into some hairy Matlab. Next week I will be the only one here in my lab, so I need to make sure that I am set up to go through all my data before the day's out.

Week 5:

I cannot believe we are actually half way through; it's a little nerve wracking. Well this week was a pretty successful week. I was able to write a code that plots my manually selected points with ellipses that another code generated. Now I am working on essentially throwing out data that will not be useful, such as when jellies are too close to the camera and only the tentacles can be seen, so the orientation cannot be accurately identified. This I have found to be a bit more tricky, but hopefully I will get it soon. Overall this week I have really felt that my Matlab skills are improving and I can begin to think more in the right direction when trying to develop some code!

Week 5:

 It's pretty amazing we are half way through but I definitely am feeling pretty good about the amount of work I have completed and the data I have collected thus far. I am so close to finishing reviewing the Pulse images and I am waiting for the most recent one to be uploaded to VARS. It was a pretty exciting week with the activities and the speaker. I am looking forward to Lobos tomorrow but I definitely know I will have to work on Saturday and Sunday in order to meet the goals set out by my mentors. I had a lot of good conversations with MBARI employees and visiting researchers this week. Since this project started I feel like I have a much better sense of the direction I would like to go with my future research and the kind of PhD programs I need to apply for.

Can't believe we are already half way through, but I love my project and the progress has been decent . In the last two weeks I have been working on building my classifier and get my classification system in place and running . I am employing a classification algorithm called as SVM(Support Vector Machines) for my classification purpose. Currently my dataset includes a bunch of fin whales and blue whales(Total 5214 samples). Running the classifier on these samples has given me a classification accuracy of 96% and an average precision of 99% which are really amazing results( I expected the classifier to fail at first ) . I will be bringing in more data( or classes) , specially high frequency data (sperm whales,dolphins etc). I might be trying an image based classification on the spectrogram images(using deep learning) of the signals parallel to my acoustic classification method.

It's been a pretty productive but short week here at MBARI. I had to work from home today because
of a meeting I had to attend on campus and tomorrow is a MBARI off-day so no work; I'm still working
on my poster for the MBARI open house and plan on showing it off throughout that day. The less text the
better I'm starting to find out! Next week I will be spending a lot of time reviewing literature
and learning as much as I can about whale benthic feeding habits. I still have yet to match an
actual whale species to my data.

Week Five:

Well I must say it's great to have Brett and Ben back from their vacations, I feel as though I have had a productive week with significantly more direction than the last two or so weeks.

As far as my project this week, I started off by calculating the scale resolution needed given a 1/32" or 1mm shift in CG and wrote this up in a word document.

I Looked at the hydrostatic moment equations and determined what we could derive from the pitch moment and what we would need to measure.

I Worked on a draft procedure for locating Xg, Xb and Zg of the long range, and will work to generalize the procedure for other AUV's considering Dorado for example does not have a shifting mass or a buoyancy engine.

In the next week, I hope to have the scales purchased, work on the design and material selection of the vehicle chocks, and some adjustable format to account for the LRAUV with and without the ESP, as well as possibly making accommodations for use with Dorado.

I'd also like to clean up my procedure and make it applicable to the non-long range vehicles and create a spreadsheet and/or code which has these equations built into a sequence following this procedure.

Week 5:

I now have access to live radiolarians! As soon as they get uncomfortable, they retract into their capsules, though, so it's still difficult to work with any in their natural, "goopy" state. I've been staining them with an actin stain to see if/how far the animals extend into the colonial sphere, but so far there hasn't been much to see. The other stain that I wanted to do was a silica tracker, but I failed to pack a necessary solvent on Monday, so that experiment will be limited to the days following our return and will depend on how long the rads live. It's pretty disappointing to miss out on that, but there's nothing I can do now, and at least it's a learning experience :/ Plus, I can still work on the other aspects/subprojects of my project, so I won't be wasting any time! On the coding front, I'm trying to reconstruct a 3D model of the lattice from 2 SEM images taken at different angle. There are established methods for this, but they require linear algebra knowledge (that I'm trying to learn) and I've temporarily switched over from R to MatLab since MatLab is, y'know, for matrices. There's also the genomics subproject to return to after the cruise. Since we're halfway done, I'm starting to get concerned that my project isn't focused enough and I'm trying to look into too many aspects of the same thing at the same time, but Steve said that it's alright since they're thematically connected, so I suppose it's fine? I'm pretty happy with it because it means I'm getting to learn so many different things! I still need to think more about synthesizing my subprojects into a coherent paper/presentation though...

That said, it's kind of hard to worry about that when the cruise is so awesome It's nonstop activity every day, and I've learned how to use the video annotation software, operate the suction sampler on the ROV, transfer live samples, set up the trawl net, and so much more! There are blue water dives and trawls every day, and the ROV is out 12 hours a day, so there's a lot to see and so many specimens to observe all the time. Doesn't hurt that most everything collected is gorgeous and bioluminescent, either. They're even targeting the prettiest ones because there's two guys on board filming for a documentary. The time spent in the control room also means I'm learning more species names! Everything is very exciting, and I even managed not to get seasick

Week five:

I am finishing to get all my data. I would have liked to have it ready this week, but you never know what will come in the next dive. Now, I have to watch videos from one more dive and so I can complete my data. However, as I have part of my data, I began to organize and do some analysis using ARGIS. It is very interesting and amazing how much information I can get from my study.

In addition to learn new things in ARGIS, I spent a few hours in MATLAB too. Although I had some knowledge of these two programs, it is always good to remember and learn more. I am very grateful to have this opportunity here.

Closing week, I will have an excellent dive in the Point Lobos!!!.

Yesterday I Spent the entire day out at the with Gene Massion of Kne Johnson's lab. Gene is working on a new design for a coastal profiling float.
We took the paragon out about 4 miles and then proceeded to hoist the float overboard. We also deployed a line with a weight attached that the float would traverse down to 60M before returning to the surface.
The line simply served as a fail safe should something go wrong. The float made 10 trips down to 60m and back, collecting ctd, oxygen, and fluorescence data.

Apart from that this I spent this week working on the algorithm that I will use to correct the rovers heading, which while progressing slowly, is progressing.

I've been incredibly busy setting up the second iteration of my abalone development experiment. I finally overcame the battle I've had with my abalone being impinged on the outlet mesh and am looking forward to running this next experiment with a completely different design. I've also been busy with REU activities so I'm trying to find a good balance between both of these programs. Diving at Point Lobos was absolutely AMAZING and I intend on diving there again sometime during the school year. I also went to the sleepover at the aquarium where I saw the one and only, George Matsumoto, give a great impromtu presenation in front of the kelp forest exhibit. Memories of sleeping under the deep ocean will be cherished for my entire life; it was absolutely incredible. Thank you, George, for making this happen! Looking forward to week 7 is so exciting and I can hardly wait!

Sorry for the late post!

Last week before leaving for the long fourth of July weekend, Brian sent me an algorithm that I was suppose to add new parameters to this week.
I was expecting it to be as easy as copy and pasting the code, inserting the new parameters, and running it in Matlab.
However, I really underestimated how difficult it would be...the language was a lot different than I expected. I am almost certain it is in a
different language and it has been very difficult for me to get a hold of Brian while is away on vacation.
In the mean time, I have been trying to learn more about the code and potential new language. I have also decided to take more length measurements, in
which I isolate length from angle so we can just focus on the relationship (or lack of relationship) between length and error.
I finished the measurements on Friday and ran the stats on it. Turns out that there is a significant affect of angle on error (which we expected),
a significance of length on error, and no significance of angle on length (...which means I successfully removed angle as a factor so we could look at the
relationship between length and error!). In graphing the data, it also appeared that as the lengths got larger so did the error and the precision decreased.

Now the next few steps are to: create equations that define the relationships between error and length, as well as, length and error. Then incorporate them into the
algorithm.