Genpact Nasscom Social Innovation Honours 2011

We had sent abstract of Blind Helper Project for the GNSIH2011. It got selected, and we were amongst the 3 team in student concept category selected all over India for the finals. Finals were to be held at Gurgaon, Pune and Bangalore on 3rd Jan. Jury included even Mr. Narayan Murthy! But, it seems that we won't be able to participate as it is coiciding with our datesheet of End Semisters. :-(

Alrite, so its now a pretty long time since my last post..
We have now been working on a new concept of interactive content development assistance in the project. This would lead to content being developed by the care taker or the content developer in being highly interactive in nature. We have also been working on Speech Recognition through limited grammer to increase accuracy. The on-board sound card facility needs lots of improvement. At present just some sort of creaky audio can be heard. There seems to be problem in synchronization in the sending and receiving of packets. The project is paused at present and we will set back to work on it again in march.

1second long Audio Data plays from from ckt! :-) And guess wat! I walked 10 km today in rain! :-)

Ah! My day started with my mom telling me to get some work done in market.. So made plan to meet one of my old friend as well! Who! I met him today after around 5 years! But as I left from his house, it started raining! With such a beautiful whether, how could I have not enjoyed it! Its fun to get wet in rain! :-) Enjoyed that after sooo many years! I walked back home around 8-10 km away in rain! Felt awesome! :-) Met two of my school friends on the way as well! And samosa and lassi on the way were too delicious! Yummy! :-)

But, throughout, mind was still in the circuit and the concept and thinkng why I was getting error and why I was not able to hear a piece of music stored in the MCU. So set out to work soon after reaching home after a 1:30-2 hour walk! :-)
And yipieee, after few adjustments, it worked! Though it was just a 1 second clip of a telephone bell ringing and extended by putting it in loop, yet the feeling of its first running was gr8! Adjusted some other parameters and worked out to interrelate them and find the optimum...

So, things do seem to be moving on the right track! :-) Now, will be working on streaming a complete song from PC to the circuit and then decode and play it Live.

Back to Home Sweet Home!

well, well.. so its been a looooong time since I posted here..
first jst a small recap of what all has been going on since past so many days... I'm back to my home sweet home after spending about a month at IIT-K. So, now, I'll be continuing the project now from my home town.. Anurag has also gone to his hometown for some days... Ah.. its been so long since he last spent time with his parents! Anyway, have a great time there dude!

Meanwhile, I set back to the project just the next day when I returned from Kanpur. Components from US have also come.. (But, damn... I had to pay Rs 1200 of Import Tax!)..
Now, whats dat tax for! I don't even earn yet!!!

But, it was great to have the components, and starting work over them.. And using these componets require lots of care as well! So, expensive, and so less in number as well.. No scope to fry any either! ;-p

Anyhow, I set out to work, and after 3-4 days of continous efforts, Ive Finally been able to hear a tuuuuuuuuuuuuuuuuuuu Tone! lolzzzz.... Its a simple Sine tone though of around 5500Hz frequency, yet its seems so wonderful to hear it! I'd been fiddling with the circuit and hardware and software since past 3-4 days... have been reading the datasheets over n over again... going through the forums, and sites, and tutorials, and finally!!! Yippieeee!!!!!

But, the coming days will see lots more challenges... But looking forward eagerly! Now next job in queue with this hardware is to play a song being streamed from PC.... Hopefully this should be done in another few days... :-) But this requires overcoming lots of complexities.. But, I will! :-)

I have been searching alot on the net about the modules that can be used for effecient data transfer. Cost and budget are putting lots of limitations.. we need to transfer mp3 or wave files wirelessly. Even if I speak of an ok - ok mp3 quality, then it means around 32kbps stream. i.e. 32 thousand bits need to be transferred wirelessly from main hub to the hand held device each second... Now, that is controllable for 1, 2 or at most 3 devices working simultaneously on XBee... But using more than that will cause problem, and the audio being heard would be broken.. Now, I've come across different counter strategies- like to keep larger data streaming when the device is not is use, and then to store that data inside the device itself in a memory card. So, when the request for that data is actually made, then it will automaticly be available in the device. But this adds on to the cost.. Cost of a memory card, and also of another MCU, because Memory card works on SPI, and we are already using MCU's SPI in other things...

After day full of calculations and net searching, n data sheet readings, and cost comparisions, I'm glad that the solution seems quite feasable... Luckily a module comes at cheaper price than XBee, and allows pto2Mbps data transmission rate! And also provides 160 seperate frequency channels! Yippie! Coz that means, that we can now cater to around more than 50 requests with LIVE streaming @ 32kbps MP3 files! It means sending LIVE individual Music Files from 1 server to 50 different people who listen to it wirelessly! Or, alternatively sending Audio Speech file @ around 16kbps to more than 100 different people simultaneously! Seems amazing! Seems, we're going on the right direction! :-) Feeling lil relieved after 2 days of calculations and analyses in lab, and even in dreams! phew! But at the end... it does seems fruitfull! :-)

So, initially we'll work on the XBee modules that we have in Lab, and work out for the functionality of MP3 related ICs, and other ICs.. Once we go for fabrication, then we'll be modifying the code a lil bit and will use the newly found RF Module... Theoritically things are appearing fine.. The calculations, and datasheets show things to be possible... Now just waiting for the components to arrive to test it out practically...

Blind Helper Till now...

Blind Helper project which aims at making a device that can help the Visually Handicapped people in operating computer was initiated by Mohit Khanna and Anurag Mishra. Till now, the system has been developed with self support and less resources. But, now, the project has got a new spark with the back support of Dr Priya Ranjan from IIT-K. Now, working at Dr Ranjan's Lab, we are very much hopeful to make this project much more effecient and usefull for the Blind people.

Initially we are interested in making the system capable of handling multiple users. So, that means solving some major problems in the way. One of which is like a milestone, and is to allow multiple users hear their own requested audio from the same PC. For single user it was damn easy by just sending the audio out the audio port, but now audio port can't be used, else all the audio streams of different users would be heard all together, superimposed. So, some other technique has been though about which includes sending the audio stream in byte form to respective devices with address. This includes lots of 'think-tank!' as well as this is quite a complex task and as said before, this is one of the biggest milestones of the project... (but not the only one though! ;)

Components have been ordered and we are waiting eagerly for them to arrrive, so that we begin implementing the test codes that we have written..

Meanwhile, the software is also being developed and currently work is being done to integrate the front end (java) with the backend (my sql). The basic low level UART communication from the software has also been implemented and tested. Some basic functionality of checking user password, etc for authentication purpose has also been implemented. These days main 'think tank' about the software end is to device a method of sending audio data in byte form to corresoponding hand held device, and to kind of synchronise the whole system.