G'day.

We're working on our thesis proposal right now. Our thesis would be to create a door lock, which you can call, and thru voice recognition can lock/unlock the door.

Is this feasible?

We already have a vague idea on how to go about this thesis. Basically, a telephone circuit will receive calls. The caller's voice would then be converted into digital format. A microcontroller will be programmed to do the voice recognition and will drive a motor to lock/unlock the mechanical door lock.

Will our idea work?

Any critique, suggestions, and inputs will be greatly appreciated.

 

You make a door lock that is equipped with Blue Tooth (wireless), and this will be used to activate a solenoid for unlocking the door.

You then have a PC with a telephone monitor and voice recognition software. Whenever the phone rings, the PC will answer and perform voice recognition on some phrase or password. If the result is positive, the PC will send an encoded signal via Blue Tooth to the door lock for it to unlatch.

 

I would avoid any implementation of voice recognition on a microcontroller. It is somewhat complex and requires some processing power, which microcontroller lacks. You'd need a fast processor, and even then, the algorithm might not be simple to develop.

The easiest way is probably to use a voice recognition IC, there are several from Philips, Sensory or OKI. These can be interfaced to a microcontroller and then to your lock mechanism.

 

A thesis for what? A PhD, or a bachelors degree. If it involves a requirement for original research then I think the proposal falls a bit short. I'm not up on current research trends, but I think you need to try and break some new ground. What does your advisor think?

 

A thesis for what? A PhD, or a bachelors degree. If it involves a requirement for original research then I think the proposal falls a bit short. I'm not up on current research trends, but I think you need to try and break some new ground. What does your advisor think?

this is just a bachelors degree proposal. our advisor gave the go signal on our topic. a similar thesis has not yet been made in our university, so our topic was approved.

we can still add more features to our thesis before the term ends.

 

nomurphy and n9352527: thank you for the inputs. this will surely help us.

 

wouldn't the bluetooth idea be vulnerable to hacking?

 

G'day.

We're working on our thesis proposal right now. Our thesis would be to create a door lock, which you can call, and thru voice recognition can lock/unlock the door.

Is this feasible?

Yes, but see the Dr. Who episode for how reliable it is. The simplest I've heard of was a very specific word stored in an eprom and compared with some 'slop' allowed. That was done back in the 80s but I can't recall where I read it.

We already have a vague idea on how to go about this thesis. Basically, a telephone circuit will receive calls. The caller's voice would then be converted into digital format. A microcontroller will be programmed to do the voice recognition and will drive a motor to lock/unlock the mechanical door lock.

Will our idea work?

Yes, it's been done you need a pretty stiff microcontroller. You might want to look at a TI microcontroller with an integrated DSP. I've only read of their capabilities not used them and know that Raytheon, a defense contractor bought them just for those capabilities. There are others that can do the job and you should review all of them. I'd suggest writing (not emailing) the right department and tell them what you want to do and ask for all the information on their development kit they can send and if there is a student discount. They might send one for free.

Any critique, suggestions, and inputs will be greatly appreciated.