Transcript for:
Drone Building Crash Course and Campus Tour at Scaler School of Technology

now I see why you guys went offline this is beautiful today I am at the scaler School of Technologies offline campus I'm about to get a crash course in how to build a drone by using off-the-shelf components from a team of brilliant 20-year-old students these students will walk us through every step of the process from selecting Motors and propellers to integrating liar sensors and AI for autonomous flight capabilities not just this at the end of this we we will test it by actually taking a flight so without wasting any time let's get right into it yeah man we would love to show you the campus sure uh so yeah uh this is the campus like you know the concept of ss right by industry for industry so when we were thinking of the campus we thought of how uh I mean from a curriculum design point of view what Microsoft wants what what Amazon wants let me start with showing you the classroom oh this is huge this is much larger than I thought it would [Music] be now this entire space that you see we have equal amount of space on that side which is the innov just the Innovation lab that looks really exciting this part is a classroom but that looks interesting hi how you doing to you likewise this looks damn good this actually looks good this feels futuristic too this is your robotics drone lab arvr a IML and that's your iot lab this is drones hi guys sir I'm going to be seeing a drone today you're going to be teaching me how to hi sir this is Dave make and ride a drone I know some of it but hopefully you'll give me a full lesson on this hi hi how long you've been working on this for the last two months nice I'm excited to see how it works so today we're going to find out like whether drone the Drone actually flies and what it's made out of so we're going to do three things one is you're going to show me that the Drone flies then there are two ways I learn right one is either I build something or I break it down breaking it down is much faster so I would love to have that broken down because otherwise building it I think will take a few hours so you'll break it down layer by layer until we get to the motors and the esc's and whatnot you're going to explain it to me and then hopefully I'm able to understand it okay I'm a drone Noob though I know nothing about drones except for the very Basics so you have to treat me like a lay man okay def shall we go yeah come can I fly it yeah do you have any experience so I have zero experience flying a drone but let me try I'm good with an Xbox controller so this is the throttle so this will this will dictate the amount of thrust it throws down and it will go up this is left and right this is called roll technically it's called roll roll P correct you know Lord man you're not a noob so this is this is basically your this is here this is this is for the pitch and this is for the rooll okay got it yeah that's it yeah that's it okay let's start when we connect the battery and then there is an arm switch on which will arm the cter so there are two uh two arm switches one is on the copter and one is on the remote you guys should move away yeah they should move away because someone will die and then I'm not responsible very slightly give thrt yeah see it goes up then you can control it two in your take it right yeah perfect perfect take it right oh but this will be an autonomous drone yeah this will be an autonomous and it's driving how through La Raspberry Pi and there will be a flight control system on board okay which will control it and uh give the FX hog the input which we give it through the radio it will basically tell the Drone to like go no no but if it's autonomous why are you giving it instructions we are not giving it the instructions the Raspberry Pi on board will give it the instructions so so so play play play it for me so it is the world through lar is a 3D representation of the world can you explain what lar is to the people watching lar is basically an instrument through which we can uh sense the range of the object uh in front of the Drone what's the full form for the audience uh light detection and ranging okay nice so it's seeing the world it's giving you a 3D representation of the world then you're bringing that world into uh the Raspberry Pi yeah right and then you're deciding where to navigate purely on the lar on the camera exactly okay but what it can the camera see behind itself no for that the Drone will just rotate 180° okay okay you want to break that down for me yeah yeah yeah yeah sure sure sure hey you want to quickly take a detour and go see the Vision Pro folks hello you have to be honest how many times have you used the device in the last one month probably daily for 3 to 4 hours really yes Apple loves you so basically when we renovate our home so we are not able to visualize which panel will look good or how with my furniture how this color will go correct or not so for that we have built a visualization where you can place panels on your walls with your furniture applied means like uh if there is a TV and table so panel will be placed behind that so you can uh take a feel of all home with the new panel uh with the extra thing you are adding and you can decide which panel is look can good or wall panel like uh nowadays uh know people are applying wall panels like wooden panels or some acrylic or something something like that so Urban company has started a new segment uh in the in the business where they're replying wall panels so you're doing some kind of wall detection yeah W detection and whole construction of mesh around the surfaces nice uh so if you touch the any panel it will be placed so if I touch this panel yeah you have to tap it from the index finger yeah so it's placed oh nice this is actually cool this is way better in the reception area out there no this is this is this is still really nice this is still really nice so that part is really great because the more complex it gets yeah segment all of this out you can use Sam you know what Runway ml does something very similar you can do it in real time so can I also tap that no we have uh we have nots dis the ceilings are disabled nice I'm already here to destroy the lab and how are you doing this uh this we have used a combination of plane detection and mesh reconstruction nice so to achieve the and how are you texturing the mesh why don't you use stable diffusion and texture the mesh so that I can I can just say make it look like flowers and then there's a flower wallpaper so we trying to work on that because uh there's a mesh resource we can place the mes So currently we are trying to uh Implement a video on it so it will look like the Matrix oh nice so on the whole layer it will place Matrix so zeros and W like an LED wall yes like an LED wall so in of wall will be the whole en so you'll be inside the Matrix so trying to build that right now nice this is pretty impressive okay back to the Drone folks let's go see what they're up to start with this these are the motors yeah correct these are all the motors correct so then you got these so how much does this cost just the fan the fan do the fan 50 rupe that's it yeah that's how much does these motors cost uh 600 rupes nice and they're driven I by an ESC ESC yes you're right so the esc's are here can you explain to the audience what an ESC is ESC stands for an electronic speed controller what it does is basically takes signal from the flight controller it gives the motor the current According to which um the motor should spin at what rpn should it spin that is decided by the pixhawk what is the pixhawk the flight controller yeah the Pix Haw is the flight controller which we're using so the pixhawk is sending data to the esc's which is then controlling your motors which is then making your fans rotate at for a certain speed the job of the ESC is to control the RPM of the motor in simple words nice and what does this do what is radio telemetry radio telemetry through with through this we can we can make the pixo communicate to our laptops using a Telemetry um communication device so what you can tell the laptop uh what the pixhawk should do yeah I mean the laptop can tell can tell the what the Pix can do what is this this is the uh receiver for the remote for the radio radio controller uh receiver radio receiver can you show me the liar in the camera module yeah yeah yeah sure this is the liar this is the liar module so basically how how it works is you see there are two two black patches on the top so basically one sends light so it's only forward liar usually a liar model is 360 no this one is that's what we said this is a one point lar okaying so it's only looking in front yeah it's only looking in front creating a map of the world in front of it yeah nice and the Drone can rotate 360° so we can create a 360 map as well nice so how it works is it sends this light beam and this this one receives the light beam they are they are pointed like this so uh it calculates the time which it uh took to send and receive and it multiplies that by the speed of light to calculate the distance got it and what does that create a 3D map of the uh mapping is done on the software side it'll just tell the um range of the object in front of it distance of the object what is this this is the brain of the Drone like the where the ml models and all the object detection model fire model you're running everything locally nothing on cloud no nothing on cloud like there would be problem if it would run on cloud like it would it would need internet to access the cloud right so if we do that the Drone might fail at any point we don't know so basically you're saying with this combination of the frame uh your pixhawk your Telemetry devices uh your esc's here your Motors and your fans anyone can build a drone at home yes yeah uh but with a little bit of knowledge about about the calibration procedures nice which go which is which is very easy to find do you want to run me through the ml model that you guys have on the system yeah like the fire detection model it will take the visual inputs from the camera and the light up like from the L it will calculate distance but it will take the visual inputs from the camera and it will send to the RP and RP that it will take the one frame out of that video and it will process the image it will detect the fire and if there is a fire it will do the PID control that is like minimizing the error for it yeah whereas the because the camera is sitting on top of the Drone and there's Jitter is that why there there is that to to combat that thing we have something called as a jimble which will stabilize the camera feed and what it'll do is like right now the fire is over to this part of the frame so what it'll try to do is to bring it over to the center so for that there is the this much error so you see like this much error is there so to correct that there is something called as pit controller so the pit controller will determine how exactly will it will it move on to the will it take this fire to the center and once it's in the center you just have to take the Drone towards the fire and then uh that's how we we we'll we'll be on top of the fire at so you're right on top of the fire then yeah then there is like then then there will be a custommade dropping mechanism which we're developing right now and it'll drop the fire drop the ball onto the fire and that's it got it so basically like the Drone has to go forward so what we we give the input of forward let's say we give the so it translates that into exact rotation so you don't have to map out what has to rotate at what point exactly like if we if if we tell it to manually ding in RPMs you're just saying forward and that's translating into how should will basically sync all the motors with itself the anticlockwise with the anticlockwise with the clockwise with the clockwise so basically like go forward so so how figure yeah it it'll tell these two Motors just spin faster and and go forward got it yeah I think everyone has a good view of how it works so hopefully audiences at home you learned like me how a drone works and how it's able to put out fires with a combination of AI and uh 85,000 rupees roughly for the Drone yes nice very very very impressive lovely this is super impressive thank you so much for having the patience to answer all my stupid questions every question is important and uh very very impressive uh are you guys planning to start a company out of this are you planning to commercialize this soon nice we have plans but not commercialize this like do more of like nonprofit kind of a thing still very very impressive congrats how old are you uh I'm 20 you're 20 years old I'm 18 I'm 18 oh my God superb stuff my audience is used to the fact that we'll cover something and then they'll double click on certain things as they like some people will be like no I will build my own flight controller some people will be like like frame 3D print some people will be like no you know I want to use some other type of Motors whatever like you can feel this is the Basel line and you can feel free to modify it as you want but what I'm also really impressed with is this nice Iron Man looking laboratory that you guys are sitting in and what would you like them to build next yeah what I would really like what we are working on I'll tell you so again we've just started a few weeks ago but uh we are working on a quadruplet bot right so we got this bot from China called the yaha boom Dogzilla and then many of these available there's the unitary go1 and stuff right so the cool thing about the bot is it's built it so there's a open-source platform called Ross Robot Operating System where I can control the bot now through tele operation yes but I can also you know write my own software for it to be my friend for it to hang out with me and it has all the devices built in so I need to buy like a separate lar module I didn't need to buy a separate camera module every everything came in right there's this unique opportunity Boston Dynamics did to put a skin on top of the bot yeah for I don't know if you saw this they put yeah they put like a dog skin on top of so the question I often wonder is is there a boatlike business possible but for Robotics and drones and for anything right like where China produces or whatever in the early days China will produce it we will build a business on top of it reskin it whatever and then over over time you can build bring the manufacturing back to India cuz the other way around is super risky right where you build this and then you commercialize this and then what if it what if you take it to Market and nobody buys it it's scary and you've lost a lot of money you've lost your money investors money whatever it is the other way around where you're just buying a product and reselling it to see is there a market first and then once you've seen if there's a market first and if there is a market then you bring everything in-house right so yeah like people keep saying India needs a lot more manufacturing but actually India also needs appetite for this because I don't know who would buy consumer robotics right now in the country yeah right so we are also experimenting it take like 3 4 5 years for us to and might not even work like 99% it might not even work but as an engineer that intellectual curiosity never leaves you see I mean like people if they get introduced to robotics then they they'll have that Curiosity of learning so that introduction part is most important I guess yeah I mean one thing that people don't get which you know I'm glad that this one lab kind of proves it right that software is not just one thing it's not just react development right it's like if you're good at software you can do XR then you could say okay I understand XR like he's working on Unity right there yes right so it's not that different Unity is I mean it is different but it's not that different from what you're doing like pitch raw uh pitch Your Role I first learned it in flight simulator Microsoft flight simulator then I've used it in Unreal Engine and then I found the the same Concepts here because you're moving in 3D space right so it it's obvious you eventually come through pitch role you so all the concepts translate but the way regular engineering colleges is like very um it's very oh you know it's you have to learn only this thing or you have to learn only that thing my question is if you're a good game developer right why then you can also translate some of those skills to robotics if you're a good Rob because a lar like especially the 3D lar not this one but the 3 lar is creating a map of the world and you're navigating based on the map in Unreal Engine there used to be this thing it used to be called AI but it's not really AI it used to be called something called Blackboard where you would make decisions you would say hey here's a decision tree right all animations unreal are State machines so these Concepts translate across platforms across things you can do and one of the cool things I really love about this lab is there's no differentiation like it's not like this is the center for react Engineers India right it's like okay if you feel like going there and doing some XR if you feel like like you're also using AI in this yeah we are right so there's no real distinction so what we feel when we do stuff is um it's experimentation and maybe we might not build a successful robot or build a successful robotics company most likely that's going to be the case but we learn something from it that we can apply somewhere else and it's also good content right it's all curiosity and the hunger to learn things and Tinker things and it's fun like what will you do in life otherwise exactly uh I mean I hate saying this because I I I'll get belted for this but you should keep it anyway but I think Engineers have like good Engineers generally have the most fun in life cuz that that I think scientists have even more fun I I wouldn't know but just that feeling where something just works and you've sat all night I'm sure you all know the feeling you you sat all night and like so I have one last question around this how did you guys get in did you did you just do does any 18-year-old just walk into this lab like why are there so few of you how how did you get here basically like there's first of all there's an exam called as net and we have to what's it called net how spell that for me NSE okay yeah uh you have to first get clear that exam and then there will be two sets of interviews one will be math interview and one will be dreams and personal interview in which they'll see your dreams and aspirations and after these two interviews and this test you're done you're you're inside this inside this institution that's and the exam itself is kind of a learning thing like there is a part called learnability where they put the concept itself in real time and they ask you the question what's what will be the answer of it so adaptive learning ability not not mugging yeah so they'll give they'll teach you they'll give you a paragraph which will teach you the concept so in my case it was like binary uh conversion so I didn't know about that concept but like I learned it on the spot and I was able to like clear the questions on the spot so you have to learn it on the spot and clear the questions nice super impressive and it's not only like learnability like if you have the the logical reasoning and Mathematics thing you'll also get selected into this like college like the math thing in computer science also in AIML whatever thing you do majorly in computer science it's only match what dat yeah that's the only risk with this place right the minute your Entrance Test becomes too easy yeah you get like someone bad who's not contributive in a group you know you have that right in a group project even if one guy doesn't contribute it's done yeah so with offline institutes like the entrance test is very important I'm glad there's there's an interesting there's a good Entrance Test how how how do people apply for the nset nset is like completely online you just have you'll have to fill a form you'll have to go on to scal School of Technology website you have to register it there and you can give the exam so that's been a look at the fascinating drone lab and the impressive offline campus at Scala School of Technology we got to see some brilliant students helping me understand how autonomous drones work by breaking things down what really stood out was how they blend different Tech skills it's not just about coding or building things it's about combining everything to solve real world problems if you're into Tech and want to learn by doing Scala School of Technology might be worth exploring so click the link in the description that's it for this video and I'll catch you guys in the next one bye