Automatic and robotic quadricopter

[w_t]

Hello everyone,

I am new to this forum, but I'm sorry for my English with the translator.

I apologize if I'm wrong section in the forum, I do not know about that topic, section or group post, please moderator if he can move in the due section.

I have a very ambitious project that I would like to see if I can finish it,
I would like to build a quadricopter able to fly without operator, in a given perimeter and that is able to avoid obstacles.

eg: in my house I have a big garden, I want the drone flights to 10m in height along the perimeter of the garden, but should scanzare obstacles like trees or if the route intersects with a bird like a dove or a raven, must dodge so as not to be damaged.

I mentioned this in other forums, the answers were all negative, "it is not possible, it is very difficult, it takes too many instant calculations, there must be a computer on board too powerful, etc ..."

I'm here because I have talked about the forum very well, I know that there are many technical professionals.

I kindly ask if you can help me out to see how far I can go, then it is really impossible, I can say I have tried until the end.

As a first, second and Step clear what would you do?
What are the problems insurmountable?

I sincerely thank you for your help.

 

[GJH105775]

Welcome to quadcopter forums

It is possible, ans has been done, but assuming you have a good bit of experience with a computer language like C, it will be somewhat hard.

What is your native language? I would like to find some tutorials that will help, but they would need to be in a language you speak.

It is fairly common for people to add obstacle avoidance, this uses ultrasonic sensors.

Another more common needed function that will be needed is a way to know altitude, usually this is a combination between an ultrasonic sensor and a barometer for a little higher.

If you are going to do this I would suggest you get a simple flight controller like the KK2.1.5 and use it for stabilization, and have something like an Arduino or Beaglebone to control the direction of the quad.



The goal in the video below is 3 dimensional mapping of a room, it has been done for sure 4 times, probably more. You may be able to enable captions in your language.

There is also a way that it can be done with a regular camera, this is visual navigation, the quadcopter would use just the information from a regular camera, this would be harder.

 

[w_t]

First I want to thank you for the answers precise, sincere, no one has ever said so, that just was not possible.

For the C language does not believe what the problems are, are a programmer for 15 years management and development sites, working in C #, PHP, Python, JavaScript, ActionScript, SQL, I think learning C or C ++ is the last problem.

Raspberry know, I already did some testing for another project, I used modules such as accelerometer and GPS,I think we need a raspberry for my project, since the drone has to do video capture, I do not think it's possible with Arduino.

My language is ItalianI

have seen all of the first video and a part of the second.

I have some questions:

That computing power it takes to manage a scenario like what I saw in the first video?
Just an Arduino?
The speed of the drone to intercept the obstacle and to avoid this is what I saw in the video?
That prototype could operate at a speed of 60Km/h ?
This system is the most secure and accurate, or there may be others with other types of sensors?

 

[GJH105775]

First I want to thank you for the answers precise, sincere, no one has ever said so, that just was not possible.

For the C language does not believe what the problems are, are a programmer for 15 years management and development sites, working in C #, PHP, Python, JavaScript, ActionScript, SQL, I think learning C or C ++ is the last problem.

Raspberry know, I already did some testing for another project, I used modules such as accelerometer and GPS,I think we need a raspberry for my project, since the drone has to do video capture, I do not think it's possible with Arduino.

My language is ItalianI

have seen all of the first video and a part of the second.

I have some questions:

That computing power it takes to manage a scenario like what I saw in the first video?
Just an Arduino?
The speed of the drone to intercept the obstacle and to avoid this is what I saw in the video?
That prototype could operate at a speed of 60Km/h ?
This system is the most secure and accurate, or there may be others with other types of sensors?

The first video with the 3D model of the room is defiantly more powerful than an Arduino, at least a Beagle bone, could easily be an Arduino with a board computer like a Raspberry Pi. If you are just moving forward you may only need two sensors. You could make one operate about as fast as your motors will be able to react. The system in the first video is not trying for obstacle avoidance, but it is mapping it's surroundings, yours would probably just need to tell if something is in a certain direction and how far.


Glad to see that you have a good bit of experience with programing, the reason you were probably not given a good answer on other sites is that they assumed you a were some kid just out of high school looking to kill time with a cool project, and most people would not know where to start on a project like this.

You may not need to learn much new syntax for another language as C# and Python could get the job done depending on what control board you world decide on.

 

[GJH105775]

What are you planning to do with this. The reason I ask is so I know what you will need.

 

[w_t]

And a very futuristic design of safety for a particular person who will make a personal use on his property.

The operation is as follows:

There is a villa, along the perimeter of the villa there are intrusion sensors, an intruder enters private property, the sensor is triggered, it sends the signal to the drone, the drone aside and go to that point, he started shooting video, you Place near the intrusion sensor and does a scan of all objects in motion, the intruder began to escape, is intercepted by the drone, then the reed alarm, is positioned to 3 meters in height and performs a series of rotations around the intruder with a radius of 2 meters, obtaining a 360 degrees, then alsa to 8 meters and takes everything to the end of the run.

Would you give me a hand real development or just out of curiosity?

 

[GJH105775]

This is a little beyond my current skill and free time, I will help you with what I can. One thing to consider though is that this would a be a serious step in a security system, and if someone breaks in to a place they could easily down a noisy quadcopter with the swat of a stick/their hand or throwing a pebble.

If this is your end goal I would suggest that all of your processing is done at a computer that would act as a ground station to tell the quad what to do. This would allow you to have a smaller quadcopter.

Another consideration is the short battery time these have, it would be no problem to make some sort of docking station that it may take off from, but it's ability to pursue may be hindered by the short battery life.

These would be your long term goals, the project would probably be at least a year for a one to two man team.

What type of budget do you have?

 

[w_t]

I realize you step in safety, noise is not a problem, we are in private ownership.

For the board of the computer processor to the ground, then we are faced with another problem, the distance, as the drone communicates with the PC on the ground at a distance of 20,000 meters that is two hectares?

What about the drone passes to the PC of the earth?

The PC sends the drone of what land?

There we return with the communication speed or generates delays in the movements?

Battery life with a drone medium / small as it may be, a few hours? Then recharge to return to the docking station

These should be the objectives in the short term a few months of work

As I told you is a particular person, if the project works the money is there.

 

[GJH105775]

Well yes you could get 20km with a LRS and external antenna, you would have to minimize sensors and just use video but your quadcopter will never be able to get that far, and especially back, in such a case that this is what you need you may want to make a quad that can convert to a plane once in the air, this is not uncommon, but for a quad to be able to fly 40km is almost unheard of on a single battery.

 

[w_t]

Sorry but I have the wrong number, are 2 hectares of land, and 20,000 m / 2 then in a straight line on a square area from side to side are roughly 7,000 meters, we consider this as the maximum limit, I think that on average there are about 3,500 m flying right.

What are the implications in terms of space and costs a station LRS + Antenna?

P.S. I thank you for your support, it makes me really like that we're here at 21 in the evening to talk about a project to say the least sci.

 

[GJH105775]

Cost is dependant, I can get you a price when we figure exactly how to make it work with a PC.

As for size, take a look at this video, he is using an [Orange] LRS, [Orange] being the company.

Here is the transmitter module.

Here is the receiver.

You may have to get a tx to run the module, you will need a tx for testing anyways, but you may need the more expensive one that can handle this module. Ideally you could make it run directly off from a PC.

 

[w_t]

The transmitter is so small ??
With 433MHz and 100mW that we reach?
How he copes with obstacles or bumps, tops them quietly or there is possibility of pertita of receipt by the receiver?
I do not know if I understand correctly, this LRS we must apply for example in the roof, and connect it to a computer that processes all the data sent by the drone to respond with the data for the movements that have to do the drone?

Last night I thought a little to the dynamics of communications

1) developing a drone with a ground station will have more autonomy and more computing power
2) developing a drone without ground station will have less expenses, less autonomy, as more futuristic design, more charming.

Do you agree? Give me some ideas too kind strengths and weaknesses of the two methods?

 

[GJH105775]

Here is a better module. This is a 1W (1000mW)

Ground station
less cost if crash (Much of expensive stuff is on the ground)
longer flight time
maybe easier

Everything On "Drone"
No loss of control signal
there are a few other things I did think of but I can't remember them haha

 

[w_t]

OK, we decide we do with ground station,
if you're okay we can begin to do the project, type:

- Calculate the weight of the component parts of the drone:
1) GPS module
2) GPRS/GSM module
3) Raspberry PI 2 + Raspi Cam
5) Sonar module
6) Receiver module
7) Engines
8) ESC
9) Batteries
10) Chassis

- Choice of engines (as we want to speed movement believe with internal rotor)
- Choice of batteries (imagine two)
- Choice of ESC
- Choice of propellers
- Choice of board flight
- Choice of dimensions and material of the frame

I am not able to calculate all of this, I have some very basic notions, they are not sufficient.

 

[GJH105775]

I'll give you a link to a calculator for quadcopters, when I get back I am on my phone right now. But being able to actually fly that far is really a job for a plane, thouh a bigger quadcopternhas a better flight time than a smaller one. Thouh you could not operate a bigger quad in an enclosed space as they will need room for error.

 

[GJH105775]

Once you decide how you want to do this I can choose the physical parts mostly from memory, that will be the easiest part.



http://www.ecalc.ch/xcoptercalc.php?lang=fr

 

[w_t]

At this point you have to take action, as procederesti for the various analysis?

If I do it from programmer to management and Web sites, mentally I worry of the following points:

1) As it is the dynamics of the communications between PC ground and drone? So what does the one, the other, and how they do it?
2) How should the PC on the ground?
3) What operating system? (I hope Ubuntu Server)
4) in the drone there must be at least Raspberry with its IP Cam HD captures video and audio, or you thought of another method of recovery?

We can analyze in detail these points? Or is it better to first see more?

 

[GJH105775]

Can you again phrase and again translate that first sentence?

Also no IP Cam would have the required range, for that you would also need a LRS video system. I believe that anything that can be done from the ground can also be done in the air with a little modification.



I use only use Linux for even desktop, so I would be able to help you with any Linux work after you get a plan together.

 

[w_t]

At this point we decide, as we proceed with the analysis?

You have the big picture of how to the project?

Take the first step to begin designing and I will follow with my comments and ideas.

 

[GJH105775]

Perhaps, I think if you get a general design together we could work off that.