This semester we try to build autonomous quadcopter for delivery, our aim is carrying 1 kg load.We don’t have much experiences about quadcopters. For our design we selected
Frame :Tarot Ironman 650,
Motors: Sunnysky X4110S 400kv
Propellers: APC 14×4.7
ESC: 30 A Simonk ESC
Battery:3S 6000-8000 mah
Flight controller: pixhawk or nasa
We will probably use rasperi pi and camera for self landing(For pattern recognition ).Our final weight with 1kg load is approximately 4/4.5 kg.Are our choices for drone parts okay? Also which flight controller do you recommend for us?
Thank you for answers.
This sounds like an awesome project! Most modern FCs will be able to send communicate to the raspberry pi as they all tend to have UART, SPI and I2C interfaces. Pixhawk and APM flight controllers are superior for this type of interfacing though. There are a lot of easy and sturdy connections that can be interfaced to the pi along with a large community of people using them. They are also designed to be interfaced to external devices while managing flight control systems. Take a look at some info on them here. There is also some more open source packages and community info for this type of tech and hardware here too.
Also, if you are aiming to train for real time object recognition for computer vision look at the new TensorFlow Object Detection API released back in July.
It was made for mobile deployment and I have seen some people integrating it with Raspberry Pis. I am using it currently for some traffic recognition systems and it works wonderfully. Accurate and easy to train.
Hi, I think you are on the right track- Remember that a good rule of thumb is to engineer the quad so you have 4x the weight in thrust. So if your quad with full load weighs 1.5KG, then you want 6kg of thurst for a good flying & maneuverable quad.
this motor outputs about 2030g at 14.8 (4S) voltage with a 16×5.4 prop. So, 4 motors will output 8 kg of thurst- which would be a great choice.
We are having a massive sale on low-kv motors- you might find something better than sunnysky in our motors area, and we can offer an Edu discount if this is for a student project and will be purchased by your school.
I think @AlexGrand can weigh in on the Rpi and computer vision- along with flight controller recommended as he is the resident expert on the subject.
Let me know if that helps to answer your questions!