To reach the goal of the project, we have 3 main technical challenges : 

The first technical challenge : Deploy in a safe way a parachute canopy 

In its initial state,  the whole system fit in a tube of an approximative size of 40cm by a diameter of 7cm. 

And what we want, is to, from this tube, deploy, when we want, a steerable, parachute canopy, also known as a ram-air parachute canopy. 

We also want this parachute canopy to be able to fly in a sufficient enough way, which mean with a high enough glide ratio, and flight speed. 

To solve the first challenge, I first started to use a very small paraglider canopy, it flew very well, but I quickly found an important factor for the deployment of the canopy, the aspect ratio, a paraglider or parachute canopy with an aspect ratio > 3 is very difficult to deploy. 

So I decided to start sewing my own parachute canopies to solve this problem. 

After few canopy sewed (you can count a week of work for the fabrication of a parachute canopy) the deployment of the parachute canopy even in bad conditions (bad folding, no forced deployment etc.) started to go perfectly well (see this deployment test video :

Then, once we have a parachute canopy that deploys properly we have the second technical challenge, to be able steer it

For this we have to be able to pull more or less in a coordinated way on the 2 brake lines of the parachute. 

By pulling on the left rear of the parachute and releasing the right rear, we break the left part and accelerate the right part of the parachute, this makes it turn left for example. 

Pulling both lines at the same time will break the parachute, and releasing both lines at the same time will make it fly faster.

It was therefore necessary to find a system allowing to pull more or less 2 ropes/ lines, each one on about twenty centimeters to have enough control over the parachute canopy. 

The solution chosen was to 'hack' servomotors originally designed for rc sailboats, they are sold with a drum spool, and modified inside to be able to still with a closed loop control, make up to 4 turns in each direction. You can see them in action on this video :

Few pictures of the last prototype of the system including the mechanical part with the two servomotors 

Finally the last challenge is to automatically steer the parachute canopy, and this challenge is not yet fully fully solved. 

The idea is the following :

By using a GPS device, we can know where we are in real time, and we know in principle where we want to go, so we can know the angle in which we want to move. 

But then to really move in this direction, we need to know in which direction we are actually moving in reality, so that we can then calculate an error, and a correction to get closer to the direction objective. 

And so the main hard point is, from a certain number of sensors, to deduce in the most precise way possible, the direction in which we are really moving. 

For this there are several solutions : 

- The first is that when there is a little wind, the parachute can drift with the wind, and therefore to go in a given direction, it may for example have to point in another direction. 

- The second problem is that the tube in which the magnetometer could be located moves a bit in all directions, and therefore the magnetometer can very quickly become useless. 

However, this technique is totally useless at a speed slower than about 10km/h, speed for which the GPS points are too close, and the GPS error makes the angle too often wrong. 

Maximum GPS measured orientation error (in °) vs horizontal speed in km/h : 

- A solution to this problem could be to make the parachute fly faster, for that we have to increase the wing loading of the parachute, this is actually in test with a smaller canopy. 

- Another solution would be to mix the magnetometer and the GPS according to the speed, so that the magnetometer is used when the parachute is not flying very fast, and the GPS when the parachute is flying faster. 

Yet another solution would be to incorporate the magnetometer into the sail itself to suppress the parasitic movement of the tube which moves in all directions. 

Then in terms of the control algorithm, it's actually very simple, it's just a proportional controller. It has already been successfully tested on a small remote-controlled car that drives itself to a GPS point (at the end of the street) as you can see in this video :

A picture of the prototype of flight computer (Ft' Teensy 4.1) : 

That's it for the actual technical details ! 

You will soon find a project update on this page talking about the actual project state !

Thanks for your interest !

And if you want to support this project, take a look at it's gofundme page :

For more update you can also follow the project on social media : 

Youtube :

Twitter :

Instagram :

Facebook :