handle high speed jet

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

handle high speed jet

russell neemo
i want to handle my jet model by kroozSD . my jet goes over 350 km/h 
what kind of servos,airspeed sensors and AHRS sensors is compatible for my model.
if it possible please give me some example. 

thanks in  advance
Russell nimo

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Michal Podhradsky
Hi Russell,

I will be delighted to hear what others think, but putting Paparazzi based AP on such a jet sounds pretty risky to me. I assume there will be lots of vibrations due to jet engine, so you will need higher quality sensors for start. Also at this speed you might get easily out of the linear flight envelope during maneuvers (somebody who knows more about aerodynamics might correct me), so a simple PID control might not be sufficient in auto1/2 mode. Gain scheduling should help though.

Also I am not sure if the current AHRS algorithms can handle such velocities and accelerations.

Besides, if this plane looses control, it might be difficult to ever find it again...

What jet model are you using?

Regards
Michal


On Mon, Sep 23, 2013 at 3:04 AM, russell neemo <[hidden email]> wrote:
i want to handle my jet model by kroozSD . my jet goes over 350 km/h 
what kind of servos,airspeed sensors and AHRS sensors is compatible for my model.
if it possible please give me some example. 

thanks in  advance
Russell nimo

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel



_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Markus Bina
Hi Michal, Russel,

> I assume there will be lots of vibrations due to jet engine, so you will need higher quality sensors for start.
That might be a problem. Maybe the influence on the acceleration sensors and gyros can be minimized by cushening the IMU
such that the most dominant frequencies get attenuated (just enough)!?
Also I'd use an airflow sensor to improve the control loops.
Another way around the vibrations would be to use filters or even adaptive filters.
Anyway imho the plane should be designed such, that any vibrations or flutter is minimized.

> so a simple PID control might not be sufficient in auto1/2 mode
Imho: It should be well tuned.
I think auto 1 (attitude control) should work just fine as long as the aircraft is dynamically stable.
Although at 300 km/h one will surely loose the line of sight to the aircraft pretty soon.
Thus I'd be more concerned about the telemetry, since most of the radios for RC-craft (I've used and seen)
are made for reliable transmissions in a line of sight.

> Gain scheduling should help though.
I was also thinking of this, but I'm lacking experience on this matter.

> Also at this speed you might get easily out of the linear flight envelope during maneuvers (somebody who knows more about aerodynamics might correct me)
Well afaik even the simplest system of equations describing fixed-wing UAVs or MAVs are non-linear.
Imho that's why the control loops are normally split into two loops (horizontal and vertical).
In case the "flight mechanics" of this model plane are truely "grossly non-linear", it would be - in my opinion - a waste of money to
even buy it as an RC craft since as a human being you are surely unable to fly it well without any autopilot.
On youtube there are quite a few - imho genuine - videos of (experienced?) people flying RC jetplanes manually:
  eg.: http://www.youtube.com/watch?v=cqCo9QZKy80

@ Russel
If you are a beginner when it comes to ppz, I'd suggest that you start with a slower aircraft.
I recommend the Bixler/EasyJet :)
It is very robust, mine still flies well although I introduced it to quite a number of trees.

Cheers,
  Markus

Am 23.09.2013 18:49, schrieb Michal Podhradsky:
Hi Russell,

I will be delighted to hear what others think, but putting Paparazzi based AP on such a jet sounds pretty risky to me. I assume there will be lots of vibrations due to jet engine, so you will need higher quality sensors for start. Also at this speed you might get easily out of the linear flight envelope during maneuvers (somebody who knows more about aerodynamics might correct me), so a simple PID control might not be sufficient in auto1/2 mode. Gain scheduling should help though.

Also I am not sure if the current AHRS algorithms can handle such velocities and accelerations.

Besides, if this plane looses control, it might be difficult to ever find it again...

What jet model are you using?

Regards
Michal


On Mon, Sep 23, 2013 at 3:04 AM, russell neemo <[hidden email]> wrote:
i want to handle my jet model by kroozSD . my jet goes over 350 km/h 
what kind of servos,airspeed sensors and AHRS sensors is compatible for my model.
if it possible please give me some example. 

thanks in  advance
Russell nimo

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel




_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Michal Podhradsky
Hi Markus,

I will make a promo for imho a very good book about advanced controls for UAVs, as it might benefit others.
Lawretsky E., Wise K.: Robust and adaptive control with aerospace applications
http://www.springer.com/engineering/control/book/978-1-4471-4395-6

It came out in 2013 and describes optimal, robust and eventually adaptive control strategies for UAVs and manned airplanes. With lots of practical examples. You will find there more information about gain scheduling too.

I guess you are right and I should correct myself - the RC jet plane will be stable so it is possible to fly it manually, and for autopilot you can linearize the flight dynamics equations around some operation point, so you'll be able to get the right gains for a given part of flight envelope. And if you are lucky enough you might be able to fly just with one set of gains (even though the performance wont be great).

There is some gain scheduling implemented for Quadshot from Transition Robotics, but I don't know the details (you can check the code though).

M


On Mon, Sep 23, 2013 at 11:34 AM, Markus Bina <[hidden email]> wrote:
Hi Michal, Russel,


> I assume there will be lots of vibrations due to jet engine, so you will need higher quality sensors for start.
That might be a problem. Maybe the influence on the acceleration sensors and gyros can be minimized by cushening the IMU
such that the most dominant frequencies get attenuated (just enough)!?
Also I'd use an airflow sensor to improve the control loops.
Another way around the vibrations would be to use filters or even adaptive filters.
Anyway imho the plane should be designed such, that any vibrations or flutter is minimized.


> so a simple PID control might not be sufficient in auto1/2 mode
Imho: It should be well tuned.
I think auto 1 (attitude control) should work just fine as long as the aircraft is dynamically stable.
Although at 300 km/h one will surely loose the line of sight to the aircraft pretty soon.
Thus I'd be more concerned about the telemetry, since most of the radios for RC-craft (I've used and seen)
are made for reliable transmissions in a line of sight.


> Gain scheduling should help though.
I was also thinking of this, but I'm lacking experience on this matter.


> Also at this speed you might get easily out of the linear flight envelope during maneuvers (somebody who knows more about aerodynamics might correct me)
Well afaik even the simplest system of equations describing fixed-wing UAVs or MAVs are non-linear.
Imho that's why the control loops are normally split into two loops (horizontal and vertical).
In case the "flight mechanics" of this model plane are truely "grossly non-linear", it would be - in my opinion - a waste of money to
even buy it as an RC craft since as a human being you are surely unable to fly it well without any autopilot.
On youtube there are quite a few - imho genuine - videos of (experienced?) people flying RC jetplanes manually:
  eg.: http://www.youtube.com/watch?v=cqCo9QZKy80

@ Russel
If you are a beginner when it comes to ppz, I'd suggest that you start with a slower aircraft.
I recommend the Bixler/EasyJet :)
It is very robust, mine still flies well although I introduced it to quite a number of trees.

Cheers,
  Markus

Am 23.09.2013 18:49, schrieb Michal Podhradsky:
Hi Russell,

I will be delighted to hear what others think, but putting Paparazzi based AP on such a jet sounds pretty risky to me. I assume there will be lots of vibrations due to jet engine, so you will need higher quality sensors for start. Also at this speed you might get easily out of the linear flight envelope during maneuvers (somebody who knows more about aerodynamics might correct me), so a simple PID control might not be sufficient in auto1/2 mode. Gain scheduling should help though.

Also I am not sure if the current AHRS algorithms can handle such velocities and accelerations.

Besides, if this plane looses control, it might be difficult to ever find it again...

What jet model are you using?

Regards
Michal


On Mon, Sep 23, 2013 at 3:04 AM, russell neemo <[hidden email]> wrote:
i want to handle my jet model by kroozSD . my jet goes over 350 km/h 
what kind of servos,airspeed sensors and AHRS sensors is compatible for my model.
if it possible please give me some example. 

thanks in  advance
Russell nimo

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel




_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel



_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Christophe De Wagter
In principle a fast jet is not much more difficult for the autopilot.

But:

- Jets are dangerous. Do not fly it until you are very experienced. (Your question about the servo's does not suggest this)
- Beware of high frequency vibrations in the range of 27kHz. The are deadly for gyros as they directly interfere with the measurement
- Beware of prolonged longitudinal acceleration: enable ahrs longitudinal acceleration correction.
- Beware: many ahrs do not correct anymore with high gforce. If that happens like most of the flight you have a problem. Enable ahrs health checking.
- Finally you will need to add gain sheduling if you need to be able to fly at high and low speeds. Start with low gains as it is easy to disintegrate a jet in air by steering to roughly.

On Monday, September 23, 2013, Michal Podhradsky <[hidden email]> wrote:
> Hi Markus,
>
> I will make a promo for imho a very good book about advanced controls for UAVs, as it might benefit others.
> Lawretsky E., Wise K.: Robust and adaptive control with aerospace applications
> http://www.springer.com/engineering/control/book/978-1-4471-4395-6
>
> It came out in 2013 and describes optimal, robust and eventually adaptive control strategies for UAVs and manned airplanes. With lots of practical examples. You will find there more information about gain scheduling too.
>
> I guess you are right and I should correct myself - the RC jet plane will be stable so it is possible to fly it manually, and for autopilot you can linearize the flight dynamics equations around some operation point, so you'll be able to get the right gains for a given part of flight envelope. And if you are lucky enough you might be able to fly just with one set of gains (even though the performance wont be great).
>
> There is some gain scheduling implemented for Quadshot from Transition Robotics, but I don't know the details (you can check the code though).
>
> M
>
>
> On Mon, Sep 23, 2013 at 11:34 AM, Markus Bina <[hidden email]> wrote:
>>
>> Hi Michal, Russel,
>>
>> > I assume there will be lots of vibrations due to jet engine, so you will need higher quality sensors for start.
>> That might be a problem. Maybe the influence on the acceleration sensors and gyros can be minimized by cushening the IMU
>> such that the most dominant frequencies get attenuated (just enough)!?
>> Also I'd use an airflow sensor to improve the control loops.
>> Another way around the vibrations would be to use filters or even adaptive filters.
>> Anyway imho the plane should be designed such, that any vibrations or flutter is minimized.
>>
>> > so a simple PID control might not be sufficient in auto1/2 mode
>> Imho: It should be well tuned.
>> I think auto 1 (attitude control) should work just fine as long as the aircraft is dynamically stable.
>> Although at 300 km/h one will surely loose the line of sight to the aircraft pretty soon.
>> Thus I'd be more concerned about the telemetry, since most of the radios for RC-craft (I've used and seen)
>> are made for reliable transmissions in a line of sight.
>>
>> > Gain scheduling should help though.
>> I was also thinking of this, but I'm lacking experience on this matter.
>>
>> > Also at this speed you might get easily out of the linear flight envelope during maneuvers (somebody who knows more about aerodynamics might correct me)
>> Well afaik even the simplest system of equations describing fixed-wing UAVs or MAVs are non-linear.
>> Imho that's why the control loops are normally split into two loops (horizontal and vertical).
>> In case the "flight mechanics" of this model plane are truely "grossly non-linear", it would be - in my opinion - a waste of money to
>> even buy it as an RC craft since as a human being you are surely unable to fly it well without any autopilot.
>> On youtube there are quite a few - imho genuine - videos of (experienced?) people flying RC jetplanes manually:
>>   eg.: Jet Cat Engine - RC Plane (ORIGINAL SOUND)
>>
>> @ Russel
>> If you are a beginner when it comes to ppz, I'd suggest that you start with a slower aircraft.
>> I recommend the Bixler/EasyJet :)
>> It is very robust, mine still flies well although I introduced it to quite a number of trees.
>>
>> Cheers,
>>   Markus
>>
>> Am 23.09.2013 18:49, schrieb Michal Podhradsky:
>>
>> Hi Russell,
>>
>> I will be delighted to hear what others think, but putting Paparazzi based AP on such a jet sounds pretty risky to me. I assume there will be lots of vibrations due to jet engine, so you will need higher quality sensors for start. Also at this speed you might get easily out of the linear flight envelope during maneuvers (somebody who knows more about aerodynamics might correct me), so a simple PID control might not be sufficient in auto1/2 mode. Gain scheduling should help though.
>>
>> Also I am not sure if the current AHRS algorithms can handle such velocities and accelerations.
>>
>> Besides, if this plane looses control, it might be difficult to ever find it again...
>>
>> What jet model are you using?
>>
>> Regards
>> Michal
>>
>>
>> On Mon, Sep 23, 2013 at 3:04 AM, russell neemo <[hidden email]> wrote:
>>>
>>> i want to handle my jet model by kroozSD . my jet goes over 350 km/h 
>>> what kind of servos,airspeed sensors and AHRS sensors is compatible for my model.
>>> if it possible please give me some example. 
>>> thanks in  advance
>>> Russell nimo
>>> _______________________________________________
>>> Paparazzi-devel mailing list
>>> [hidden email]
>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>
>>
>>
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>
>

--
-Christophe 


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Iman
Hello

I want to build a luncher for my AC by air pressure or other way, but I have a question:

How much is the highest first acceleration could applied to plane(or sensors) to keep sensors safe?

if I lunch it by a large G value , could it cause problem in filtering algorithms?

Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Reto Büttner
Hi Iman
 
I recommend traditional bungee launch. This works just fine both for small and large planes, is robust, cheap and maintenance-free. This simple solution lets you focus on more critical parts of the UAS.
 
We always launched in manual mode, but many others have done it autonomously, e.g. OSAM. They might provide some tips and tricks.
 
Regards
Reto
 


2013/9/25 Iman <[hidden email]>
Hello

I want to build a luncher for my AC by air pressure or other way, but I have
a question:

How much is the highest first acceleration could applied to plane(or
sensors) to keep sensors safe?

if I lunch it by a large G value , could it cause problem in filtering
algorithms?





--
View this message in context: http://paparazzi.517.n7.nabble.com/handle-high-speed-jet-tp13617p13641.html
Sent from the paparazzi-devel mailing list archive at Nabble.com.

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Ben Laurie



On 25 September 2013 09:36, Reto Büttner <[hidden email]> wrote:
Hi Iman
 
I recommend traditional bungee launch. This works just fine both for small and large planes, is robust, cheap and maintenance-free. This simple solution lets you focus on more critical parts of the UAS.
 
We always launched in manual mode, but many others have done it autonomously, e.g. OSAM. They might provide some tips and tricks.

Who are OSAM?


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Reto Büttner


2013/9/25 Ben Laurie <[hidden email]>



On 25 September 2013 09:36, Reto Büttner <[hidden email]> wrote:
Hi Iman
 
I recommend traditional bungee launch. This works just fine both for small and large planes, is robust, cheap and maintenance-free. This simple solution lets you focus on more critical parts of the UAS.
 
We always launched in manual mode, but many others have done it autonomously, e.g. OSAM. They might provide some tips and tricks.

Who are OSAM?


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel



_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Chris Gough-2
In reply to this post by Ben Laurie

> Who are OSAM?

http://www.engr.usu.edu/wiki/index.php/OSAM

This is their bungee launch stuff

 http://paparazzi.enac.fr/wiki/Advanced_Navigation_Routines#Bungee_Takeoff

Chris Gough

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: handle high speed jet

Christophe De Wagter
For the AHRS filter a 100g 10 millisecond launch is much better than a 1m/s2 10 second launch.

atan2(1g,1g) = 45deg error
atan2(1g, 10g) = 85 deg error
atan2(1g, 1000g) = 89 deg error

-Christophe 


On Wed, Sep 25, 2013 at 1:48 PM, Chris Gough <[hidden email]> wrote:

> Who are OSAM?

http://www.engr.usu.edu/wiki/index.php/OSAM

This is their bungee launch stuff

 http://paparazzi.enac.fr/wiki/Advanced_Navigation_Routines#Bungee_Takeoff

Chris Gough

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel