Beagle Zero G Project

Hi, Everyone,

Being inspired by Heath Robinson’s amateur high-altitude balloon project “Icarus” [1], I’m proposing a similar project based on Beagleboard for the Google Summer of Code this year. I know there’s a separate mailing list for GSoC topics but since this may be interesting for a broader audience, I’d like to post my idea here.

The Icarus project used an AVR microprocessor to control a reprogrammed Cannon A560 camera to take photos every few minutes. There are also a few sensors in the payload to monitory varies environmental parameters. The whole system was wrapped in a thermal insulating box to resist the minus sixty Celsius degree temperature in the high-altitude orbit. Replacing the microprocessor with a powerful ARM processor on the Beagleboard would give the system more horse power to process more images in real time, and to reduce development period by utilizing third-party software for sensors on Linux.

This is a mix of hardware, software and meteorology project. There will be several engineering and non-engineering problems to solve as listed below. I must have missed tons of things and you are welcome to add comments. A project like Beagle Zero G requires good knowledge on multiply areas so if you are experienced in one of any related fields, please don’t hesitate to join the discussion. Thanks!

Issues I can think of right now:

  1. Cameras.

The Icarus project used a reprogrammed consumer camera to take and store all pictures. In the Beagle Zero G project, we will achieve this by utilizing three (pointing to three directions which separates by 180 degrees) or four (plus one pointing downward) webcams through USB connection. The Beagleboard will control the cameras by a fixed interval, or an adjustable intervals according to sensor data (ie shorter interval at higher altitude) and store the pictures in memory card. Both inexpensive high quality cameras and free drivers under Linux (such as this one [4]) are easy to obtain.

  1. Sensors, sensors and sensors.

The payload may need to include sensors of temperature, barometric pressure, altitude, humidity and GPS. Some sensors can be sealed in the box and others must be exposed outside. One of barometric pressure and altitude sensors may be alternative since there’s an exponential relation between the altitude and atmospheric pressure [2]. For a compact and robust design, we may need to integrate these sensors on a daughter board.

  1. Communication between the orbit module and the ground station.

The Icarus used radio teletype (RTTY) to transmit position back to the ground station. I’m not familiar with this area so need to study more on how to implement both the hardware and software. Do you have any suggestions?

  1. Balloon.

The highest altitude the Helium balloon can reach could be calculated by combining the altitude/pressure relation and the largest stress the balloon can take. The maximum stress could be tested on ground.

  1. Parachute.

I saw from a media picture [3] that in the Icarus project a parachute was placed between the balloon and the payload, as shown below,

@ balloon

^ parachute

payload box

so, we need to figure out a way to avoid the parachute being entangled by the imploded balloon.

The area of parachute needs to be carefully calculated to control the impact speed to a certain range. If the landing speed is to high, it may damage the system and harm security of beagles and beagles’ friend (aka human) on the ground. An insufficient falling speed may let the wind blow it too far away from the launch site.

  1. Working at low temperature and low air pressure.

I don’t have data on how low temperature and pressure the beagle board can work under, but this can be tested on ground before the launch. It’s possible to completely seal the box to prevent losing air pressure.

  1. Insulating box.

An insulating box is required to resist the cold (due to high altitude) and heat (on the side of direct exposure under sunlight). Cosmic ray may bring error bits to the memory or even destroy part of the circuit so a metal shield may be needed (but it would reduce useful payload and the altitude the balloon can reach).

  1. Power.

Depends on the time of flight, the Beagleboard itself may not require much battery power but it could be a problem if we put too much sensors and cameras on board.

  1. Launch permission.

It’s reported that Heath Robinson had to obtain permission from Civil Aviation Authority to launch the balloon in UK. Not sure if there’s similar regulation in the United States where I’m currently living in. Giving the fact that I’m not an US citizen, it might be even much harder to get such permission in the states. I know in some countries, like China where I’m from, lifting a payload of less than 4kg and 3.2 m^3 don’t have to ask for permission from the government [5].

References:

[1] http://www.robertharrison.org/icarus/wordpress

[2] http://en.wikipedia.org/wiki/Barometric_formula

[3] http://www.timesonline.co.uk/tol/news/science/space/article7074839.ece

[4] http://mxhaard.free.fr

[5] http://www1.china.com.cn/chinese/law/264778.htm

Sincerely,

I like this idea and list of issues to address, but it seems to fall short on generating new reusable design materials. If the sensors are widely available and you do some good collection/analysis software, that might be good enough. The RTTY software might be of enough value, if it is something new and repeatable. You might even write modem code on the ARM or DSP. Just be sure to generate something that several people can reproduce based on your code/design.

Have you been able to identify a mentor?

It is strangely apt that you have mistakenly referred to Robert
Harrison as Heath Robinson throughout this message:

http://en.wikipedia.org/wiki/W._Heath_Robinson#In_popular_culture

Oops. I got the first name wrong. It should be Robert Harrison. Please
forgive my bad recognizance of English names. There was no pun
intended.

Sorry for the repost but since my reply to Jason on the beagleboard-
gsoc list might be useful to others:

Hi, Jason,

Thanks for the comment. The first stage of the project will be
selecting
sensors/cameras and writing drivers for them (if the drivers are not
available), which I think is reusable (especially the GPS module) in
other
projects thus good enough as a GSoC project. And as you have said,
data
analysis, RTTY and modem software could also be done as a part of the
final
deliverable. I'm not sure how much percent will the project to be
reused in
other things but there would definitely be a lot folks wanting to
reproduce
it with Beagleboard if we made it a success (showing off pictures
taken by
budget devices from earth orbit in front of friends is hard-core geek,
isn't
it?)

To answer the mentor question: I haven't got a mentor yet. This is the
third
idea I proposed in this mailing list but obviously the first two
(emulator
and JTAG) didn't interest anyone. I believe the way of getting ideas
in GSoC
projects is a mutual selection that both students and mentors should
work on
the project they are mostly interested in.

br,
hui

Do you really mean RTTY? That stands for radio teletype, and pretty
much by convention is limited to speeds in the 5.62 - 10 bits/second
range. if you mean to apply that term to mean radio binary frequency
shift modulation, then it makes more sense, as you probably can't
send more than one color image per HOUR by real RTTY!

I seriously doubt the Beagle would be harmed by low pressure, we
routinely run electronics in low-pressure environments (in chambers,
however). The one area of concern is running higher-voltage circuits
in relatively low pressures, where corona discharges can cause
damage. But, the beagle runs mostly on 1.8 V. Batteries could be a
problem at low pressure. Steel-walled cells like AA should handle it
fine, big plastic gel-cells might vent or burst at low pressure.

A GPS gives true (geodetic) altitude. But, you can measure pressure
with an encoding altimeter. General Aviation altimeters usually only
read to 20,000 feet, so you need a jet or military version. For
coarse readings, an absolute pressure transducer will work fine.

You most CERTAINLY will need FAA approval to launch a balloon with a
significant payload. You will need to provide an FAA transponder and
enough battery life to keep it running throughout the flight. You
will need to file a flight plan and give contact info so the air
traffic controllers can contact you if you need to "shoot it down".
And, you will need to provide a shoot-down system to bring the balloon
down by remote command. Systems I've seen used a radio-controlled hot
wire taped to the balloon. On command, or by end of flight timer, the
heater melts a hole in the balloon to bring it down. These systems
were used in air pollution experiments in the 1970s, and had a payload
that was built into a picnic cooler. There was no parachute, the
location of the shoot down wire was calculated so as to let out JUST
enough helium the make the system less than bouyant. The helium above
the wire was mostly retained during the descent.

Jon

Hi, Jon,

Thanks for the information! It’s very helpful!

Do you really mean RTTY? That stands for radio teletype, and pretty
much by convention is limited to speeds in the 5.62 - 10 bits/second
range. if you mean to apply that term to mean radio binary frequency
shift modulation, then it makes more sense, as you probably can’t
send more than one color image per HOUR by real RTTY!

There’s probably no need to send back images/videos to ground station in real-time. We are not going to broadcast it live to the whole world. (Or maybe we should? :slight_smile: ) Only vital information will be transmitted, such as position, temperature, and the “shoot-down” signal you have mentioned.

I seriously doubt the Beagle would be harmed by low pressure, we
routinely run electronics in low-pressure environments (in chambers,
however). The one area of concern is running higher-voltage circuits
in relatively low pressures, where corona discharges can cause
damage. But, the beagle runs mostly on 1.8 V. Batteries could be a
problem at low pressure. Steel-walled cells like AA should handle it
fine, big plastic gel-cells might vent or burst at low pressure.

Thanks for the good news on low temperature performance. Yes, I doubt there will be a high-voltage situation if it’s launched on a clear day. What I have read from past high-altitude balloon projects (HALO2 and Icarus) is that a flight usually takes less than three hours and the internal temperature dropped to minus Celsius degree for much less time. Some data from Icarus project:

http://www.flickr.com/photos/30721501@N05/4111728280/
(Fig. 1)

A commenter in the page gave a good reference on the temperature as a function of altitude:

http://www.classzone.com/books/earth_science/terc/content/investigations/es1702/es1702page05.cfm
(Fig. 2)

So in the lifting phase, the external temperature shown in Figure one is about 40 Celsius degrees higher than that in Figure two which was measured by a radiosonde (not sure how it’s working). I guess the reason of the temperature difference is that the box was heated by sunlight and low pressure still air is a good heat insulator. My explanation to the great dip of temperature in the falling phase as shown in figure two is

  1. contrary to the “common sense” that a landing object will be heated up, though still low pressure air is a good heat insulator because the air particles have low speed so the capability to bring out heat from the box (through collision with the atoms on the box surface) is limited, a wind with high speed (~70mph) would actually be able to effectively bring box temperature towards environment temperature because of gained dynamic energy of air particles.

  2. Quite unlikely, the imploded balloon or parachute covered on the box and prevented it being heated by sunlight.

BTW, the “Wind chill is cooler” theory only works for animal skin where evaporated water can bring heat out. It doesn’t work for dry objects.

Maybe we can preheat the battery and isolate it from other components. I know they discharged quicker at low temp but I still need to do some research on the details.

A GPS gives true (geodetic) altitude. But, you can measure pressure
with an encoding altimeter. General Aviation altimeters usually only
read to 20,000 feet, so you need a jet or military version. For
coarse readings, an absolute pressure transducer will work fine.

You are right. GPS only works under certain altitude, and will fail when it reaches the plane of medium earth orbit satellites (about 20km from sea level). Beyond this limit, the altitude could be inaccurately derived from other parameters such as air pressure, like this one

http://www.flickr.com/photos/54644437@N00/1782292839/

You most CERTAINLY will need FAA approval to launch a balloon with a
significant payload. You will need to provide an FAA transponder and
enough battery life to keep it running throughout the flight. You
will need to file a flight plan and give contact info so the air
traffic controllers can contact you if you need to “shoot it down”.
And, you will need to provide a shoot-down system to bring the balloon
down by remote command. Systems I’ve seen used a radio-controlled hot
wire taped to the balloon. On command, or by end of flight timer, the
heater melts a hole in the balloon to bring it down. These systems
were used in air pollution experiments in the 1970s, and had a payload
that was built into a picnic cooler. There was no parachute, the
location of the shoot down wire was calculated so as to let out JUST
enough helium the make the system less than bouyant. The helium above
the wire was mostly retained during the descent.

I’m worried by this part. Is it possible for a non US citizen to get the permit? Maybe I must find a US mentor? Thanks for telling me the “shoot it down” command thing since I haven’t heard about it before.

Another thing I’m think is the reliability of the project. The system may need to reload some sensors (such as GPS, which probably takes much longer time than others) or even reboot itself (then my past fast-boot swiftbeagle project will be useful :slight_smile: ). The shoot-down system may also need backup plan, that if the operation system is not working, the explosion could be trigger by an independent timer, or a fire arrow shot with a cross-bow by my mentor from the ground.

BR,
Hui

There is no permit necessary to launch a balloon in the US. You might
be able to ask someone from the FAA about a good launch point though,
just use common sense, like not launching in or near controlled airspace.
There also is no requirement for a transponder or an early descent trigger
since it would be totally impractical for this application. Hundreds of radiosondes
are launched every day worldwide; to the best of my knowledge, none have ever collided mid-air with aircraft.

As for GPS tracking; if you’re launching in the US, I would recommend using APRS which uses a network of amateur
radio repeaters at 144MHZ to relay tracking data to the web.

This device will encode the GPS data to an APRS signal for you:

http://www.byonics.com/tinytrak/

Then you can track via the web here:

http://aprs.fi/

What altitude are you hoping to reach?

There are some balloons that can theoretically reach 30 km (unlikely in reality)

Hope this helps,
Tristan

Hi, Tristan,

Thanks for the good news on launch permit and I will have a look at the websites you mentioned. I expect it to reach about 20km-25km so height shouldn’t be a problem for the balloon.

However I haven’t got any comments under my proposal in Google’s site yet. It would be really better if a mentor can pick it up and help me to get fund from Google for this project.

Best Regards,
Hui

Hi, Jason,

Thanks for the comment. The first stage of the project will be selecting sensors/cameras and writing drivers for them (if the drivers are not available), which I think is reusable (especially the GPS module) in other projects thus good enough as a GSoC project. And as you have said, data analysis, RTTY and modem software could also be done as a part of the final deliverable. I’m not sure how much percent will the project to be reused in other things but there would definitely be a lot folks wanting to reproduce it with Beagleboard if we made it a success (showing off pictures taken by budget devices from earth orbit in front of friends is hard-core geek, isn’t it?)

To answer the mentor question: I haven’t got a mentor yet. This is the third idea I proposed in this mailing list but obviously the first two (emulator and JTAG) didn’t interest anyone. I believe the way of getting ideas in GSoC projects is a mutual selection that both students and mentors should work on the project they are mostly interested in.

br,
hui

I have put the draft of my proposal with answered questions in Google
Summer of Code's website. Possible mentors can go there and have a
look. Thanks!

Great project idea. A good mentor would be important here.