Posted by Tom Yochum on December 11, 2009 at 2:03pm
I read in Aviation Week (Nov 16th issue) about the International Astronautical Federation and the Minseeker Foundation teaming up to study landmine detection from space. Landmines are a terrible leftover from war in many developing countries. There are an estimated 100 million landmines in the world, and they kill or maim 15,000 to 20,000 people each year. Many international organizations are working on ways to efficiently remove these landmines.This got me thinking about the feasability of using an inexpensive UAV as a minesweeping platform.The biggest hurdle, or course, is developing a payload that could detect landmines from the air. I figure if someone thinks they could do this from low earth orbit, they could make a smaller package that could do it from 100 feet. Does anyone know of anyone doing research into this sort of thing? Anyone have a grad student friend looking for a research project? I am sure there is research money out there to fund this sort of project.What would be required from an aircraft perspective? Again the payload size and weight would ultimately dictate the airframe, propulsion, and other parameters. But what about the avionics? Could ArduPilot provide the navigation, sensor orientation and geolocation information with sufficient accuracy? Could it maintain a desired altitude above ground (using a laser or sonar altitude sensor)?Is this just a science-fiction fantasy?Tom
Funny you mention an Orbiter Ryan I know somebody flying one this week, will be interested to see see how he gets on. Deep stall land is possible with Atto, also pretty cool is something played with which if its windy enough is a virtual pole mode.
Just put the Atto airframe into landing assist mode and reduce throttle keep it into wind and catch.
An EPP wing with an ATTO and 5 different microSD cards with the centre NSEW stuff on would work. Change one card for the next after landing and the battery and off you go. With Atto you can add a tag to any pattern you create, ie a box search or some such and it will do that wherever you next switch on the autopilot, it re maps the coordinates. Make sure you add an R tag to the file though else the aircraft will try and fly back to wherever you set that up. Some middle eastern countries lost a very grown up UAS when they did exactly that, took off, switched the autopilots on and it headed home!!
There is a wing with a 2 hour plus endurance out of the molds now so that could do all the areas in one sweep, its designed for a BTC 88 and vertical camera for stitches.
Mission planning is good, If you can tell the UAV to go photograph an area 2km away and come back the that's great, however it does up the support and technical requirements a bit along with required training and tech support and budget. Also you may not know the long/lat of the area. In the case of a survey if you have to go there to find out the long/lat then you might as well take the UAV and do the photographs.
I should not worry too much about finding a clear area for landing. It would be rare to have an area with no gap in the cover. 90+% of the time there will be an available piece of open ground. It just might not be flat.
The simplicity of turn on, throw and have a survey based on your current location is very appealing. It totally removes the need for a ground station or planning computer with GIS support. If proximity to the affected area is an issue then we just get the UAV to photograph a bigger area. The ideal solution would to have a number of flight plans built in that could be selected prior to launch. i.e.
5 programs:
Survey with me in centre
Survey area to my North
Survey area to my South
Survey area to my East
Survey area to my west
A switch on the aircraft itself would select the area to be surveyed.
Yes, an almost disposable UAV would be ideal.
What camera do you think would be good? Can you give me a link to an example?
Remember we are not looking for mines with this system. The quality needs to be to recognise the difference between roads, trees, bushes, areas cleared, buildings etc. Number plate recognition may come late :)
After additional thinking, localized recovery is really not an option (jungle etc).
the solution appears to be a few km range, and operating 500m-2km away from 'safe' landing site.
This must be close to local area of operations for better communication, less weight, less infrastructure.
Electric for reliability and easy deployment, as a consequence below 5-8kg.
What is really needed is good mission planner and UAV cheap enough to be 'almost disposable' in case of failure: so no unnecessary equipment onboard. But the lift must be around heavier cameras if this has to be a pro platform, small cameras imply long exposition times and require calm weather.
The key is good mission pre-planning and data processing post.
Image stabilization of the cameras themselves is almost useless in the air due to centrifugal forces - most of them are based on accelerometers alone, you simply need short exposition times or very hi-tech stabilized mount.
Raven is only good if you have hardened camera, classic one will fail due to landing accelerations.
Some very good points gentlemen,
One of the biggest challenges here is going to be recovery. I was keen not to have a traditional landing as often a large enough strip will not be available, there is also the training issues. Also the vegetation in many of these areas are very tough and would destroy a UAV flying with any horizontal speed. Parachute landing would not work due to the size of parachute needed to reduce speed significantly, however that does not rule out a parachute as a guidance tool to make the UAV descend more predictably.
The reason I was going for a stall landing was I have seen it used successfully with a number of small hand launched military UAV's.
I went to an AUVSI airshow some years ago and saw many small UAV's descend to around 30 ft and then stall. There were more recovered this way than a more traditional landing and with no parachutes (although one, the Skylark, had a big airbag). Most of these UAV's had video pan-tilt units fitted and were all around the same size if not smaller as the easystar. https://www.youtube.com/watch?v=NiJv6B82DO8 https://www.youtube.com/watch?v=ty2EVkig3EU (this one is a bit extreme)
These videos are not very clear I have been looking for a better one.
I don't see this been used so close to a minefield that drift of a few meters would put it in a mined area. There has to be some operational rules applied, i.e. if the wind is towards a mined area from the launch point then you double your safety distance.
This typically would be used every day for progress reports or surveys. Most likely at the beginning or end of the day when wind conditions would be ideal (none).
The camera does not have to be anything special, we are not looking for IR or high end. As small commercial 3-5M pixel stills camera with fixed focus would be enough. I have seen many in my local shop for less than 80uk pounds that include image stabilisation.
@Richard, After going on marathon reading session in DDAaccident records ,am back here for a break :) . i agree with Bosak in both the case. The case starting from 277 might use some help with blimp :). Any way, like Bosak said, after reading around 100 cases , clearly 10 c.m of stepping on the wrong side is deadly , you may want to consider the fact that auto landing is still not a precise art in DIY , so recovery is one area you want to focus on, also keeping in mind backup plan incase the models drifts or lands behind the enemy lanes /lines :) , experienced Human eyes with stereoscopic sights backed up with most powerful computer ever misses a PMA2-3 at a distance of less than a meter :)o( , now that is clearly a great challenge of any machine( i know machine vision has advantages), i am not sure if your small foam plane will have any serious lifting power left over for serious photography after all the AP. propulsion electronic are fitted. But this could be a learning experince for you and HD team. Good luck , I am going back for more reading :)
There are 2 contradictory requirements: lifting of pro, or even IR camera AND landing in very limited space.
You also need easy in-field repairs so helicopters are not the best choice.
This calls for a plane that is probably catapulting and parachuting, but typically larger than small styrofoam RC we fly here. The best would be full manual landing (most precise, least 'collateral damage') - but this time we would add mixed skills requirement to field operators.
How often should it be used? once per week/village? Once per day documenting progress? Just thinking.
Why not to buy Raven from AeroVironment?
Hi Morli thanks for the comments,
Krzysztof I think you have hit the nail on the head.
Once you get past the misconceptions of minefields as shown in films like 'behind enemy lines' and others then the reality is very different. What I am trying to achieve is very straight forward and has been done before with UAV's but not for this application. The way the system would be used would be as follows:
1.Operator goes to a point as close as possible to the area to be photographed.
2.The UAV is assembled and hand launched under full autonomous control
3.The UAV would climb to an altitude say 300 feet and take a photo over the launch site
4.The UAV would descend to say 200 feet and photograph a 2 x 2 grid
5.The UAV would descend to say 100 feet and photograph a 4 x 4 grid
6.Finally the UAV would return to the launch point and stall at low altitude
This would drop the UAV down close to the operator who would recover the UAV and possibly a SD card containing the photos. These would then be sent to the operational office for analysis/reporting.
The UAV would then be recharged for its next flight.
This would mean aerial photographs can be taken using operators with minimal training. The autopilot would mean repeatable photograph of areas.
My proposal was sort of photomapping. Because you have to compare visible light (yeah, we have trees there), local maps (near to the road, next to the army's kitchen..) then look at IR (where things look so differently and some feature-rich areas might look uniform, others to the contrary).
I don't think you can do this reliably with FPV, too many things move around. Really, if you want to overfly the are to find 'the worst', you can take a ride with classic airplane once per month and 'get the idea' to pick the village. Yet, you can just ask locals. I think the demining operations are so time consuming, that the primary use of UAV is repetitive, reliable documentary action allowing planning of the future moves. It could also provide 'initial area scan' of all existing or abandoned trails etc but we are talking about a scale of 1x1km, resolution around 2cm per pixel, typical altitude over 50m, preferably 100-200m etc.
Comments
Just put the Atto airframe into landing assist mode and reduce throttle keep it into wind and catch.
http://vimeo.com/3507464
An EPP wing with an ATTO and 5 different microSD cards with the centre NSEW stuff on would work. Change one card for the next after landing and the battery and off you go. With Atto you can add a tag to any pattern you create, ie a box search or some such and it will do that wherever you next switch on the autopilot, it re maps the coordinates. Make sure you add an R tag to the file though else the aircraft will try and fly back to wherever you set that up. Some middle eastern countries lost a very grown up UAS when they did exactly that, took off, switched the autopilots on and it headed home!!
There is a wing with a 2 hour plus endurance out of the molds now so that could do all the areas in one sweep, its designed for a BTC 88 and vertical camera for stitches.
Flying in better weather makes for better photos!
Mission planning is good, If you can tell the UAV to go photograph an area 2km away and come back the that's great, however it does up the support and technical requirements a bit along with required training and tech support and budget. Also you may not know the long/lat of the area. In the case of a survey if you have to go there to find out the long/lat then you might as well take the UAV and do the photographs.
I should not worry too much about finding a clear area for landing. It would be rare to have an area with no gap in the cover. 90+% of the time there will be an available piece of open ground. It just might not be flat.
The simplicity of turn on, throw and have a survey based on your current location is very appealing. It totally removes the need for a ground station or planning computer with GIS support. If proximity to the affected area is an issue then we just get the UAV to photograph a bigger area. The ideal solution would to have a number of flight plans built in that could be selected prior to launch. i.e.
5 programs:
Survey with me in centre
Survey area to my North
Survey area to my South
Survey area to my East
Survey area to my west
A switch on the aircraft itself would select the area to be surveyed.
Yes, an almost disposable UAV would be ideal.
What camera do you think would be good? Can you give me a link to an example?
Remember we are not looking for mines with this system. The quality needs to be to recognise the difference between roads, trees, bushes, areas cleared, buildings etc. Number plate recognition may come late :)
the solution appears to be a few km range, and operating 500m-2km away from 'safe' landing site.
This must be close to local area of operations for better communication, less weight, less infrastructure.
Electric for reliability and easy deployment, as a consequence below 5-8kg.
What is really needed is good mission planner and UAV cheap enough to be 'almost disposable' in case of failure: so no unnecessary equipment onboard. But the lift must be around heavier cameras if this has to be a pro platform, small cameras imply long exposition times and require calm weather.
The key is good mission pre-planning and data processing post.
Image stabilization of the cameras themselves is almost useless in the air due to centrifugal forces - most of them are based on accelerometers alone, you simply need short exposition times or very hi-tech stabilized mount.
Raven is only good if you have hardened camera, classic one will fail due to landing accelerations.
One of the biggest challenges here is going to be recovery. I was keen not to have a traditional landing as often a large enough strip will not be available, there is also the training issues. Also the vegetation in many of these areas are very tough and would destroy a UAV flying with any horizontal speed. Parachute landing would not work due to the size of parachute needed to reduce speed significantly, however that does not rule out a parachute as a guidance tool to make the UAV descend more predictably.
The reason I was going for a stall landing was I have seen it used successfully with a number of small hand launched military UAV's.
I went to an AUVSI airshow some years ago and saw many small UAV's descend to around 30 ft and then stall. There were more recovered this way than a more traditional landing and with no parachutes (although one, the Skylark, had a big airbag). Most of these UAV's had video pan-tilt units fitted and were all around the same size if not smaller as the easystar.
https://www.youtube.com/watch?v=NiJv6B82DO8
https://www.youtube.com/watch?v=ty2EVkig3EU (this one is a bit extreme)
These videos are not very clear I have been looking for a better one.
I don't see this been used so close to a minefield that drift of a few meters would put it in a mined area. There has to be some operational rules applied, i.e. if the wind is towards a mined area from the launch point then you double your safety distance.
This typically would be used every day for progress reports or surveys. Most likely at the beginning or end of the day when wind conditions would be ideal (none).
The camera does not have to be anything special, we are not looking for IR or high end. As small commercial 3-5M pixel stills camera with fixed focus would be enough. I have seen many in my local shop for less than 80uk pounds that include image stabilisation.
How did you find iraq? It is an interesting place ain't it. Bomb dogs arte good but there just isn't enough to go around.
Krzys
Whats this orbiter thing? I have heard of that. Is it just a small hand launched aircraft with a ground station in the following vehicle?
You also need easy in-field repairs so helicopters are not the best choice.
This calls for a plane that is probably catapulting and parachuting, but typically larger than small styrofoam RC we fly here. The best would be full manual landing (most precise, least 'collateral damage') - but this time we would add mixed skills requirement to field operators.
How often should it be used? once per week/village? Once per day documenting progress? Just thinking.
Why not to buy Raven from AeroVironment?
Krzysztof I think you have hit the nail on the head.
Once you get past the misconceptions of minefields as shown in films like 'behind enemy lines' and others then the reality is very different. What I am trying to achieve is very straight forward and has been done before with UAV's but not for this application. The way the system would be used would be as follows:
1.Operator goes to a point as close as possible to the area to be photographed.
2.The UAV is assembled and hand launched under full autonomous control
3.The UAV would climb to an altitude say 300 feet and take a photo over the launch site
4.The UAV would descend to say 200 feet and photograph a 2 x 2 grid
5.The UAV would descend to say 100 feet and photograph a 4 x 4 grid
6.Finally the UAV would return to the launch point and stall at low altitude
This would drop the UAV down close to the operator who would recover the UAV and possibly a SD card containing the photos. These would then be sent to the operational office for analysis/reporting.
The UAV would then be recharged for its next flight.
This would mean aerial photographs can be taken using operators with minimal training. The autopilot would mean repeatable photograph of areas.
Hope this makes the project clearer.
I don't think you can do this reliably with FPV, too many things move around. Really, if you want to overfly the are to find 'the worst', you can take a ride with classic airplane once per month and 'get the idea' to pick the village. Yet, you can just ask locals. I think the demining operations are so time consuming, that the primary use of UAV is repetitive, reliable documentary action allowing planning of the future moves. It could also provide 'initial area scan' of all existing or abandoned trails etc but we are talking about a scale of 1x1km, resolution around 2cm per pixel, typical altitude over 50m, preferably 100-200m etc.
-
1
-
2
-
3
-
4
-
5
of 7 Next