BARNARD MICROSYSTEMS LIMITED

generating ideas, creating  software and designing innovative products...

introduction

Grand Challenge

Here is some of the data we collected at the second  meeting of the Teams in the UK MoD Grand Challenge at Copehill Down on Monday 15th October, 2007.

• Data collected
• Lessons learned
• Meeting at Barnard Microsystems London office
• EDA signs contract for ...  anti-sniper system 

Josh Levine did an excellent job of generating a large composite map of Copehill Down from Google Earth. The final image was 6,912 x 4,864 pixels in size, with the JPEG version being 30.909 MBytes and the PNG version being 53.515 MBytes in size. Below is a reduced resolution version of the map (the full version would lake a long time to download).

A  magnification of the central region (shown above within a yellow box) of the map, shown below, illustrates the high resolution one can obtain using Google Earth for some regions on earth.

However, not all features are discernable from the Google Earth map. From the superposition of a GPS track on the Google Earth map shown below, the deviation seen around the middle of the image  in walking from the car park (lower open area) was caused by the need to bypass a barrier across the road. Each green dot represents a data point taken each second.

The various overhead wires going from building-to-building are also not visible in the Google Earth images, so we spent time locating all the overhead wires (shown in thick yellow in the above image) we could, and have added this information to our maps. Our conclusion on viewing a superposition of our GPS tracks (shown in green) on the map with the location of overhead wires identified is that the wires do not degrade the GPS location information in any noticeable way. Basically, we do not see any discontinuity in indicated location as we pass underneath any of the overhead wires. Note that our view is based on the use of a high sensitivity GPS unit from u-Blox.

Google Earth imagery is  not up to date. In the image below with the superimposed, green,  GPS track showing where I walked, there are clearly a set of containers at eight o'clock on the overall image of Copehill Down that have since been moved to a new area. On 15th October, 2007, this was an open road.

The raw accuracy of the GPS location information is shown in the image below where I used a high sensitivity u-Blox unit attached to a PC and walked within 1 metre of the walls of various houses and buildings, always remaining on the outside. The thin green tracks show where I walked, clearly erroneously indicating me moving into and out of some houses and buildings.

In the portion of the Google Earth image of the car park, we show the tracks for a physically stationary u-Blox high sensitivity GPS receiver with the antenna fixed on top of  the roof of our  car. The tracks, shown in green,  are for a period from 10:37:10 am to 12:59:37 am and the wander in the reported location lies within a box shown below.

The potential exists to improve the accuracy of the location reported by a GPS  receiver  through the use of the data from a fixed GPS receiver. This technique is often referred to as " differential GPS" . This is ongoing work, and we hope to be able to report on progress later.

We defined 92 locations from which to take a set of panoramic images using a professional, 10 MPixel,  Canon DSLR camera. Josh Levine and Chris spent from just after 10:00 am to 4:30 pm generating many gigabytes of imaging information, with all the images taken in the RAW format. The location coverage order is as shown in the map above, with some locations turning out to be in out-of-bound areas. The panoramic images from each location will be corrected for light level, converted to JPEG format and then stitched together to form one panoramic image per location.

We identified the unique buildings and houses, and photographed  the four sides of each unique building. For example, below we show the front side of Building 6 (marked as " X" on the left in the map above) with me holding a Super A3 (or A3+) sized copy of the map shown above. This information, together with the detailed panoramic images, will help us (and others, such as MBDA)  construct an accurate,  3D, " synthetic environment" for Copehill Down.

What were we not able to do?

All the above work took us longer than I envisaged. Tasks we did not have time to perform were:

• mapping the signal coverage  for GSM 900 and GSM 1800 phones
• mapping the signal coverage  for WLAN 802.11 signals

We my need to rely on the construction of a 3D model to enable us to predict the signal coverage we did not have time to measure.

return to top

1. Plan the task to be performed in detail. For example, we had drawn up a map indicating the route to be followed for the close-to-building GPS mapping, simply showing the order in which to proceed from building to building. In reality, we needed / wanted to walk round each building, recording a GPS track. The detail of the walk then threw our plan into a bit of disarray, requiring some ad-hoc planning which is no substitute for more detailed planning in advance. Below you can see the simple plan we had and the reality of the walk around. Due to time pressures and the absence of a detailed plan to manage the walk around, the actual route started to deviate more and more with time from that planned. The more we deviated, the more time we needed to think where to go next (as evidenced by a cluster of dots on the green GPS track), wasting valuable time.

An example of good planning, execution and versatility in the face of unexpected obstacles was the panoramic imaging work. All the locations that could be  covered, were covered before 4:30 pm. Where Josh and Chris encountered " out-of-bounds" areas not obvious from the Google Earth map, they made changes and recorded them on the laminated map using an indelible ink pen.

2. Everything takes longer than planned. I thought the panoramic photography taken at 92 sites would be completed in about 2 hours: three hours maximum. Well, in reality Josh and Chris paused only to download massive amounts of data and grab a roll to eat, and worked solidly from 10:00 am to 4:30 pm, being among the last to leave. I reckon I need to  double the time from the maximum I estimate  to get closer to reality. We ran out of time to perform any communications tests.
3. Be careful. Sometimes more haste leads to less progress. I set the logging computer in the car to record GPS data, whilst using another laptop to gather data on a walk around. Well, guess what. The laptop in the car was set to sleep after 20 minutes when on battery power, so we only have 11 minutes and 30 seconds of time in which the logging data has been taken at the same time as the data on the walk around. A little more thought and preparation would hopefully have led to me setting up the power saving conditions on the laptop more appropriately. Bummer!
4. The Army sniper was most interesting and professional. He reckoned that the only way he personally would be able to  detect another sniper is through movement. This was our early strategy, but we started to have doubts about our strategy when informed by other Army personnel that snipers do not move. Thanks to Andy Wallace, the Project Leader for the MoD Grand Challenge for arranging the presence of the sniper, and thanks to the sniper for his engaging and informative discussions.
5. Rehearse the plan. We brought along a precision Leica GeoSystems Disto A5  laser-based distance measuring unit, only to find the laser dot it projects to the distant location could not  easily be seen in broad daylight. This left us with trying to peer down the sighting system, leading us to take more time than we had budgeted. We had to drastically curtail our use of this instrument, and omit useful information such as the dimensions of the various unique buildings.

Another reason to rehearse the plan is that it is quite stressful getting everything done by a Team in a limited time, and we are not even in a military conflict situation!

6. Don't believe the weather forecast. The forecast was for rain and gusty winds (" military weather" , in Andy's words). In fact, as can be seen from the photographs, we had lovely sunny weather with some winds, thankfully.
7. Work together where you can. There is a huge amount of work to cover. We plan to  trade some of the information we have gathered for information, such as WLAN coverage, that we simply did not have time to cover.
8. Leave the car engine running if you are using a mains inverter. We used a mains inverter to power our laptop logging GPS information in the car, and did not run the engine. With only one car left in the car park, we prepared to move out, only to find nothing happening since we now had a flat battery. Thanks to Stephen Crampton for having jumper leads and providing us the power to start our car. As folks jokingly commented, we would not have been short of a house to stay in that night: shame about having to sleep on the floor. I would not like to think of the reception we might have faced the following morning from  military personnel with this display of ineptitude on our part.
9. Google Earth can provide surprising high resolution images, almost to high altitude aircraft imaging quality. Unfortunately, many of the images are out of date so you need to be careful when using these maps.
10. We should have brought along more people so I could have spent more time with the sniper, spent more time looking inside the houses and buildings and spent more time seeing what the other Teams were up to. Unfortunately, two of our Team members experienced last minute problems: one had an urgent business meeting that he needed to attend, and the other  was caught up in a massive traffic jam. Having two  reserve Team members who could have been called on at short notice would have helped us with the communications work.

Anyway, it was a tremendously useful day. We collected a load more information that can be applied to the detection of military threats in any village. We will be posting more results from our analyses to benefit those developing Unmanned Vehicles for use in the detection of threats to military personnel and those involved in the monitoring of installations of all descriptions,  throughout the world.

We are using the MoD Copehill Down village as a generic village in which to exercise our threat detection strategies.

From left to right, Lee Smiley from DESG, Joseph Barnard from Barnard Microsystems Limited, Andy Wallace, the Project Manager for the MoD Grand Challenge 2008 and Rob Frampton, from DSTL and a  judge at the MoD Grand Challenge 2008 at All Bar One in Crouch End, North London. Prior to this, we discussed the Barnard Microsystems tactics and  progress to date, and Andy updated us all on what he had learned after a visit to the DARPA staff who have been managing the DARPA Grand Challenges.

return to top

return to top

© Barnard Microsystems Limited 2006 - 2008