This article details how satellite remote sensing, geographic information systems (GIS) and surveying methods are being used on the Bloodhound Supersonic Car project to ensure that a suitable site is used for the World Land Speed Record challenge.
The Bloodhound SSC (SuperSonic Car) is a rocket and jet powered car that will challenge the laws of physics in order to achieve 1610 km/h, (1,6 km in just 3,6 seconds) in an attempt to break the current World Land Speed Record (WLSR) which stands at 1228 km/h held by Wing Commander Andy Green in 1997 who will also be driving Bloodhound SSC.
Fig. 1: The Bloodhound Supersonic Car.
The Bloodhound SSC project is an iconic engineering and education adventure that is pushing technology to its limit, providing us with a once in a lifetime opportunity to inspire the next generation of scientists and engineers, which is the main aim of the project.
In 2009, working together with Prof. Adrian Luckman of Swansea University; Wing Commander Andy Green identified the best land speed track in the world to be the Hakskeen Pan in the Northern Cape Province. This was done using the latest satellite and Landsat imagery from the Space Shuttle together with basic weather data. Having identified the best track in the world for Bloodhound SSC, a partnership was established between the Northern Cape Provincial Government and the Bloodhound SSC Project Team which resulted in the employment of over 300 people from nearby villages who helped scrape every square metre of the surface, remove all debris however small and ensure that the 20 km by 1,1 km (22-million m2) track is free of rocks and small stones totalling 20 000 tonnes.
Fig. 2: A team of around 300 local villagers have been employed to scrape the pan.
What does geomatics have to do with it?
Not to mention the obvious, it only makes sense to think that running a supersonic car at 1610 km/h you need a flat, smooth and hard surface that provides lateral grip for the Bloodhound V-profile forged aluminium wheels. The selection of such a surface requires the combined knowledge of satellite remote sensing, geographic information systems (GIS) and surveying methods.
Identifying a suitable surface for Bloodhound SSC
Fig. 3: Clearing small stones at the pan.
Data sources
The primary criteria of size and surface smoothness demand data which can describe such topographic and geographic characteristics for the whole land surface of the globe, apart from clearly inhospitable regions such as the Arctic and Antarctic. The source of near-global topographic data is the Shuttle Radar Topography Mission (SRTM). SRTM employed a technique called satellite radar interferometry, using data acquired by the Space Shuttle in February 2000, to build a digital elevation model (DEM) of the land surface between the latitudes of 60° North and 56° South. This data has been made freely available at a spatial resolution of approximately 90 m and forms one of the main sources of data for this study.
Fig. 4: Selected areas for topographic analysis.
Searching for a site
This was done in three phases in order to apply the geospatial analysis methods discussed later, with the following hierarchical requirement approach of finding suitable sites:
Flat ground with centimetre level surface smoothness.
A large area with minimum dimensions of 19 m x 5 m.
Reliable surface dry-out period during the rainy season.
Easy access from major routes, political and non-political security.
Local government friendliness and the potential for publicity and constructive competition.
Phase 1
Prof. Luckman wrote an automated computer programme that enabled the Bloodhound SSC team to search the surface of the earth for possible sites to run the supersonic car. Details on the site search process are outlined here:
Fig. 5: Blue Marble composite satellite data to illustrate the Australia map area in the UTM map projection.
Select areas for geographic data analysis
This involved a global search for perfectly flat areas, with no vegetation and in excess of ≈ 20 km long. For the purpose of the Bloodhound SSC project, using the Universal Mercator Projection (UTM) map projection, (which is not suitable for large areas; because of an increase in error and distortion when moving farther from the zone from which the projection is defined), the relevant part of the global land surface was divided into seven map sheets (Fig. 4) covering Africa, Australia, Central Asia, East Asia, Europe, North America and South America. This led to the development of a GIS database and processing tools for the project.
The aim of the automated part of the site selection process was to locate large flat areas of the global land surface using an analysis of topographic elevation data represented by the STRM DEM and to refine the list of possible targets by rejecting those which global satellite land cover analysis showed to be covered in vegetation.
Fig. 6: Shade relief representation of the complete digital elevation model divided into 20 map tiles constructed at 100 m spatial resolution.
Topographic and land cover data analysis for selected areas
Within each tile of each geographic map area, the variance was measured in 10 km by 10 km DEM patches. This represents the approximate scale of the required smooth surface on which the Bloodhound SSC is to run. The algorithm essentially took all the DEM height values within the 10 km square patches (100 by 100 DEM samples) and calculated the variance in topographic height. The patch variance was sampled on a grid of 1 km; that is the centres of the patches were spaced at 1 km intervals to produce a map of topographic variance with a nominal spatial resolution of 1 km (recognising that neighbouring samples are not truly independent of one another). Thus, in the output maps, areas of 20 or more adjoining 1 km pixels with low variance indicate regions both large enough (20 km ≈ 12 miles) and flat enough to be investigated as potential lake bed sites. To further refine the selection process, the two sources of land cover data were also re-projected into the same map projection at the output spatial resolution of 1 km. Fig. 6 shows one of the datasets used.
Areas with land cover classes of “bare”, “short vegetation” (to allow for the possibility of mis-classifications) and “water” (to pick up seasonally-flooded lakes) were included and all other classes were excluded as being unsuitable. The result of the combined topographic variance and land cover analysis was a series of maps for the seven geographical regions presented at a spatial pixel size of 1 km by 1 km and highlighting large (> 10 km) areas of low topographic variance (< 10) with minimal vegetation cover.
Fig. 7: Land Cover map for Australlia reprojected from data from the Global Land Cover Facility. variance with suitable land cover classes.
Refinement using Google Earth and other satellite data
The first stage in the refinement process was to use Google Earth to inspect each point at which 20 or lower variance pixels of 1 km (~12 miles in all) were bunched together. Google Earth was chosen because it provided ready access to aerial imagery of sufficient spatial resolution to identify the extent of lake beds, and to allow obstacles such as roads to be spotted. It was discovered that Google Maps occasionally included different aerial imagery to Google Earth so this was used in conjunction to gain a second view of each target on a separate date, and therefore at a potentially different part of the seasonal cycle. From the refinement process, a list of 35 possible Bloodhound test sites was produced, nine of which were previous sites for land speed record attempts. These sites are illustrated in Fig. 9. Further prioritisation resulted in a list of 13 sites requiring further investigation.
Fig. 8: Potential Bloodhound driving environments of Australia highlighted by combining low topographic. variance with suitable land cover classes.
Phases 2 & 3
The Phase 1 global search produced a list of 35 possible sites for Bloodhound to run on. Nine of these are “known” land speed record sites that do not require Phase 2 study. Phase 2 therefore concentrated on the 13 priority “unknown” sites from Phase 1 which might prove to be suitable. This involved collecting additional images, together with an indication of regional weather. The multiple images were intended to show the variability in conditions, and include Google Earth images, Google Maps images (unless identical to Google Earth) and Landsat images (from the year 2000, date unspecified). These Landsat images have been made into maps with scale and geographic information.
For the last phase of the desert selection, additional data on the surface conditions of the 13 priority sites was collected. Other factors of access, security, geographic location, local infrastructure, etc., were also considered to reduce this to a list of sites for visits and ultimately Hakskeen Pan emerged as the best site to run the Bloodhound SSC.
Fig. 9: Locations of 35 test sites identified by the topographic variance technique as worthy of further consideration.
Surveying the pan
At this stage, the number one question about the run site for the Bloodhound SSC is how flat it is or perhaps more accurately how bumpy it is. The team needed to know if there would be any nasty surprises when running the Bloodhound SSC at 1610 km/h.
A Northern Cape team surveyed a 2 km stretch of desert, adjacent to the chosen run track, to give an indication of vertical variation over this length. At 1610 km/h, the car will travel 1,8 km (≈ 2 km) in 4 seconds hence the 2 km survey task. Measure the vertical height (as accurately as possible) at 200 points, at 10 m intervals in a 2 km long straight line, on a representative part of the desert which resulted in a total vertical variation of 61 mm over the 2 km proving the desert to be flat enough for the run.
Fig.10: The Hakskeen Pan in the Northern Cape Province.
Progress
Communication tests
Emcom Radio Communications has deployed and tested its DMR Tier 3 radio network operating on TDMA technology for voice communications at the Hakskeen Pan. This network has voice and data capability. It will primarily be used for communication between the driver of the car and the control room as well as by the operational staff handling logistics to do with the safety and control at the pan (Police/Fire/Ambulance/Helicopter/Track marshalls etc.). Furthermore the 300 data channels uploaded from the car will be covering all important information such as the speed of the Bloodhound SSC, fuel consumption, all three engine temperatures and performance, forces on the chassis, intake and wheels, systems condition, plus of course the driver’s (Andy Green) heartbeat. All this information will be pushed via a satellite link into the cloud for public access and live tracking of the event from across the world from the Bloodhound website for the run.
Fig. 11: Surveying the pan.
MTN has deployed an LTE (4G) communications network for voice by the general public in the area but primarily for data and high definition camera live streaming from the vehicle using an antenna system specially developed by Poynting Antennas based in Johannesburg. Images taken during the communication tests that were carried out in November 2014 at the pan can be seen in Fig. 13.
Fitted with the same radio equipment as the jet and rocket powered Bloodhound SSC, the F-Type R All Wheel Drive (AWD) was driven head-to-head at top speed with a similarly equipped L39 Jet flown at 500 knots and just 15 m above the ground. The combined closing speed of almost 800 km/h enabled the successful test of the system that will allow communications between the ground crew and driver Andy Green in Bloodhound SSC, which will run for the first time on the Hakskeen Pan desert in 2015. Bloodhound SSC will run low speed on a runway in the UK in August 2015 and then head for South Africa in a heavy-lift aircraft and commence running in September. The target speed for 2015 is 1300 km/h and the car will return in 2016 when the target speed is 1610 km/h (1000 mph).
Fig. 12: Vertical variation over the 2 km surveyed length.
The future
As previously mentioned, the number one aim of the Bloodhound SSC project is to inspire the next generation of engineers and scientists by sharing all the research, design, manufacturing and testing of Bloodhound SSC while setting a new World Land Speed Record (WLSR) of 1610 km/h. This is done through preliminary test runs at the Hakskeen Pan and a network of Ambassadors throughout South Africa who visit schools to deliver exciting presentations on the Bloodhound SSC Adventure.
Fig. 13: The Jaguar F-Type R Coupé and Jaguar XF sedan carrying communications equipment needed to stream voice, data, and video from the Bloodhound SSC at the Hakskeen Pan in November 2014. Above them an L39 aircraft flies past over the track, reaching closing speeds of about 800 km/h.
References
[1] Tony Sipho Sibanda: Director: Business Development, Emcom Wireless
[2] Dave Rowley: Bloodhound SSC Education Director – South Africa
Contact Teboho Maphakisa, Department Rural Development and Land Reform,
Tel 021 658-4460, teboho.maphakisa@drdlr.gov.za
The post Geomatics and its role on the Bloodhound Supersonic Car project appeared first on EE Publishers.