Finding flatland


Away from the engineering hothouse, Bloodhound Supersonic Car itself, data gathering and analysis have been equally important in selecting a flat site for the speed attempt. A novel progressive elimination method was used. Endorheic salt pans provide the population from which to sample potential sites, and candidates were initially identified using digital elevation models constructed from Shuttle Radar Topography Mission data. At this stage, spatial resolution is about 90 metres and allows elimination of many candidates.


Te simulator runs a virtual mathematical


Example of sampling buffers for one iConnect case study site. From Ogilvie et al[7]


component has its own data capture system, but these also have to be able to read and respond to each other’s information flows. A Matlab and Simulink environment


contains design, modelling and implementation, allowing numerous design iterations to be modelled and simulated quickly. Real-time code is automatically generated using model-based design, and xPC Target soſtware performs control and data acquisition from within Simulink. ‘Te graphical nature of modelling,’ comments Coorous Mohtadi, ‘allows a multi-disciplinary team approach to design and review that would be impossible with traditional document-based systems.’ Wheels nearly one metre in diameter help


to minimise the effect of surface irregularities but, for a car travelling at these velocities, the flattest possible site is needed. Here, again, innovative new data-driven approaches have been used (see box: Finding flatland). Looking more widely, transport design


depends heavily on feedback from test data into the design process. Tis is increasingly so as vehicle engineering becomes more complex and multidisciplinary, but also becomes more difficult as new methods compress timescales. All too oſten, insights come too late to be incorporated as fully as the designers would like. Maplesoſt, a soſtware publisher with well-known relationships across the transport manufacturing sector, is one of the contributors behind VI-grade’s VI-DriveSim Dynamic simulator, which aims to provide test drivers with an intermediate stage between virtual prototyping and physical testing.


10 SCIENTIFIC COMPUTING WORLD


model of the vehicle and track linked to a ‘six degrees of freedom’ Ansible Motion platform, with Maplesoſt’s MapleSim providing the inverse relationships acquisition which underpins realistic response. Driver actions are the inputs and the platform provides real-time inertial response from the vehicle model, allowing natural interaction between human driver and virtual vehicle. Te resulting experience provides data that can be analysed and returned to the design cycle for refinement and validation. Most of us, of course,


Land cover data at kilometre resolution further winnow the selection by revealing those with unsuitable temporal flood patterns. Likely locations were then analysed by an algorithm that computed elevation variance (using data from LandSat and other sources) at more than one kilometre pixels. Any consecutive series of 20 pixels (reflecting the necessary length of the speed record attempt site) with low variance emerged as sites justifying detailed investigation through hands- on human data gathering.


of pedestrians with disability scooters, and almost anything else you can imagine. One study[1]


(which I haven’t space to


are transported at speeds somewhat lower than Bloodhound’s 480m/s, or even a racing car’s 85m/s. On the other hand, we also move around at spatial densities considerably greater than Bloodhound’s one passenger per hundred square kilometres. A great deal of data analysis effort has to do with planning, improving and optimising mass transport infrastructures against one set of criteria or another: safety; efficiency; economy; productivity; volume of throughput; and quality of outcomes. Tere are statisticians out there studying the problems of motorway traffic, rail scheduling, the mixing


TOO OFTEN, INSIGHTS COME TOO LATE TO BE INCORPORATED


do justice here), for example, analyses the comparative sleep patterns of heavy goods vehicle drivers at home and on the road to develop research on safe driving legislation. At sea, Formal Safety Assessment (FSA) of evaluation of maritime regulations, for instance, is highly dependent upon analysis of accident data, which makes it sensitive to incompleteness in the databases used. Examination of historical data shows[2]


that, even


in highly developed jurisdictions such as Britain


and Norway, true figures suffer dramatically from under reporting with upper bounds in those two being <30 per cent and <41 per cent respectively. Right down at the opposite extreme of the


automotive scale from Bloodhound, one of my students, inspired by a combination of Te Straight Story[3]


and DARPA’s quadrupedal LS3


robot packhorses to examine low-powered transport options, is conducting a diligent data analytic investigation into the success/failure probability components of riding modified


Bloodhounds old and new www.scientific-computing.com


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32