Exploiting Neglected Features To Solve Known Problems
Talk by Prof Ralph Benjamin
CB, PhD, DSc, DEng, FIEE, FCGI, FREng.
9 August 2006
Having read the preliminary write up to this talk, we waited with trepidation for Ralph to begin. But his presentation was both straight forward and concise. For more than 60 years almost all of his projects were triggered by "XYZ can't be done", given the standard assumptions and established approaches, and completed by thinking outside those assumptions.
For each of his 5 topics he first explained why there were problems. He then showed how the problems could be solved by using a different approach.
Smart Mobile Base Stations and Dumb Terminals
There are many routes between a base station and a mobile phone due to signals reflecting off buildings etc. Signals arriving at the mobile can interfere with each other, producing poor quality sound. This is countered by having complex channel-matching, diversity antennas and equalisers at both the mobile and the base station. By interleaving the transmit and receive signals at the same frequency, a training signal from the mobile unit can characterise its propagation path both to and from the base station. The burden of processing is transferred from the many sparsely used mobiles to a few intensively used pre-processors at the common base station, thus saving the complexity, bulk, weight, power consumption and cost of the mobile. Analysis shows that providing equipment for 20 simultaneous calls produces a signal to noise improvement of 100,000.
As an aside Ralph showed that the best place to site a Base Station is at a school. Using the inverse square law, the power of a transmitted signal 50m from a base station is 1,000,000 times less than the power 5 cm from a mobile phone transmitting at peak power. In order to save on battery power a mobile phone transmits at a power inversely related to the power of the signal it receives from the base station. So the nearer the mobile is to the base station then the lower the power it uses to transmit.
X-rays with CAT-scan quality, but minimal irradiation
Ralph explained that ordinary X-rays can not distinguish between absolute and relative size, depth location, thickness and density of components within a target object. He showed how a CAT-scan gets round this problem by using multiple fan beam images. Unfortunately this requires a very high dose of radiation - about 100,000 times that of a normal X-ray. It also requires expensive fast moving equipment. He showed that a moving target produced a blurred image. The patient can be asked to hold their breath for 10 seconds but not to stop their heart beat. By using a totally different way of analysing the data, Object Based 3D Imaging, Ralph showed how the radiation level could be dramatically reduced together with an enormous reduction in data processing and storage. The equipment is relatively low cost, has no moving parts and can cope with a heart beat. This method of analysing intersecting fan beam images produces a wire frame representation of the object. Computer graphics can then be used to convert the image to a more life like representation.
Finding explosives in aircraft luggage
Object Based 3D Imaging gives a much improved determination of density. It can detect a difference of about 2% between materials, sufficient to discriminate between plastic explosive and Haggis hidden in luggage. 10 stationary fan beams are arranged along the luggage security screening tunnel. The time shift of the luggage travelling between the beams is used to simulate two beams intersecting.
Seeing "invisible" buried land-mines
Land mines kill and maim indiscriminately, make vast areas inaccessible and are being laid faster than they are being cleared. Traditional methods using a plastic probe and metal detectors, are dangerous and do not work with plastic mines. Ground penetrating RADAR fails when a mine is small, plastic or flush with the surface. RASOR - Real Aperture Synthetically Organised RADAR uses an array of antennas. By using all of the transmitter to mine to receiver paths between the different antennas and allowing for the distance between the antennas on particular paths, a more accurate position of the mine can be computed. About 30 paths are required to detect a flush buried plastic mine. This technique has many applications:
Road works - buried pipes, cables, ducts
Forensics - arms, loot, buried bodies
Terrorism - through wall imaging
There is also the potential for an acoustic variant for observing heart action and a moving belt system to detect food contaminants. But the most interesting is the detection of breast cancer.
Detecting "undetectable" tumours
Breast tumours have poor X-ray contrast, 20% are missed, and more than 20% false alarms are produced. Using 10GHz RASOR tumours as small as 2mm are detectable. The technique can also see places, such as towards the arm pits, that are difficult for X-ray mammograms. Screening time is about 2 minutes, much more comfortable for the patient and once in mass production it could be cheaper.
Ralph's last slide showed how difficult it is to get innovative solutions to market. The mobile phone developments have been sold to Texas Instruments - exploitation unknown. The Object Based 3D Imaging is waiting for the Universities to get funds for a spin-out company. It is also being investigated for adaptation to MNR imaging. The RASOR land mine detection has vanished into the MOD. The moving belt variant for QA in food processing has had a successful feasibility study and QinetiQ is negotiating for a development contract with an international company. Only the RASOR breast screening is making progress with a University spin-out company about to start clinical verification trials.
Ralph started his talk with a quote from Rutherford "If you understand your subject you should be able to explain it's essence to any sympathetic listener with reasonable intelligence, common sense, and no preconceived ideas". Ralph certainly passed this test and provided us with an excellent talk.
Michael Clinch thanked Ralph for his thought provoking talk. He agreed that "it can't be done" was a good starting point for research and development and admitted that he had been unaware that CAT-scans used such high radiation levels, citing its use on the unborn child as of particular concern.