UCL researchers take their research to parliament
In total, eight researchers from across UCL were shortlisted to present their research to a panel of expert judges and MPs at the House of Commons on March 7 in this year’s SET for Britain 2016 competition.
Run by the Parliamentary and Scientific Committee, with support from various institutions, SET for Britain aims to encourage, support and promote early career research scientists, engineers, technologists and mathematicians from around the UK whilst fostering dialogue between researchers and MPs.
Stephen Metcalfe MP, Chairman of the Parliamentary and Scientific Committee, said:
“This annual competition is an important date in the parliamentary calendar because it gives MPs an opportunity to speak to a wide range of the country’s best young researchers. These early career engineers, mathematicians and scientists are the architects of our future and SET for Britain is politicians’ best opportunity to meet them and understand their work”.
<pThis year’s UCL entrants from the department included Dr Francesco Fioranelli and Dr Adnan Mehonic. Details of their research is shown below.
MULTISTATIC RADAR SYSTEMS FOR PERSONNEL RECOGNITION AND CLASSIFICATION OF HUMAN ACTIVITIES
Dr Francesco Fioranelli and Prof Hugh Griffiths.
The presented research described the use of multistatic radar systems for personnel recognition and classification of human activities based on micro-Doppler signatures. Multistatic radar have multiple transmitter and receiver nodes in different locations, separated by significant distances, as opposed to conventional monostatic system with co-located transmitter and receiver. This can provide substantial advantages, such as additional information on targets by exploiting multi-perspective views from different radar nodes, and more resilience for the overall system in case of problems at one of the nodes.
The poster showed how micro-Doppler signatures extracted from multistatic radar data can be analysed for the identification and classification of different activities performed by people (walking, running, carrying objects, crawling, and so on), as well as for the recognition of particular individuals from their walking gait.
Potential applications include access control to restricted areas at military and commercial sites to be monitored for unauthorised personnel, and discrimination of people walking with free hands vs people carrying objects, which in some contexts could be related to potentially hostile behaviour and require further investigation using other sensors or the deployment of counter-measures.
Some presented results included the successful recognition of an individual walking with accuracy up to 99% using a single feature (a numerical parameter extracted from the radar micro-Doppler signatures and used as input to automatic classifiers), as well as the discrimination between unarmed vs potentially armed personnel with accuracy up to 97% again with a single feature.
MEMRISTORS – A NEW TYPE OF COMPUTER MEMORY AND BEYOND
Dr Adnan Mehonic and Professor Anthony Kenyon.
For many decades now, modern technology has relied on digital electronics, which has delivered unparalleled and revolutionary technological advances along with increased prosperity. The huge success of CMOS industry was realised by continuous scaling of basic electronic elements – transistors. However, in many ways >we are reaching the fundamental limits of what we can achieve with this approach. The aggressive downscaling of device size that has driven the microelectronics revolution is reaching its limit and, coupled with increased power dissipation in evermore densely packed microchips, we are nearly at the end of this particular roadmap.
Memristors (also known as Resistive RAM), two - terminal nanodevices, offer a potential solution to some of the challenges. Such devices have a wide range of potential applications including high-density memories, novel processor architectures, and artificial neural networks. In 1971, a memristor was theoretically described as the “missing” fourth passive circuit element. The first practical realisation of the memristor came 40 years later, in 2008, when an HP research team demonstrated a titanium oxide memristor. The main property of a memristor is that applying an appropriate voltage can vary its electrical resistance.
In 2012, we demonstrated a silicon oxide memristor. Silicon based memristors have a great advantage of being more easily integrated with silicon CMOS processing technology. We use a simple capacitor-like metal-insulator-metal design. The resistance contrast between the states is as high as 1,000,000, the switching time is less than 60ns, and the switching energy is 1pJ/bit or lower. Scanning Tunnelling Microscopy suggests individual switching elements could be as
small as 10nm or smaller. The device can be switched between the states at least a few thousands times. These properties are already superior to those of Flash memory. In addition we have demonstrated a room temperature quantisation of conductance in our devices making it attractive for quantum information processing. Further our device can model many different aspects of the neuronal dynamics, more specifically the integration of incoming signals and the generation of spike-like responses. This makes our device very useful for neuromorphic systems.
The impact of developing a novel CMOS compatible memristor device is to contribute in solving common problems such as storing and computing in more efficient ways. Academically this work has the potential to benefit a number of different communities, including, but certainly not limited to those working on resistance switching, memristors, neuromorphic applications, unconventional computing, bio-inspired electronics and oxides in general.