We often hear about the rankings and ratings of institutes which are considered the best in terms of research. Unsurprisingly, they’re often located in developed countries. Yet, a more nuanced view of the global research landscape reveals that many institutes in developing countries have competitive research strengths in as well. The reason these don’t come up in global research indexes is that usually such institutes excel in niche areas. The methodology for research ranking commonly looks at the comprehensive research strengths of institutes and as a result overlooks these small pockets of competitive research.

For example, a leading university in the developed world will usually have an integrated Computer Science and Electrical Engineering which will be producing a lot of research in this broad discipline. The approach to rank the quality of universities’ research (done through bibliometric indexes of quality and volume of research) will consider such integrated departments as a term of comparison. This method will not take into consideration a university in the developing world which may not have such an integrated department or comprehensive strengths overall as an institution, but might well be doing excellent research in, say, the sub-field of Artificial Intelligence.

To create recognition about niche research strengths, ITU’s Scientometrics Lab, under the patronage of Dr. Saeed ul Hassan, is building a data analysing tool. This is a pioneering project launched to address the fact that research quality measurement models for the developing world need to be differentiated.

The idea behind this project is that the identification of such research strengths will help institutes and universities in the developing world to strengthen their programs, focus on their strengths, improve their research profiles and be in a better position to apply for funding and research grants. The project is at an initial stage at the moment but it will eventually allow universities to compare their departmental and niche research strengths through various indicators and benchmark their research strengths with better analytical tools. The scope for this project will be global.

Technological advances in communications technology over the past decade or so have rendered devices with sensing and computing ability, such as cell phones, affordable and widely available. This universality presents an interesting phenomenon known as Opportunistic Networks; these are networks which are formed when devices come into contact opportunistically through physical proximity, and communicate wirelessly to share each other’s content, resources and services. Opportunistic networks very closely resemble social networks; they are primarily human centric for they come into being when humans come into contact. As Marco Conti and Mohan Kumar put it, “opportunistic networks are tightly coupled with social networks and can exploit human relationships to build more efficient and trustworthy protocols.” Whereas human mobility in ad-hoc networks is seen as problem, opportunistic networks in fact see at is a productive opportunity which can be exploited to forward data.

With the pervasiveness of devices with built-in wireless and wired technologies, the infrastructure for opportunistic networks is in place and has great potential. Some of the most useful applications of opportunistic networks include establishing communication in situations wherein regular networks are disrupted; such as in the case of natural disasters, emergencies or in far-flung areas. However, there are no guarantees that data sent through opportunistic networks will reach the intended recipient, and a there may also be a great deal of delay. Farah Amjad, a Research Assistant at ITU’s TRENDS (Technology and Research in Emerging Networks and Distributed Systems) lab is currently completing a paper in which she is identifying the main challenges which bar opportunistic networks from being more successful.

Among the challenges Ms. Amjad has identified in her paper is the problem of routing; i.e. in opportunistic networks, routing algorithms have to provide reliability even if connectivity among nodes is intermittent or temporarily unavailable. Another problem is that of increasing the efficiency of devices- as the devices are small they don’t have very long lasting batteries.  Similarly another issue is the due to high experimental cost most of mobility models are simulated and designed for small scale. To overcome this there is need to simulate mobility models on a large scale while considering real life mobility patterns. Additionally, congestion is also a problem in opportunistic networks as the overflow of data exceeds the storage and sending capacities of devices. Ms. Amjad has also highlighted the key challenges of security, privacy and trust in opportunistic networks as well as the fact that the nodes in opportunistic networks are often small devices functioning on low bandwidth and thus network protocols need to be designed accordingly.

The future of opportunistic networks is extremely promising. Ms. Amjad’s paper will make a significant contribution to the field by highlighting the need for networking solutions focussed on opportunistic networks. This paper will soon be sent for publication, and is being written under the supervision of ITU faculty member and TRENDS Lab patron Dr. Adnan Noor Mian.

Using technology to make contemporary living easier and hassle-free is a given, but what about using technology to prevent crime? Usman Bukhary, an ITU MS student and a Research Assistant at the Business Analytics Lab (BAL), is currently using data-mining and machine-learning techniques to analyse data gathered on crimes taking place in the district of Lahore.

GIS software and data-mining technologies can help immensely in organizing data to build systems which can prove useful in improving efficiency and ensuring better resource utilization. In this case, Bukhary is using data on crime to provide actionable feedback to local law enforcement agencies to help them improve crime-prevention measures.

Data collected on crime from every thana (police station) in the seven divisions of the Lahore district has been mapped and analysed in both spatial and temporal dimensions. It has been organized across different variables including the location, time and date at which crimes have taken place, and the type of crime.

The results, including thematic maps, have yielded useful information; for example, the cluster map below shows that crime takes place mostly on roads and public places in Lahore.

Such information can be used to follow crime patterns, which can be extremely helpful in the efficient deployment of personnel by law enforcement agencies. The map below shows areas where motorcycle thefts usually take place in the city. Temporal data reveals that the probability of motorcycle theft in these areas is highest on Tuesdays. Police could easily use such information to ensure that appropriate and specialized preventive measures are taken to counter the type of crimes an area is most prone to at any given time.

However, the data Bukhary is working with at the moment comes exclusively from the police stationsof Lahore and accounts only for reported crimes. This data very likely does not represent all the crimes which take place in the district. To make the system more comprehensive, Bukhary plans on analysing information from social media platforms such as Twitter.  For example, if someone is robbed, they may not report it but if they publicly tweet about it, the information could be recorded in the database.

This project is being supervised by ITU faculty member and BAL patron Dr. Faisal Kamiran.

Three postgraduate students at ITU have initiated research in Sensor Networks in Technology and Research in Emerging Networks (TRENDS) lab as part of their MS and PhD theses requirements. Under the supervision of ITU faculty member Dr. Adnan Noor Mian, two MS students; Mohammed Umer Zahid and Ali Imran, are working on applied research on sensors while PhD student Raees Khan is involved in developing protocols.

Umer Zahid’s research is geared towards developing a Smart Parking system wherein a mechanism could be developed to place sensors in parking lots which will be able to detect which lots have been filled and thereby allow for an efficient guiding system to be installed in parking lots. The most significant hurdle for Zahid in this project is to evolve a cost-effective system for this project using cheap, off-the-shelf sensors and microcontroller board such as Arduino board. It is also necessary that this system be energy-efficient and so Zahid is working to minimize the energy consumption of the system and looking to solve such problems as how to prolong the battery life of sensors so that they could run for years.

Ali Imran is working to develop the coexistence of Wireless LAN and sensor network protocols. Wireless LAN technology works in the 2.4 GHz ISM band, which requires no license for usage. Most wireless technologies such as Bluetooth networks, wireless mouse, wireless LANs, etc. operates in this band. However, due to interference in the band, data loss often takes place. Imran’s research is to focus on ensuring that these different sensor network protocols work in harmony within the band; with minimal interference and data loss. This research is being done in collaboration with the software house Qairus.

Raees Khan is doing his PhD in developing transport layer protocols. Reliability in transport networks is centred on source to destination routing through multi-hop transmission. In end-to-end reliability systems, the destination sends back acknowledgement of the receipt of data to the source; however, in these systems, intermediate nodes are not concerned with reliability. Khan’s research is focussed on implementing protocols to develop hop-by-hop reliability wherein intermediate nodes will also be involved in ensuring data reliability.