Data analytics, big data, machine learning, artificial intelligence etc. are words dominating our research and industry discourses these days. We believe that if we had enough data, to understand our biggest problems like air pollution or climate change, we would be able to better tackle these menaces. AI for Social Good (Google) or AI for Earth (Microsoft) are example grant programs trying to develop AI algorithms to process environmental sustainability datasets. However, creating the necessary environmental datasets in developing countries is hard. Delhi-NCR, for example, covers 55K square KMs, but has only 35 air quality monitoring stations, even with the whole world’s attention focussed on this city’s pollution problems. Measurement infrastructure in other parts of the country is worse. Budget constraints, lack of domestic instrument production increasing procurement and maintenance costs from foreign countries, lack of broadband network so that deployed sensors can easily send data from the field to remote servers for processing — all play a role in this data paucity problem. In this talk, I’ll highlight how embedded systems/edge computing/IoT (same thing, different names) can augment this data generation process for sustainability problems in developing countries. I’ll use two examples of air pollution monitoring using Delhi public buses and traffic monitoring on Delhi Ring Road intersections. The talk will touch upon the low level sensing and embedded processing pipeline including embedded deep neural networks, software verification and remote attestation methods to ensure device security, privacy issues with the collected data data when different industry partners are involved, and AI/ML on the aggregated datasets for urban policy analysis like odd-even traffic rule.
Short Bio:
Rijurekha is an assistant professor in the department of computer science and engineering at the Indian Institute of Technology Delhi. Her research interests are in problems at the intersection of information technology and society, and in particular on building distributed, networked and privacy aware systems. This includes work on building systems for road traffic monitoring, human mobility measurements, public policy audit and privacy enhancement in ubiquitous systems, among others. Before moving to IIT Delhi, Rijurekha did her PhD at IIT Bombay and was a postdoc at the Max Planck Institute for Software Systems at Saarbrucken in Germany.