“No research without action, no action without research” - Kurt Lewin 

The above-mentioned words rightly portray the value research entails. Working on AI problems define the technology of tomorrow, impacting everyone, hence, calls for multiple research work across many domains. Indian academicians are no exception, they challenge conventions and work towards reimagining technology for the larger good.

India has multiple problems at hand, and smart solutions hold the key. To that end, researchers across India publish hundreds of research papers, conduct fundamental research, and make datasets and tools available to the broader research community for a collaborative approach. To recognise their work, we have listed down the top five AI research projects by Indian academicians that can help solve some of the most pressing issues we’re facing today.  

1. Synchronized Multi Scale and Multi Sensor Traffic Data from Indian Urban Roads 

Researchers: Bhargava Rama Chilukuri, an Assistant Professor in the Department of Civil Engineering, Dr Gitakrishnan Ramadurai, an Associate Professor and part of the transportation systems group, both from IIT-Madras. 

About the project: Congestion and pollution, the two biggest dilemmas for the Indian urban world require an alternative transportation model. The availability of a high-quality traffic dataset is critical for the model to work. However, the Indian datasets suffer from three shortcomings that make them unfit to train AI models: limited Spatio-temporal extents of the collected data a limited number of parameters observed and sources used, and  limited to a single resolution of data  

To overcome these challenges, the research is oriented towards creating a precise traffic dataset for a 25-kilometre urban road system with data from multiple sources and measuring several parameters at both the macroscopic and microscopic levels for the first time in Indian settings. The data collected via sensors including GPS sensors, LIDARs, video cameras, Wi-Fi sensors and even drones will be accessed using RBC-DSAI’s in-house ML and deep-learning techniques. 

2. Indian Urban Data Exchange (IUDX) Project 

Researchers: The project is carried out by the Robert Bosch Centre for Cyber-Physical Systems (RBCCPS), an academic and research centre at the Indian Institute of Science, Bangalore. It is an initiative of the Smart Cities Mission under the Ministry of Housing and Urban Affairs (MoHUA) and the Ministry of Electronics and Information Technology (MeitY). 

About the project: With a need to enable data exchange between government agencies, city departments, citizens and the private sector – IUDX was created to utilise massively available data intelligently towards solving complex urban challenges. The open-source software platform clears the way for authenticated and secure data exchange. IUDX research project was initiated in 2019, with an initial reference code. From managing a green corridor for emergency vehicles to solid waste pickup and health management information systems – IUDX presents a greater opportunity for data-driven innovations. 

3. Deep learning assisted multi-omics integration for survival and drug-response prediction in breast cancer 

Researchers: Vidhi Malik from Brigham and Women’s Hospital, and other two researchers namely Yogesh Kalakoti and Durai Sundar from IIT-Delhi have contributed towards this project in healthcare. 

About the project: Around 1.8 lakh breast cancer cases were reported in India in 2020, as per the International Agency for Research on Cancer, thereby raising an alarm. In cancer research, the ability to predict the likely development of the disease – survival and drug response are two vital parameters. Here, the team proposed a late multi-omics integrated framework for robustly quantifying breast cancer patient survival and drug response, with an emphasis on the relative prognostic potential of available omics data types. With an accuracy of 94%, the proposed strategy provides an effective way of extracting crucial information from diverse omics data types enabling estimation of prognostic indicators. Such integrative models with high predictive power have a significant utility in precision oncology. 

4. Object Detection Project by Team Autonomous Ground Vehicle (AGV)  

Researchers: A multi-disciplinary research group from IIT Kharagpur with an aim to build a fully operational self-driving car is working on varied projects involving deep learning, computer vision and reinforcement learning. 

About the project: Object detection is one of the primary requisites for autonomous vehicles. Hence, the team developed vision algorithms for tasks such as dynamic obstacle tracking, pedestrian detection, lane detection, etc. The project will help cope with the various requirements of an AV in a diverse environment. Moreover, the reliability of vision algorithms is absolutely critical for real-time decision making, hence ensuring safety. The team worked to solve two important problems including, detection of objects in an image, and identification of an object using convolutional neural networks (CNN) to a large extent. 

5. Aurora – Intelligent UAV Design 

Researchers: A team of BTech, MTech and PhD students at IIIT Delhi with a background in CSE, and experience in fields of aeromodelling, robotics, autonomous systems, computer vision, etc. The team is being provided guidance by faculty. 

About the project: The goal is to create an intelligent UAV with capabilities such as autonomous flying, navigation, and landing, as well as the ability to perform complex tasks such as automatic airdrop, detecting and localising targets of interest, reaching a specified position or target, and so on, with manual remote controls. UAVs having these qualities can be used for a variety of tasks. The project assembles off-the-shelf airframes with the required components to create a basic functional UAV. This is then combined with an autopilot system, which adds sensing and autonomous navigation capabilities.