Absolutely! I possess a strong inclination towards research projects. Wondering why? Some of the reasons have been nicely summed up here.

I deeply believe in the value of collaborative research. Throughout my journey so far, I’ve been fortunate to work with supportive mentors and incredible collaborators who have enriched my experience and broadened my perspective. I’m truly grateful for their guidance and the opportunities they’ve provided, as none of these projects would have been possible without their contributions! I’m passionate about sharing ideas and receiving feedback, so if you have any thoughts or comments about my current projects, I would love to set up a quick call to discuss them. Engaging in meaningful conversations not only helps refine my work but also fosters a sense of community in our field. Let’s connect and explore new possibilities together!

What I'm working on right now?

I find it fascinating that starting with the building blocks of Probabilistic and Statistical Thinking, and incorporating some Linear Algebra and Calculus, we can achieve extremely powerful function approximators that emulate knowledge-based reasoning. Currently, I am focusing on laying down firm roots in these fundamentals while taking an applied approach to Artificial Intelligence, Machine Learning, and Robotics through my software development skills. Consequently, I am inclined to study various fundamental problems that enable the autonomous flying of UAVs, autonomous multi-agent systems, and the advancement of control systems. Furthermore, I enjoy building robotic systems to practically demonstrate various ideas, aiming to span the entire world with optimized solutions.

Past Work and Projects

Multi-agent Systems

In what I consider past life now 😄, I spent considerable time working on topics in ML and control in multi-agent systems. As a part of my research fellowship at Moon Labs, IISER Bhopal, I worked on a novel problem on priority based colliion avoidance in different types of UAVs, flying in a common dense airspace. Along with my team members Agamdeep Singh, and Prof. P.B. Sujit, we worked on a series of projects around this topic:

• PANDA: Priority-Based Collision Avoidance Framework for Heterogeneous UAVs Navigating in Dense Airspace:

  In a preliminary study (submitted to AAMAS 2025), we presented PANDA, a novel potential-field-based approach that addresses priority-based collision avoidance constraints in a unified way by augmenting tangential potential fields. We evaluated PANDA through simulations, and the results showed that it achieves a 54% improvement in average completion time over existing algorithms in dense airspace. Additionally, PANDA achieves a 21% faster completion time for the highest priority UAVs compared to a no-priority baseline and a 60% faster completion time for the lowest priority UAVs.

• DenseUAV: Decision-making Collision Avoidance Framework for Diverse UAVs with Dynamic Priorities, No-fly Zones, and Environmental Condition:

  In a follow-up work (submitted to RA-L 2024), I developed a decision-making collision avoidance framework for diverse UAVs with dynamic priorities, taking into account no-fly zones and environmental conditions such as a wind disturbance model. To the best of my knowledge, there has been no work so far that addresses dynamic priority in this context. Engaging deeply with this research provided me with the motivation to pursue a PhD! 🙌

Computer Vision and Robotics

I had been fortunate enough to begin my research through computer vision in robotics through my research assistantship at IIIT-Hyderabad. I wroked on a serious of projects with my teammates (Neel Adwani, Prof. Madhava Krishna, Prof. Harikumar Kandath, and Prof. Ravi Kiran Sarvadevabhatla)

• RHFSafeUAV: Real-Time Heuristic Framework for Safe Landing of UAVs in Dynamic Scenarios:

  I worked on detecting safe potential landing zones (PLZs) and identifying the best one for landing. Initially, PLZs were detected using the Canny edge algorithm and diameter-area estimation, labeling spots larger than the vehicle's clearance as safe. In the next phase, I calculated the velocities of dynamic obstacles approaching the PLZs and their time to reach these zones. I also computed the UAV's estimated time of arrival (ETA) and executed dynamic obstacle avoidance during descent. This approach showed better results in real-world testing compared to existing methods.

• UAV-based Visual Remote Sensing Automated Building Inspection:

  The initial work at IIIT Hyderabad where I proposed a methodology to automate the building inspections through UAV-based image data collection and a software library for post-processing that helps in estimating the seismic structural parameters. The key parameters considered here were the distances between adjacent buildings, building plan-shape, building plan area, objects on the rooftop and rooftop layout. This made me explore the different aspects of the computer vision techniques.