Few of the research projects that I have been involved are summarized as follows (to be updated): 

I am currently working on applications of computational topology and statistical inference to develop hybrid models for unknown environments using the data gathered from biobotic unlocalized insects.

Imagine a scenario in which an earthquake has left several individuals trapped under the ruins of collapsed buildings. Sending teams of human rescuers to search and extract survivors may put the rescuers and the survivors in danger. Instead, we can choose to send teams of robots to accomplish this search and rescue operation. Since large robots may not be able to explore all desired locations due to size and safety limitations, a team of small autonomous agents can be considered.  The first requirement for such systemwould be to explore and built a map of the environment and then localize anyone in need of assistance. The mapping in such situations becomes extremely challenging due to hardware limitations, unstructured nature of the environment, and lack of GPS underground or at indoor locations. Present day technology falls short in offering robotic agents that function effectively under such complex conditions. Insects, on the other hand, exhibit an unmatched ability to navigate through complex environments and overcome perturbations by successfully maintaining control and stability. Swarms of biobotic insects, consisting of cockroaches with small electronic backpacks attached to them for sensing and control, can be employed for mapping and exploration of unstructured environments in emergency response situations with gaining access to locations that may not be reachable otherwise. Since application of traditional mapping techniques fail due to existing restrictions, we propose a robust approach to obtain a topological map of the unknown environment using the coordinate free interactions among a group of biobotic insects, which act as search and rescue agents.

In our study, we are developing fundamental building blocks of environment mapping with a cyber-physical sensor network consisting of team of Madagascar hissing cockroach “biobots” equipped with electronic backpacks. 

- Integration of input controllers into virtual reality applications, Samsung Research America

Virtual reality engages users in environments similar to the events and objects of the real world. Users interplay in a virtual environment with a sense of immersion and presence. One of the key components of user presence is their interaction with the objects in the virtual world that can be realized with properly designed hardware and software input.  Human interaction with machines, which is considered as an essential part of human machine interaction (HMI), basically can be realized through hand tracking and hand gestures.

In this project, we developed algorithms and software in order to bring human hands into the play as the most natural and intuitive way of the interaction with the virtual world by the means of wearable motion sensors and controllers. 

- Coverage control for multi agent systems with nonlinear constraints (AUT)
This project deals with the problem of covering an unknown environment using a group of mobile robots with nonlinear constraints (e.g. car like robots). We designed algorithms for mobile robotic sensors to relocate themselves to find the best deployment under various kinds of situations in order to cover the largest area. We took into account physical and motion constraints of mobile robots in the coverage control algorithm. This includes constraints on their kinematic and dynamic properties as well as their physical shape, all of which had not been addressed before. We designed controllers that make the mobile sensors converge to optimal configurations for coverage while they are mathematically guaranteed to operate in a stable manner. Furthermore, we developed a distributed adaptive approach for mobile sensors to be able to optimally cover an area with no a priori sensory information (e.g. the intensity of light). The robots learn and estimate the distribution of sensory information over the whole environment as they relocate themselves for optimal coverage. The estimation in our framework is based on distributed estimation from current measurements fed into an artificial neural network estimator.     

- Evolutionary optimization for wireless networks (AUT)