Prabhat Kumar Mishra

Policing agencies increasingly face complex challenges requiring technological solutions to enhance operational efficiency and safety. Dogs are the oldest friends of humans and therefore they helped human civilization to evolve by providing emotional companionship and alarming about the possible threat or intruders. The natural instincts of dogs have also helped police personnels in surveillance and investigation. In the modern era, dogs are replaced by quadruped robots. This project aims to develop an advanced quadruped robotic system tailored for police use, capable of efficient autonomous navigation in rough terrain. The project combines cutting-edge robotics, artificial intelligence, and control systems to address critical demands in modern law enforcement, including search and rescue, surveillance, and hazardous environment assessment.

Deep learning based tools are getting tremendous popularity in decision making and planning problems but they also receive criticism from the mathematicians due to the lack of their interpretability and the automation industry professionals due to the lack of safety guarantees, which further prohibits their quick adoption in real-time applications involving safety of humans and hardware components. This project aims to advance the theory of safe learning by providing mathematical guarantees and to overcome the practical challenges through experiments. We are interested in combining optimization based control theoretic tools with deep learning based data-driven tools in such a way that the impact of uncertainties can be minimized. This project also aims to develop a laboratory equipped with autonomous ground and air vehicles, and a simulation environment for the benchmarking and validation of the theoretical outcomes of this project.