You can read my paper on Modeling of Dust Cleaning Robot and work on it further as a control research problem. We can discuss and publish together incase you are interested. Please email me.

Dear Venkata Mahesh,

naturally, scientests thinks that they do the most important research. I am not exception and recommend you the direction dubbed Fully probabilistic control

of Fully probabilistic design of decision strategies (and an independently developed special case called KL controll).

Best regards M. Karny

See

https://www.utia.cas.cz/biblio?author=cav_un_auth%2A0101124

especially works:

Kárný Miroslav, Böhm Josef, Guy Tatiana Valentine, Jirsa Ladislav, Nagy Ivan, Nedoma Petr, Tesař Ludvík :  Optimized Bayesian Dynamic Advising: Theory and Algorithms, Springer, (London 2006) [2006]

Kárný Miroslav, Kroupa Tomáš :  Axiomatisation of fully probabilistic design , Information Sciences vol.186, 1 (2012), p. 105-113 [2012]

Kárný Miroslav, Böhm Josef, Guy Tatiana Valentine, Nedoma Petr :  Mixture-based adaptive probabilistic control , International Journal of Adaptive Control and Signal Processing vol.17, 2 (2003), p. 119-132 [2003]

Kárný Miroslav :  Towards fully probabilistic control design , Automatica vol.32, 12 (1996), p. 1719-1722 [1996]

it all depends on what you want to do! Trending topics cover a large range - starting from Robotics and its applications as one of the respondents indicated, and various control applications - where you try to get a better handle on the system through the use of control theory, to developing control theory - and establishing new results - new control method, property of a control metho (robustness etc) to Fault Tolerant Control and reconfiguration. One area, as Prof Karny, has indicated in Probabilistic control and its applications for FTC. Indeed elements of probablistic control ideas have been used very efficiently in a wide range of applications and domains. The choice is yours to make