Suraj Kumar

Ongoing Projects

Design and Control Co-Optimization for Quadruped Locomotion in Low Gravity

This work explores unified optimization of quadruped robot mechanical design parameters and control laws to achieve robust locomotion in low-gravity environments relevant to ISRO's upcoming south polar lunar missions.

Physics-Informed Machine Learning for Real-Time Hamilton–Jacobi–Bellman Optimal Control for Aerospace Applications

Learning based Adaptive Powered Descent Guidance Algorithms for Planetary Landing

In this work, we employ a learning-augmented guidance framework in which the functional form of classical model-based guidance laws are preserved, while data-driven learning layers adapt gains and problem parameters that are difficult to obtain analytically.

Real Time Planning and Control of Quadrupeds For Unstructured Environment with Safety Guarantees

This research is focused on developing real-time planning and control framework for quadruped robot to ensure safe exploration of uneven terrains of moon and mars.

Completed Projects

Descriptive Model based Learning and Control for Bipedal Locomotion

This research focused on developing descriptive model reduction based approach for control and learning architecture for bipedal locomotion and established that learning in reduced state and action space is sufficient to realize robust locomotion.

Chandrayaan-3: Guidance, Navigation and Control

This research focused on developing the end to end guidance, control and navigation algorithms and flight implementation for Chandrayaan-3, India's 3rd lunar mission with objective of soft landing and surface mobility.

Closed loop Planning and Control of Low Thrust Orbit Transfers Using Physics Informed Evolutionary Machine Learning

This research focuses on the development of closed-loop planning and control laws for low-thrust orbital transfers by integrating classical Lyapunov-based control methods with evolutionary machine learning techniques. The proposed framework leverages physics-informed learning to adaptively optimize control policies while preserving formal guarantees on stability and convergence.

Pseudospectral optimal control for trajectory optimization in Aerospace Systems

This research focused on developing hp adaptive pseudo-spectral collocation methods solving optimal control problem and utilize it for trajectory optimization for Entry, descent, landing, ascent and rendezvous applications

Template Model based Model Predictive Controllers for Bipedal Locomotion

This research focuses on the design of hierarchical controllers with receding horizon model predictive control for balance and task space control for trajectory tracking

Mobile Rover Path Planning with Non-holonomic Constraints

This research focuses on the development path planning approaches to mobile rover that incorporates vehicle properties and non-holonomic dynamics constraints in the planning layer.

System Identification and Learning based Dynamics Modelling

This research focuses on the development Learning based models for simulation and control

Page updated

Google Sites

Report abuse