Manan Gandhi

Papers

Total Citations

H-Index

About

Manan Gandhi’s research lies at the intersection of optimal control, variational inference, and machine learning for autonomous systems. His most influential work introduces a generalized framework for Variational Inference-Stochastic Optimal Control by leveraging the non-extensive Tsallis divergence. In his 2021 paper, Gandhi reformulates the optimality likelihood function using a deformed exponential, creating a novel Tsallis Variational Inference-Model Predictive Control approach that bridges information theory and trajectory optimization. This contribution, which has garnered 17 citations, offers a more flexible and robust alternative to traditional KL-divergence-based methods, enabling better exploration and uncertainty handling in complex control tasks. Gandhi also explored trajectory optimization under partially learned dynamics, addressing the critical challenge of model fidelity in robotics. By combining pseudospectral methods with machine-learned models, his 2017 work provides a pathway for autonomous systems to operate effectively even when underlying parametric models are incomplete. Together, these contributions demonstrate Gandhi’s commitment to advancing the theoretical foundations of model predictive control while tackling real-world autonomy challenges. His work is particularly valuable for researchers developing learning-based controllers that must balance exploration, safety, and computational efficiency.