Ashish Joglekar

Papers

Total Citations

H-Index

About

Ashish Joglekar is a robotics researcher specializing in legged locomotion and reinforcement learning, with a particular focus on developing practical, deployable control strategies for quadrupedal robots. His work sits at the intersection of deep learning and robotic motion planning, addressing one of robotics' most challenging problems: enabling four-legged machines to walk robustly across varied terrains. Joglekar's most notable contributions center on the quadruped robot Stoch 2, for which he has pioneered multiple learning-based locomotion frameworks. His 2019 paper on trajectory-based deep policy search introduced an innovative approach to reinforcement learning that optimizes policies at the walking cycle level rather than individual timesteps, representing a meaningful departure from conventional methods. Complementing this, his work on linear policy approaches demonstrates a commitment to computationally lightweight solutions suitable for low-cost hardware — a practically significant consideration for real-world robotics deployment. His gait library synthesis research further expanded these capabilities by generating diverse locomotion behaviors, including forward and backward trotting, side-stepping, and turning, using augmented random search techniques. With citations accumulating across his publications, Joglekar's research has contributed meaningfully to making reinforcement learning-based quadrupedal locomotion more accessible, efficient, and hardware-ready for the broader robotics community.