Fasih Ud Din Farrukh
Papers
8
Total Citations
36
H-Index
3
About
Fasih Ud Din Farrukh is a robotics and embedded systems researcher whose work bridges reinforcement learning, hardware acceleration, and simultaneous localization and mapping (SLAM). His primary research areas include robot locomotion control, bipedal gait optimization, and energy-efficient hardware design for autonomous systems. Farrukh's major contributions include developing a portable accelerator for Proximal Policy Optimization (PPO) that enhances reinforcement learning efficiency for robotic control, and creating the LORM framework for biped gait control that simplifies complex dynamics through RL. His work on motion sequence learning for robot walking, which combines traditional control with deep RL, has been cited 8 times and addresses critical convergence issues in bipedal locomotion. In hardware design, Farrukh has achieved notable milestones, including a 325 FPS corner-detection accelerator for SLAM systems and a 197-μJ/frame single-frame bundle adjustment accelerator for mobile visual odometry. His FPSNET architecture for neural-network-based feature point extraction in SLAM demonstrates his expertise in hardware-software co-design. With over 30 citations across his publications, Farrukh's work is advancing the practical deployment of intelligent robotic systems, particularly in resource-constrained mobile platforms.
Research Focus
Key Achievements
Top Papers
- 1A Portable Accelerator of Proximal Policy Optimization for Robots11 citations · 2021
- 2Motion Sequence Learning for Robot Walking Based on Pose optimization8 citations · 2020
- 3LORM: a novel reinforcement learning framework for biped gait control6 citations · 2022
- 4
- 5
- 6
- 7
- 8