Feng Tan Email & Phone Number

Cambridge, Massachusetts, Us

Cambridge, Ma, Us

Research Area: Multi-Agent Synchronization, Nonlinear Control, Machine Learning, Robotics▶Research Project I(2011-Present): Bio-inspired Data Clustering Algorithm • Designed a clustering algorithm inspired by quorum sensing, successfully bridging modern Machine Learning techniques with adaptive control and estimation.• Achieved better performance on static data benchmarks than existing classical algorithms.• Smoothly incorporate the algorithm into dynamic system control problems.• Realized mobile robot clustering, adaptive controller clustering, adaptive control parameter identification and novel multiple model switching control, which cannot be solved by traditional methods.▶Research Project II(2010-2011): Image segmentation with oscillator network • Studied oscillator networks attributes and synchronizing properties using Contraction Analysis.• Designed an image segmentation algorithm inspired by spring-damper dynamic system.• Improved the algorithm using sub-segmentation and mosaic strategy to utilize parallel computing powerPotential projects: Adaptive control parameter identification

Woburn, Ma, Us

• Proposed improvements on bicycle model. Provided data analysis on pressure rippling in the hydraulic system• Designed algorithms for estimating leakage coefficients of the hydraulic pump using least squares and filtering

Billerica, Ma, Us

• Designed a metric measuring linearity of boundary sensors, which influences line following performance.• Designed, set and committed a series of tests concerning robot performance following fluorescent boundary tapes contaminated by water and mud by different level. Found root cause of tapes being water-logged.• Designed new algorithms for Lidar sensor to detect and evaluate shape and radius of objects in field. Proposed criterions to determine probability of pots existence.

The MIT-China Innovation and Entrepreneurship Forum (MIT-CHIEF) is an annually-held conference and high-profile platform for intellectual exchange to foster deeper and broader collaboration between China and the United States in technology, innovation, and entrepreneurship. On November 19-20, 2011, leading authorities from academia, industries, enterprises, businesses, and government institutions, from both China and the US, will convene at MIT to deliver presentations and showcase their work. We intend to address not only current developments and collaborations between enterprises in the US and China in cutting-edge fields such as clean technology, information technology, health care, and biotechnology, but also policy-making issues and future directions China should take to facilitate continued innovation in the country. Additionally, for budding entrepreneurs, we also feature a "Pitch to China" Business Plan Contest with cash prizes totalling $18,000.

▶Research on Humanoid Stable Walking with a Large Mass Torso (Final Thesis)• Designed a six-controller-collaborating reflex control strategy realizing stable walking. • Proved the proposed strategy reliable and gained a better simulation performance than peer work.• Proposed adequate analysis on control differences between biped robots with/without massive torso.▶Research on Biped Passive Humanoid Walking on Slope Changing Terrains•Conducted mechanical design, dynamic&control strategy simulation in Robotics&Automation Lab•Improved the dynamic model from point feet to arc-shaped feet in Matlab/Simulink environment.•Wrote M-files generating accurate real-time animations of robot walking.•Conducted research on walking strategy based on Central Pattern Generator.•Designed a hierarchical reflex control system, improved walking performance on changing slopes.•Authored a paper for 2010 IEEE International Conference on Robotics and Automation.

▶Designed “Tsinghua Student Robot Climber”, won the First Prize in the 9th Machine Innovative Competition of Tsinghua University, JEC Special Prize for best mechanical design. ▶Undertook a fund of National Innovation Experiment Plan of Student, designed a positioning method combining RFID and image recognition.

Ph.D. Candidate, Control, Robotics, Neuroscience

Bachelor, Meche, Robotics