Impulsive Hybrid System Optimization

Collaborators: Todd Murphey

In this project, I derived the second order switching time and magnitude optimization of an impulsive hybrid system. One application of this optimization method is in feature detection and localization of a surface traced by a sensor (e.g., a robotic finger). This optimization method allows me to detect and localize a surface feature based on the kinematics and dynamics of the sensor. The algorithm involves two parts: feature type estimation and feature localization. The first part of the algorithm takes the kinematic data of the sensor and estimates the feature types and sequence through a relaxation of an impulsive hybrid system optimization. With the information on the feature types and sequence, the second part of the algorithm localizes the boundary of each feature using the impulsive hybrid system optimization.

Phantom Surface

The figures show the PHANToM OMNI haptic device used for collecting experimental data to test the surface feature detection algorithm.


  1. Y. P. Leong and T. D. Murphey, “Surface feature localization using kinematics and impulsive hybrid optimization”, in IEEE Transactions on Automation Science and Engineering, vol. 10, no. 4, pp. 957–968, 2013.
  2. Y. P. Leong and T. D. Murphey, “Second order switching time and magnitude optimization for impulsive hybrid systems”, in American Control Conference 2013, pp. 6213–6218.
  3. Y. P. Leong, “Surface feature detection based on proprioception of a robotic finger during haptic exploration,” M. S. Thesis, Northwestern University, 2012.