Michael T. Rosenstein

AAAI Conference 2005 System Paper

Remote Supervisory Control of a Humanoid Robot

• Michael T. Rosenstein
• Robert Platt

For this demonstration, participants have the opportunity to control a humanoid robot located hundreds of miles away. The general task is to reach, grasp, and transport various objects in the vicinity of the robot. Although remote “pick-and-place” operations of this sort form the basis of numerous practical applications, they are frequently error-prone and fatiguing for human operators. Participants can experience the relative difficulty of remote manipulation both with and without the use of an assistive interface. This interface simplifies the task by injecting artificial intelligence in key places without seizing higher-level control from the operator. In particular, we demonstrate the benefits of two key components of the system: a video display of predicted operator intentions, and a haptic-based controller for automated grasping.

ICRA Conference 2002 Conference Paper

Velocity-Dependent Dynamic Manipulability

• Michael T. Rosenstein
• Roderic A. Grupen

Measures of dynamic manipulability summarize a manipulator's capacity to generate accelerations for arbitrary tasks, and such measures are useful tools for the design and control of general-purpose robots. Existing measures, however, downplay the effects of velocity or else ignore them altogether. In this paper we derive the relationship between joint velocity and end-effector acceleration, and through case studies we demonstrate that velocity has a complex, non-negligible effect on manipulability. We also provide evidence that movement near a singularity is beneficial for certain tasks.

ICML Conference 2000 Conference Paper

Combining Reinforcement Learning with a Local Control Algorithm

• Jette Randløv
• Andrew G. Barto
• Michael T. Rosenstein

AAAI Conference 1999 Conference Paper

Continuous Categories for a Mobile Robot

• Michael T. Rosenstein
• Paul R. Cohen
• University of Massachusetts

Autonomous agents makefrequent use of knowledgein the formof categories -- categories of objects, human gestures, webpages, and so on. This paper describes a wayfor agents to learn such categories for themselves through interaction with the environment. In particular, the learning algorithmtransformsraw sensor readings into clusters of time series that havepredictive value to the agent. Weaddress several issues related to the use of an uninterpreted sensory apparatus and showspecific exampleswherea Pioneer 1 mobilerobot interacts withobjects in a cluttered laboratorysetting.

AAAI Conference 1998 Conference Paper

Concepts from Time Series

• Michael T. Rosenstein

This paper describes a wayof extracting concepts from streams of sensor readings. In particular, we demonstrate the value of attractor reconstruction techniques for transforming time series into clusters of points. Theseclusters, in turn, represent perceptual categories with predictive value to the agent/environment system. Wealso discuss the relationship betweencategories and concepts, with particular emphasis on class membership and predictive inference.