EE 547: Sensing and Planning in Robotics Syllabus

Fall 1998 instructor: Dr. Gaurav Sukhatme
www-robotics.usc.edu/~gaurav
gaurav@usc.edu

(Old Title: Software Methods in Robotics)

Description:

Introduction to software methods in robotics including sensing, sensor fusion, estimation, fault tolerance, sensor planning, robot control architectures, planning and learning.

• Introduction to Sensing, Planning and Robot Control (1 week)
• Uncertainty Management with Sensor Fusion and Planning (4 weeks)
  • Integrated Sensor System Modeling and Architectures.
  • Sensor Fusion with Statistical, Geometric, and Evidential Reasoning Approaches.
  • Constraint Satisfaction in Integrated Sensor Systems.
  • Estimation theory introduction.
  • Fault Detection, Identification and Reconfiguration.
  • Optimal Sensor Configuration and Placement
• Planning and Reasoning in Robotics (2 weeks)
  • Motion Planning and Assembly Planning.
  • Path and Trajectory Planning with Static/Dynamic and Holonomic/Non-holonomic Constraints.
  • A survey of state-of-the-art research in planning and reasoning.
• Adaptation and Learning (2 weeks)
  • Robot Learning techniques survey including supervised and unsupervised approaches
  • A survey of contemporary learning systems
• Computational Intelligence in Robotics (2 weeks)
  • Introduction to Neural Networks, Fuzzy Systems, and Evolutionary Computation using contemporary research examples.
  • Neural Network Based Visuo-Motor Coordination.
• Real-Time Robot Control Architectures (3 weeks)
  • Real-time Robotic Programming, Simulation, and Control Environments.
  • A Survey of Modern Robot Control Architectures

Required Text:

Sensor-based Advanced Robotics, a collection of papers edited by Sukhan Lee. (Will be made available for purchase by students from the department)

Recommended Readings:

1. Integration, Coordination and Control of Multi-sensor Robot Systems, by H. F. Durrant-Whyte, Kluwer, Boston, 1988.

2. Computer-Aided Mechanical Assembly planning, edited by Homem de Mello and Sukhan Lee, Kluwer, Boston, 1991.

3. Robot Motion Planning, Jean-Claude Latombe, Kluwer Academic, 1991.

4. Neural Networks in Robotics, edited by George Bekey and Ken Goldberg, Kluwer Academic, 1992.

Grading:

Grades will be based on written reports of papers as well as a final project. Class participation and discussion will also count. There will be no final exam or midterm.

Paper reports: 50%
Final project: 30%
Class participation: 20%

Prerequisites:

CS 561 Artifical Intelligence or Instructor Permission

Return to the course homepage
This page was created by Gaurav Sukhatme and is maintained by him.