Neural Networks for Real-Time Sensory Data Processing and Sensorimotor Control

Abstract

One of the critical problems that we faced in modeling the neural control of escape turns in the cockroach was a quantitative analysis of the leg movements that generate the turn. Because the turn is very rapid, often spanning no more than 20 msec, conventional motion analysis techniques could not be employed. For that reason, we acquired an NAC high speed video and computer motion analysis system. With these tools, we have not attained a very good description of the movements that occur at each joint of each leg during turns in response to wind from four different angles relative to the animals body. Figure 1 is a typical stick figure generated by the computer system showing four successive frames at 5 msec intervals. The pairs of legs shown are, from front to back, legs in the prothoracic (T1), mesothoracic (T2), and metathoracic segments. In all legs the coxa, femur, and tibia segments are shown. These segments make two critical joint angles: the coxal-femur (C-F) joint and the femoro-tibial (F-T) joint. Keywords: Robotics.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 14, 1990
Accession Number
ADA223096

Entities

People

  • Randall D. Beer

Organizations

  • Case Western Reserve University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Assembly
  • Computers
  • Data Processing
  • Electromagnetic Interference
  • High Acceleration
  • High Level Language Architecture
  • Long Life
  • Neural Networks
  • Neurons
  • Potentiometers
  • Robots
  • Simulations
  • Three Dimensional
  • Video Cameras
  • Wind
  • Wind Direction

Readers

  • Robotics and Automation.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • Autonomy