Asimov Robotic Interface

Developing and implementing brain-based models of sensory perception, navigation, and decision making is an active area of research in robotics. This is because while robots can greatly exceed humans’ ability to obtain sensory information about their environment, they are barely able to approach the data processing ability humans exercise almost effortlessly. Not only will better models increase the functionality of robots, but they may also provide insight into how our own brains work.

Enter: Asimov. Asimov is a Python-based middleware application that facilitates the development and testing of brain-based sensorimotor models on the iRobot Create by enabling remote machine control. This allows access to high-power computing resources for sophisticated models and tools.


Asimov is designed around five primary goals:

  • expose control of an iRobot Create with a consistent, simple, stable network interface;
  • allow simultaneous robot control from multiple, potentially incompatible processes;
  • minimize wireless network traffic;
  • eliminate the need for manual network configuration; and
  • localize Create-specific code so additional robotic platforms can easily be added later.

Asimov achieves this through a combination of Python code running on a netbook mounted on the Create, an atomic streaming key-value store for transmitting sensorimotor signals, and interfaces to this store written for a variety of platforms.

References:
Galbraith B., Versace M., and Chandler B. (2011) Asimov: Middleware for Modeling the Brain on the iRobot Create. PyCon, Atlanta, March 11-13th 2011.

NL team working on this project: Byron Galbraith, Ben Chandler

Collaborators

The Neuromorphics Lab is highly collaborative with connections across both academia and industry.