Tino is a mobile robot platform of the
developed at SRI International,
quite suitable for operating within an office environment.
It is equipped with two independent driving wheels
and a free-wheeling castor, while
sensoring capabilities are provided by seven ultrasonic range finders
placed in the front of the platform.
Detailed hardware specifications
of the Erratic base are also available.
The robot is remote controlled from a host computer through a wireless communication. Control software runs on the remote workstation and real-time capabilities are obtained by means of a 10 Hz communication with the mobile platform.
Tino is based on a layered architecture, which combines a reactive control mechanism with a reasoning system. At the lower level a fuzzy controller is responsible for reactively executing primitive actions, while at the upper level a reasoning system is used to achieve high-level goals. The integration of planning and reactivity is obtained by means of a monitor, which both translates data representation between the two levels (allowing for a heterogeneous implementation) and checks whether the execution of actions leads to the accomplishment of the goals.
The main goal of our research work is to provide autonomous reasoning
capabilities to the robot by devising a principled and practically
feasible realization of a
KR&R approach for reasoning about actions.
Specifically, the basis of our proposal for reasoning about
actions is provided by Propositional Dynamic Logics (PDLs).
However, the implementation has been obtained by relying on the tight
correspondence that exists between PDLs and Description Logics (DLs).
By exploiting this correspondence we have been able both to develop
an interesting theoretical framework for reasoning about actions and to
obtain an implementation that uses a knowledge representation system based
In fact, a general knowledge representation system based on DLs (CLASSIC) is used to represent the robot's knowledge about the environment and the actions it can perform. The reasoning services of CLASSIC allow the robot to deductively derive plans to achieve its goals, while its reactive capabilities allow it to execute its plans in the real world.
Experiments involve navigation tasks in an office-like environment.
The robot accepts high-level goals, computes a plan whose performance
is expected to achieve the goal and reactively executes it in the world.
First experiments were focussed on using moving actions to reach a specified goal and on the integration between the planning system and the fuzzy controller. Recently we introduce in our framework sensing actions, and we test them in some new experiments.
Robotics information source
The URL of this page is http://www.dis.uniroma1.it/~iocchi/tino/tino.html.
Last updated on May 6, 1997.