Jump to the ATLANTIS Overview page.
Jump to the Comparative Capabilities page.
Jump to the Home page.

Capabilities

Planning

Planning is manifested in ATLANTIS because of the third layer: the deliberative layer.

Prediction

It wasn't explicitly stated, but prediction could be possible. Only in the deliberative layer , however, since it is the only level with world knowledge and planning capabilities.

Reactivity

This was one of the design goals for this architecture because of the types of environments mobile robots must be robust in. The reactivity of this system comes about because of the control layer.

Taskability

The agent can be instructed as to whatever goals it should perform. This is accomodated through the deliberative layer

Interruptability

The locus of interruptability resides in the sequencing layer.

Navigation/Manipulation

This was one of the design goals for this architecture. This capability has been successfully demonstrated on two agents.

Coherent Behavior

ATLANTIS agents exhibit goal-directed behavior. The behavior is also coherent in the sense that when instructed to perform a task, plans are generated based on the model of the world maintained in the deliberative layer.

Perception

Mobile robots require perception. The perception in ATLANTIS was handled by the control layer. This perception information no doubt made its way up to the deliberative layer in order to maintain the world model.

Two agents were developed to demonstrated the applicability of this architecture. The indoor agent incorporated twelve sonar sensors. The outdoor agent use stereo vision to sense its environment.