AuRA: Autonomous Robot Architecture From: Integrating Behavioral, Perceptual, and World Knowledge in Reactive Navigation Ron Arkin, 1990.

AuRA: Autonomous Robot Architecture

From: Integrating Behavioral, Perceptual, and World Knowledge in Reactive Navigation

Ron Arkin, 1990

Architecture

Hierarchical Planning1. Generate possible linear segments

Hierarchical Planning2. Find a piecewise linear plath

Hierarchical Planning3. Generate a sequence of schema groupings

Using World Knowledge & Reactive Schemas

• World knowledge can be used to determine the general area in which to move or look for a landmark

• When the landmark is found, then reactive navigation takes over

• Arkin’s claim: World knowledge is necessary for efficient, flexible, and generalizable navigation

• Do you agree?

Docking

Docking in a Cluttered Environment

Long Term Memory for persistent, a priori info

Short Term Memory for dynamically acquired info

Cartographer for building

a mapHomeostatic

control to monitor internal

conditions

Discussion Questions

• What is the best way of handling traps in vector potential fields? Noise addition, deadlines, or ??

• Can the noise schema ever cause the robot to become trapped beyond its capability to recover?

• Are there times when the methodology of low-level reactive planning will fail?

• What would happen if a motor schema failed to activate correctly, say, due to a hardware failure? Could the robot self-correct?

Discussion Questions• The output of the schemas is a velocity vector. Are we

assuming a nearly instantaneous change in velocity or a delay?

• In the experiments conducted, a particular velocity is used. What if the velocity were increased?

• How does the schema-based navigation relate to the way humans make navigational decisions?

• How would the potential field navigation be adaptable to 3 dimensions?

3D Schema-Based Navigation

Discussion Questions• If the perceptual schema confidence exceeds the motor

schema threshold for action, the motor schema starts to produce a repulsive field surrounding the obstacle. Why?

• What happens if obstacles suddenly appear or disappear?• How does the robot adapt?• What environments are idea for the use of schema-based

navigation?