Intelligent Global Vision for Teams of Mobile Robots

Baltes, Jacky; Anderson, John

doi:10.5772/4773

Cited by 5 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a more challenging scenario, we also used a more degenerate configuration (on the left, which puts the robot in a position to more easily score on its own goal while manoeuvring). The individual demonstrators were recorded by the Ergo global vision system [17] while they performed 25 goal kicks for each of the two field configurations. The global vision system continually captures the x and y motion and orientation of the demonstrating robot and the ball.…”

Section: Resultsmentioning

confidence: 99%

“…In our implementation we define five primitive actions reflective of the atomic motor commands available to the wheeled imitator: (forward, backward, left, right and stop). We assume no initial knowledge of an imitator's own actions, and begin by having the imitator collect visual data of the outcomes of its own primitive actions using the Ergo vision system [17], to create its own model of what it can accomplish in the world. Once the robot has the ability to recognize the outcomes of its own action, primitive mimicry is possible.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Vision-Based Imitation Learning in Heterogeneous Multi-Robot Systems: Varying Physiology and Skill

Allen

Anderson

Baltes

2012

AUSMT

Self Cite

View full text Add to dashboard Cite

Imitation learning enables a learner to improve its abilities by observing others. Most robotic imitation learning systems only learn from demonstrators that are similar physically and in terms of skill level. In order to employ imitation learning in a heterogeneous multi-agent environment, we must consider both differences in skill, and physical differences (physiology, size). This paper describes an approach to imitation learning from heterogeneous demonstrators, using global vision. It supports learning from physiologically different demonstrators (wheeled and legged, of various sizes), and self-adapts to demonstrators with varying levels of skill. The latter allows different parts of a task to be learned from different individuals (that is, worthwhile parts of a task can still be learned from a poorly-performing demonstrator). We assume the imitator has no initial knowledge of the observable effects of its own actions, and train a set of Hidden Markov Models to create an understanding of the imitator’s own abilities. We then use a combination of tracking sequences of primitives and predicting future primitives from existing combinations of primitives, using forward models to learn abstract behaviors from demonstrations. This approach is evaluated using a group of heterogeneous robots that have been previously used in RoboCup soccer competitions

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%