Catching ground balls: Optical control heuristics used by humans and robots support a unified fielder theory

Michael McBeath, Thomas Sugar, Michael J. Thompson, Keshav Mundhra

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In past studies we examined optical control heuristics used by fielders and robots for interception of fly balls. Our findings support use of the heuristics of maintaining optical speed constancy and a linear optical trajectory (LOT) for balls projected above the horizon. In the current study we explore behavior of robots and fielders intercepting ground balls projected below the horizon. In robotic simulations, we confirmed that the same control heuristics demonstrated for fly balls are viable ones for intercepting ground balls. In the human experiment, we measured optical position of the ball with three skilled fielders using a head cam, and measured actual ball and fielder location using stationary external cameras. Our findings support that, for most easily caught grounders, fielders select a slow running pace and maintain the same optical control heuristics that they do for fly balls. This also results in an energy-efficient, near constant-speed running path. On a number of the more difficult trials, where fielders needed to run a long distance, the optical trajectory was much better accounted for as two phase: an initial constant-speed linear trajectory at a severe lateral angle, followed by a second constant-speed linear trajectory at a new diminished lateral angle. On these trials it appears fielders overshoot the first LOT and establish a second one with a more workable approach angle. The findings also support that fielders tend to establish a running path with a fairly constant velocity that remains roughly within a vertical plane perpendicular to their initial lineup with the ball. Overall, the findings are consistent with a unified fielder theory in which the same optical control heuristics are used to navigate to the interception destination for both fly balls headed above the horizon and grounders below it. Use of these heuristics holds promise in the creation of navigating mobile robots designed to achieve or avoid collisions.

Original languageEnglish (US)
JournalJournal of Vision
Volume3
Issue number9
DOIs
StatePublished - 2003

Fingerprint

Diptera
Running
Robotics
Heuristics
Head

ASJC Scopus subject areas

  • Ophthalmology

Cite this

Catching ground balls : Optical control heuristics used by humans and robots support a unified fielder theory. / McBeath, Michael; Sugar, Thomas; Thompson, Michael J.; Mundhra, Keshav.

In: Journal of Vision, Vol. 3, No. 9, 2003.

Research output: Contribution to journalArticle

@article{8ef62bf402b546bea01888deb45d33b1,
title = "Catching ground balls: Optical control heuristics used by humans and robots support a unified fielder theory",
abstract = "In past studies we examined optical control heuristics used by fielders and robots for interception of fly balls. Our findings support use of the heuristics of maintaining optical speed constancy and a linear optical trajectory (LOT) for balls projected above the horizon. In the current study we explore behavior of robots and fielders intercepting ground balls projected below the horizon. In robotic simulations, we confirmed that the same control heuristics demonstrated for fly balls are viable ones for intercepting ground balls. In the human experiment, we measured optical position of the ball with three skilled fielders using a head cam, and measured actual ball and fielder location using stationary external cameras. Our findings support that, for most easily caught grounders, fielders select a slow running pace and maintain the same optical control heuristics that they do for fly balls. This also results in an energy-efficient, near constant-speed running path. On a number of the more difficult trials, where fielders needed to run a long distance, the optical trajectory was much better accounted for as two phase: an initial constant-speed linear trajectory at a severe lateral angle, followed by a second constant-speed linear trajectory at a new diminished lateral angle. On these trials it appears fielders overshoot the first LOT and establish a second one with a more workable approach angle. The findings also support that fielders tend to establish a running path with a fairly constant velocity that remains roughly within a vertical plane perpendicular to their initial lineup with the ball. Overall, the findings are consistent with a unified fielder theory in which the same optical control heuristics are used to navigate to the interception destination for both fly balls headed above the horizon and grounders below it. Use of these heuristics holds promise in the creation of navigating mobile robots designed to achieve or avoid collisions.",
author = "Michael McBeath and Thomas Sugar and Thompson, {Michael J.} and Keshav Mundhra",
year = "2003",
doi = "10.1167/3.9.543",
language = "English (US)",
volume = "3",
journal = "Journal of Vision",
issn = "1534-7362",
publisher = "Association for Research in Vision and Ophthalmology Inc.",
number = "9",

}

TY - JOUR

T1 - Catching ground balls

T2 - Optical control heuristics used by humans and robots support a unified fielder theory

AU - McBeath, Michael

AU - Sugar, Thomas

AU - Thompson, Michael J.

AU - Mundhra, Keshav

PY - 2003

Y1 - 2003

N2 - In past studies we examined optical control heuristics used by fielders and robots for interception of fly balls. Our findings support use of the heuristics of maintaining optical speed constancy and a linear optical trajectory (LOT) for balls projected above the horizon. In the current study we explore behavior of robots and fielders intercepting ground balls projected below the horizon. In robotic simulations, we confirmed that the same control heuristics demonstrated for fly balls are viable ones for intercepting ground balls. In the human experiment, we measured optical position of the ball with three skilled fielders using a head cam, and measured actual ball and fielder location using stationary external cameras. Our findings support that, for most easily caught grounders, fielders select a slow running pace and maintain the same optical control heuristics that they do for fly balls. This also results in an energy-efficient, near constant-speed running path. On a number of the more difficult trials, where fielders needed to run a long distance, the optical trajectory was much better accounted for as two phase: an initial constant-speed linear trajectory at a severe lateral angle, followed by a second constant-speed linear trajectory at a new diminished lateral angle. On these trials it appears fielders overshoot the first LOT and establish a second one with a more workable approach angle. The findings also support that fielders tend to establish a running path with a fairly constant velocity that remains roughly within a vertical plane perpendicular to their initial lineup with the ball. Overall, the findings are consistent with a unified fielder theory in which the same optical control heuristics are used to navigate to the interception destination for both fly balls headed above the horizon and grounders below it. Use of these heuristics holds promise in the creation of navigating mobile robots designed to achieve or avoid collisions.

AB - In past studies we examined optical control heuristics used by fielders and robots for interception of fly balls. Our findings support use of the heuristics of maintaining optical speed constancy and a linear optical trajectory (LOT) for balls projected above the horizon. In the current study we explore behavior of robots and fielders intercepting ground balls projected below the horizon. In robotic simulations, we confirmed that the same control heuristics demonstrated for fly balls are viable ones for intercepting ground balls. In the human experiment, we measured optical position of the ball with three skilled fielders using a head cam, and measured actual ball and fielder location using stationary external cameras. Our findings support that, for most easily caught grounders, fielders select a slow running pace and maintain the same optical control heuristics that they do for fly balls. This also results in an energy-efficient, near constant-speed running path. On a number of the more difficult trials, where fielders needed to run a long distance, the optical trajectory was much better accounted for as two phase: an initial constant-speed linear trajectory at a severe lateral angle, followed by a second constant-speed linear trajectory at a new diminished lateral angle. On these trials it appears fielders overshoot the first LOT and establish a second one with a more workable approach angle. The findings also support that fielders tend to establish a running path with a fairly constant velocity that remains roughly within a vertical plane perpendicular to their initial lineup with the ball. Overall, the findings are consistent with a unified fielder theory in which the same optical control heuristics are used to navigate to the interception destination for both fly balls headed above the horizon and grounders below it. Use of these heuristics holds promise in the creation of navigating mobile robots designed to achieve or avoid collisions.

UR - http://www.scopus.com/inward/record.url?scp=4243064035&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=4243064035&partnerID=8YFLogxK

U2 - 10.1167/3.9.543

DO - 10.1167/3.9.543

M3 - Article

AN - SCOPUS:4243064035

VL - 3

JO - Journal of Vision

JF - Journal of Vision

SN - 1534-7362

IS - 9

ER -