TY - GEN
T1 - Prediction of a Ball Trajectory for the Humanoid Robots
T2 - 22nd RoboCup International Competition and Symposium, RoboCup 2018
AU - Yazdankhoo, Behnam
AU - Shahsavari, Mohammad Navid
AU - Sadeghnejad, Soroush
AU - Baltes, Jacky
N1 - Publisher Copyright:
© Springer Nature Switzerland AG 2019.
PY - 2019
Y1 - 2019
N2 - Recent advances in robotics have made it necessary for robots to be able to predict actions like humans. This problem is well presented in international RoboCup competition leagues, especially for humanoid robots in challenges such as Goal-Kick from Moving Ball. In this paper, we proposed double exponential smoothing (DES), autoregressive (AR) and quadratic prediction (QP) as online methods and self-perturbing recursive least squares (SPRLS) as an offline method for prediction of the ball trajectory on ground. These prediction methods are compared in two scenarios by applying LuGre friction model. We simulated our proposed methods by Simmechanics library of MATLAB’s Simulink. By comparing results using root-mean-square error and normalized root-mean-square error, we could deduce that methods that were based on predefined models such as QP performed poorly when the friction deviated from the presumed model. Whereas numerical methods such as AR could adapt themselves to variation much better, depending on the friction force variation with time. Also offline methods such as SPRLS are good replacements for online ones when pre-training is possible.
AB - Recent advances in robotics have made it necessary for robots to be able to predict actions like humans. This problem is well presented in international RoboCup competition leagues, especially for humanoid robots in challenges such as Goal-Kick from Moving Ball. In this paper, we proposed double exponential smoothing (DES), autoregressive (AR) and quadratic prediction (QP) as online methods and self-perturbing recursive least squares (SPRLS) as an offline method for prediction of the ball trajectory on ground. These prediction methods are compared in two scenarios by applying LuGre friction model. We simulated our proposed methods by Simmechanics library of MATLAB’s Simulink. By comparing results using root-mean-square error and normalized root-mean-square error, we could deduce that methods that were based on predefined models such as QP performed poorly when the friction deviated from the presumed model. Whereas numerical methods such as AR could adapt themselves to variation much better, depending on the friction force variation with time. Also offline methods such as SPRLS are good replacements for online ones when pre-training is possible.
KW - Autoregressive
KW - Ball trajectory prediction
KW - Exponential smoothing
KW - Goal-Kick from Moving Ball
KW - Humanoid robots
KW - LuGre model
UR - http://www.scopus.com/inward/record.url?scp=85070692503&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85070692503&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-27544-0_32
DO - 10.1007/978-3-030-27544-0_32
M3 - Conference contribution
AN - SCOPUS:85070692503
SN - 9783030275433
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 387
EP - 398
BT - RoboCup 2018
A2 - Holz, Dirk
A2 - Genter, Katie
A2 - Saad, Maarouf
A2 - von Stryk, Oskar
PB - Springer Verlag
Y2 - 18 June 2018 through 22 June 2018
ER -