TY - GEN
T1 - A cross-validation approach to trajectory and shape reconstruction for rigid bodies
AU - Yu, Jieqi
AU - Zheng, Haipeng
AU - Kulkarni, Sanjeev R.
AU - Poor, H. Vincent
PY - 2010/11/24
Y1 - 2010/11/24
N2 - In this paper, a method is proposed for reconstructing the trajectory and shape of a rigid body in a damped environment from distributively collected, asynchronous data. In this problem setting, both the shape parameters of the rigid body and its trajectory are unknown. The shape/trajectory recovery problem is modeled as a minimization of energy dissipation under geometric and acceleration constraints. In order to solve this problem, a convex relaxation for the geometric constraint is introduced, and the geometric constraint is reinforced in a cross-validation stage to verify the parameters. In this manner the shape and the trajectory of the rigid body are reconstructed simultaneously. For simplicity, a two-dimensional ball is taken as the rigid body prototype and simulations demonstrate the efficacy of the algorithm.
AB - In this paper, a method is proposed for reconstructing the trajectory and shape of a rigid body in a damped environment from distributively collected, asynchronous data. In this problem setting, both the shape parameters of the rigid body and its trajectory are unknown. The shape/trajectory recovery problem is modeled as a minimization of energy dissipation under geometric and acceleration constraints. In order to solve this problem, a convex relaxation for the geometric constraint is introduced, and the geometric constraint is reinforced in a cross-validation stage to verify the parameters. In this manner the shape and the trajectory of the rigid body are reconstructed simultaneously. For simplicity, a two-dimensional ball is taken as the rigid body prototype and simulations demonstrate the efficacy of the algorithm.
UR - http://www.scopus.com/inward/record.url?scp=78449271888&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=78449271888&partnerID=8YFLogxK
U2 - 10.1109/MLSP.2010.5589198
DO - 10.1109/MLSP.2010.5589198
M3 - Conference contribution
AN - SCOPUS:78449271888
SN - 9781424478774
T3 - Proceedings of the 2010 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2010
SP - 307
EP - 312
BT - Proceedings of the 2010 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2010
T2 - 2010 IEEE 20th International Workshop on Machine Learning for Signal Processing, MLSP 2010
Y2 - 29 August 2010 through 1 September 2010
ER -