TY - GEN
T1 - A comparison of analytical and empirical controllers for the SLIP hopper
AU - Hale, Matthew F.
AU - du Bois, Jonathan L.
AU - Iravani, Pejman
PY - 2017/7/20
Y1 - 2017/7/20
N2 - The Spring-Loaded Inverted Pendulum (SLIP) model poses a challenging control problem, important for the development of legged robots, due to the difficultly in solving the stance phase of the dynamics. Multiple attempts have been made to approximate these dynamics to allow for an analytical control method; here four of these methods have been compared for controlling agile hopping, where there are large changes in forward velocity across a single stance. In addition, a new, empirical, approach has been demonstrated. In this, a simple control law is formulated, based on some simple approximations, which allows the parameters to be selected empirically through simulation. This has led to a controller able to offer similar performance to the best analytical approximation but with a much simpler form. This empirical controller may present new opportunities for controlling more complex dynamics and the development of a self-tuning method in future work.
AB - The Spring-Loaded Inverted Pendulum (SLIP) model poses a challenging control problem, important for the development of legged robots, due to the difficultly in solving the stance phase of the dynamics. Multiple attempts have been made to approximate these dynamics to allow for an analytical control method; here four of these methods have been compared for controlling agile hopping, where there are large changes in forward velocity across a single stance. In addition, a new, empirical, approach has been demonstrated. In this, a simple control law is formulated, based on some simple approximations, which allows the parameters to be selected empirically through simulation. This has led to a controller able to offer similar performance to the best analytical approximation but with a much simpler form. This empirical controller may present new opportunities for controlling more complex dynamics and the development of a self-tuning method in future work.
UR - http://www.scopus.com/inward/record.url?scp=85026759487&partnerID=8YFLogxK
UR - http://dx.doi.org/10.1007/978-3-319-64107-2_7
U2 - 10.1007/978-3-319-64107-2_7
DO - 10.1007/978-3-319-64107-2_7
M3 - Chapter in a published conference proceeding
AN - SCOPUS:85026759487
SN - 9783319641065
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 79
EP - 85
BT - Towards Autonomous Robotic Systems - 18th Annual Conference, TAROS 2017, Proceedings
PB - Springer Verlag
CY - Germany
T2 - 18th Annual Conference on Towards Autonomous Robotic Systems, TAROS 2017
Y2 - 19 July 2017 through 21 July 2017
ER -