🏃 Pole vaulting robots

Last updated on Mar 1, 2025 Project

Sports Biomechanics from the Perspective of Robotics

Table of Contents

Overview

Pole vaulting is a highly intriguing research topic in robotics. This sport involves several aspects that can be applied to robotics, including robot motion strategies utilizing a highly flexible pole, conversion of kinetic energy into potential energy using the pole, and generation of inertial movements while being supported by the environment through the pole. Among these, we focused on the aerial movements performed by athletes while hanging onto the pole, and proposed the Active Pole Bending Theory to model these movements. Furthermore, we developed robots capable of performing pole vaulting and experimentally validated the theory in a real-world setting, successfully achieving a jump height of 1.6 meters. By observing and analyzing sports from a robotics perspective and exploring various physical conditions, we aim to uncover new insights and benefit the sports.

Athlete Movements During the Pole Support Phase

In pole vaulting, it is commonly believed that after inserting the pole and initiating the jump, the athlete simply follows passive dynamics. However, in reality, while hanging onto the pole, athletes actively move their bodies to maximize their vaulting height. In fact, approx. 20% of the energy contributing to jump height is generated through the body movements performed while hanging on the pole. We analyzed these body movements using the Active Pole Bending Theory, interpreting and examining how athletes’ movements contribute to the system.

Active Pole Bending Theory

First, we proposed “Transition Buckling Model” to describe active bending behavior during the pole support phase, where the terminal conditions can be varied according to the bending input.

Toshihiko Fukushima, Satoshi Nishikawa, Kazutoshi Tanaka, Yasuo Kuniyoshi (2013). Transitional buckling model for active bending effect in pole vault. AMAM.

PDF Cite Project Poster

Simulation Study

Through extensive simulations, we revealed how initial velocity, pole length, bending amplitude, and switching time interact to affect vault height. Positive-bending helps verticalize the vault angle, while negative-bending boosts vault speed and resilience. We also found that switching between Positive-bending and Negative-bending was the most effective when done at the time of maximum pole deflection. These results indicates key strategies for improving both robot and athlete performance.

Toshihiko Fukushima, Satoshi Nishikawa, Yasuo Kuniyoshi (2022). Active pole bending effect in pole support phase. Sports Biomechanics.

PDF Cite Project Video DOI

PABRO (1st Gen. Robot)

PABRO performs pole vaulting and can exert bending moment on a pole during the pole support phase. Using this robot, we have demonstrated in the real world that active bending motion can improve vaulting height.

Comparison between non-actuation and active-bending

Vaulting over 1.2 m with the active bending

Toshihiko Fukushima, Satoshi Nishikawa, Yasuo Kuniyoshi (2014). Active bending motion of pole vault robot to improve reachable height. ICRA.

PDF Cite Project Video DOI

PAULTER (2nd Gen. Robot)

PAULTER, a single-arm pole vaulting robot, uses its own inertial force to apply a bending moment to the pole, enabling it to perform a motion more similar to an actual pole vault competition. Experimental results showed that a swing-up motion after the maximum deflection of the pole can significantly increase the vaulting height. This difference in optimal timing from the simulation results arises from the the robot’s own inertial forces when its CoM is offset from the pole.

PAULTER vaults with using its inertial force

Satoshi Nishikawa, Toshihiko Fukushima, Yasuo Kuniyoshi (2013). Effective timing of swing-up motion by a pole-vaulting robot. ICAR.

PDF Cite Project DOI

PAULTER II (3rd Gen. Robot)

PAULTER II is a dual-armed robot that can also release a pole in the air. Using this robot and a multiple-pendulum simulation, we found that the direction and trajectory of the vaulting can be controlled by the switch timing of the active bending motions. By using an optimal strategy, PAULTER II successfully vaulted 1.6 m height.