Introduction

This repository contains data and code from optimal control problems (OCP) of a planar digital twin performing a backward tuck somersault. The simulations explore the impact of limb-on-limb contact during the "tuck" phase, specifically how allowing the digital twin to pull its shanks with its hands (using holonomic constraints) affects movement compared to not having any contact or just having contact without force (kinematic constraints).

https://github.com/user-attachments/assets/d0aacbbe-1787-49c2-922c-8e6e37570ee7

Cite this work

bibtex @article{Farr2025, title = {Including limb-on-limb holonomic constraints in predictive simulation allows replicating athlete’s backflip technique}, journal = {Multibody System Dynamics}, volume = {}, pages = {}, year = {2025}, issn = {}, doi = {https://doi.org/10.1007/s11044-025-10082-0}, url = {https://rdcu.be/emfNR}, author = {A. Farr, E. Charbonneau, M. Begon and P. Puchaud}, keywords = {Holonomic constraint, Constraint dynamics, Predictive simulation, Optimal control, Biomechanics, Gymnastics, Closed-loop}

Status

| Type | Status | |---|---| | Zenodo | |

How to install dependencies

In order to run the code, you need to install the following packages from pyomeca: bash conda install -c conda-forge biorbd=1.11.1 python-graphviz matplotlib numpy scipy pyorerun bioviz=2.3.2 plotly

Installing Bioptim from source

1. Clone the repository: bash git clone https://github.com/Ipuch/bioptim.git cd bioptim git checkout ReleasePaperBackTuck
2. Install the library: bash python setup.py install # Setting Linear Solver for Ipopt To enhance optimization, configure the linear solver (e.g., ma57) in your script: python from bioptim import Solver solver = Solver.IPOPT() solver.set_linear_solver("ma57") Otherwise, use mumps, but you won't replicate the results.

Exemples

In the folder examples\, you will find the following examples (NTC, KTC, HTC): - somersault_htc.py: This script generates the optimal control problem with the holonomic constraint. - somersault_htc_taudot.py: This script generates the optimal control problem with the holonomic constraint and torque derivative driven - somersault_ktc_taudot.py: This script generates the optimal control problem with the kinematic constraint and torque derivative driven - somersault.py: This script generates the optimal control problem without the holonomic constraint. - somersault_taudot.py: This script generates the optimal control problem without the holonomic constraint and torque derivative driven