This repository contains code that demonstrates physically feasible Bayesian robotic manipulator parameter identification using Hamiltonian Monte Carlo (HMC).

• numpy
• matplotlib
• pystan 3
• pandas

If you would also like to regenerate the inverse dynamics code then you will need

• sympybotics
• sympy 0.7.5 (newer version don't work with sympybotics)

The simulated parameter identification example can be run using

python sim_3dof.py

This will use stan to sample from the posterior distribution and can take >10 hours. Hence presaved results have been included with the repository and can be plotted using

python plot_results.py

To regenerate the inverse dynamics code that is used in the model block of the stan model file you can run

python sympy_3dof.py