atsushisakai / pyadvancedcontrol Goto Github PK

View Code? Open in Web Editor NEW

488.0 19.0 165.0 1.06 MB

Python codes for advanced control

License: MIT License

Python 49.99% Shell 0.04% Jupyter Notebook 49.97%

python mpc control

pyadvancedcontrol's Introduction

PyAdvancedControl

Python Codes for Advanced Control

Dependencies

Python 3.7.x
cvxpy 1.0.x
ecos 2.0.7
cvxopt 1.2.x
scipy 1.1.0
numpy 1.15.0
matplotlib 2.2.2

lqr_sample

This is a sample code of Linear-Quadratic Regulator

This is LQR regulator simulation.

This is LQR tracking simulation.

finite_horizon_optimal_control

This is a finite horizon optimal control sample code

mpc_sample

This is a sample code of a simple Model Predictive Control (MPC) regulator simulation

mpc_tracking

This is a sample code of a Model Predictive Control (MPC) traget tracking simulation

mpc_modeling

This is a sample code for model predictive control optimization modeling without any modeling tool (e.g cvxpy)

This means it only use a solver (cvxopt) for MPC optimization.

It includes two MPC optimization functions:

1 opt_mpc_with_input_const()

It can be applied input constraints (not state constraints).

2 opt_mpc_with_state_const()

It can be applied state constraints and input constraints.

This figure is a comparison of MPC results with and without modeling tool.

inverted_pendulum_mpc_control

This is a inverted pendulum mpc control simulation.

tools

c2d

This is a API compatible function of MATLAB c2d function.

Convert model from continuous to discrete time MATLAB c2d

pyadvancedcontrol's People

Contributors

Stargazers

Watchers

Forkers

haochihlin huiyi1990 bbrito capri2014 yidann huleg sirawats robotics-mark zxsted tiagoprata donrv boyali pb-pravin kevinladlee kaustav6422 jfjensen lincolnxlw joosm theabhithakan cyy1991 fangnan99 amdshameer frankszn buanderie abeldavid weisili2016 kazimbalti liujiangjiang xeoncores ltinphan alanxu89 het369 leo123456789011 ntwong0 deepakingole menson168 stevemats aegonwei fighterxi michaelchi08 yangwei07 hankerbit ghback ke-wang1017 jaxx016 neilpradhan hyzcn zhexiaozhe whc0114 littleggghost jiaquan301 minaek kejingjing88212 jnswx weiyizhibj knut0815 otoshuki masterhou budongyue lntsh854 malwinapasz xuegangzhang megayeye liuyuan0123 sbravell boamad3 kuangenzhang tsukuyomih2 baoquanm cyan-at zgrobot zhaoningshengg zzhuuh xuenke msareebhakak nicoleorzan qifan-sailboat zhaofengwhu cyrobot githubbeinner kkmonster rittee qff892444 muhammadbakr ciniks117 jmcrobbie rafeemusabbir kevin3062 surendrasah lakshminarayanareddybn zhurp ducquang1 wangqiongse dongchen06 qishuailiu mceb hitersyw duonghieumai hebowei2000 jsung0218

pyadvancedcontrol's Issues

Guides on how to use the regulators?

I am very new to regulators and controllers in general. It would be helpful if there were some information on how the provided controllers could be applied to specific use-cases.

In my specific use-case I'm trying to replace a system that uses two PID controllers with a single LQR (or other) controller. There are two variables that can be adjusted, but they both affect a single observable variable that I want to be a set value. In my understanding I'm looking for a MISO (multi input single output) controller. Is this possible to achieve with any of the controllers in this repo or are they only SIMO?

I especially enjoyed this repo since it only uses numpy (and scipy.linalg which I could replace with numpy.linalg).

some bug ?

when i run mpc_path_tracking and mpc_sample

Traceback (most recent call last):
File "main.py", line 177, in
Main()
File "main.py", line 129, in Main
ustar, xstar, cost = CalcInput(A, B, C, x, u)
File "main.py", line 84, in CalcInput
constr = [x[:, t + 1] == A * x[:, t] + B * u[:, t] + C]
File "/usr/local/lib/python3.5/dist-packages/cvxpy/expressions/expression.py", line 46, in cast_op
return binary_op(self, other)
File "/usr/local/lib/python3.5/dist-packages/cvxpy/expressions/expression.py", line 452, in add
return cvxtypes.add_expr()([self, other])
File "/usr/local/lib/python3.5/dist-packages/cvxpy/atoms/affine/add_expr.py", line 33, in init
super(AddExpression, self).init(arg_groups)
File "/usr/local/lib/python3.5/dist-packages/cvxpy/atoms/atom.py", line 45, in init
self._shape = self.shape_from_args()
File "/usr/local/lib/python3.5/dist-packages/cvxpy/atoms/affine/add_expr.py", line 41, in shape_from_args
return u.shape.sum_shapes([arg.shape for arg in self.args])
File "/usr/local/lib/python3.5/dist-packages/cvxpy/utilities/shape.py", line 49, in sum_shapes
len(shapes)" %s" % tuple(shapes))
ValueError: Cannot broadcast dimensions (4,) (4, 1)

Exception occured when running inverted_pendulum_mpc_control.py

$ python3 inverted_pendulum_mpc_control.py

Exception has occurred: exceptions.TypeError
Variable name 31 must be a string.
  File "/home/***/code/mpc_inv/inverted_pendulum_mpc_control.py", line 67, in mpc_control
  File "/home/***/code/mpc_inv/inverted_pendulum_mpc_control.py", line 43, in main
  File "/home/***/code/mpc_inv/inverted_pendulum_mpc_control.py", line 188, in <module>

Which versions are you using?

Could you perhaps write in the README which versions of the various Python libraries you are using when running the tests?
I'm thinking of Numpy,Scipy, ecos, cvxpy, etc...
Thanks!