v3.0.1

This release fixes inequality vector data types for the wheeled inverted pendulum system, and adds some type checks to make sure generated problems use floating-point arrays.

Fixed

• Deprecation warning from using logging.warn
• Make sure stacked MPC matrices have floating-point data type
• Type of inequality vectors in wheeled inverted pendulum MPC

v3.0.0

This is a major release according to semantic versioning because the name of the project changed to qpmpc. In passing, the "cart pole" system was a misnomer and has been renamed to "wheeled inverted pendulum".

Changed

• Renamed the "cart pole" system (misnomer) to "wheeled inverted pendulum"
• Renamed the library to "qpmpc"

v2.0.0

This release makes a major API update to switch from cold-start to hot-start when solving of MPC QP problems. It meanwhile adds two examples (with live plots) taken from real-robot use cases:

• The open-loop MPC part of the LIPM walking controller: python examples/lipm_walking_controller.py
• A closed-loop MPC balancer for Upkie, which also runs on the real robots: python examples/cart_pole.py

The problem class, now called MPCProblem, allows for initial/goal state and target trajectory updates. A new MPCQP class is introduced for QP vector updates (cost vectors depend on the initial state and targets, while on LTI problems QP matrices don't depend on these changing quantities, a source of optimization for faster solve times).

The new API also introduces two new submodules:

• Live plots: a debugging tool for plotting MPC trajectories while running (uses the TkAgg matplotlib backend)
• Systems: we start gathering common systems that can be re-used outside of examples. For instance, CartPole shipped with this release is used in a closed-loop MPC balancer on Upkie.

Added

• Base class for exceptions raised by this library
• Documentation checks using ruff
• Exception: ProblemDefinitionError
• MPCProblem: setter for the initial state
• MPCProblem: target state trajectory for stage state cost
• MPCQP class to update cost vectors during execution
• Plan: first_input getter
• Plan: is_empty property
• Started ltv_mpc.live_plots submodule
• Started ltv_mpc.systems submodule

Changed

• Always add state-input inequalities (even when unfeasible)
• Don't assume a Problem is fully defined in constructor
• Initial and goal states are now keyword arguments of MPCProblem
• Refactorbuild_qp into a new MPCQP class
• Rename Problem to MPCProblem
• Rename Solution to Plan
• Rename stacked_inputs to just inputs in plans
• Rename stacked_states to just states in plans
• Solver keyword argument to solve_mpc is now mandatory
• Use problem class from qpsolvers internally
• Warn rather than raise an exception when initial state is unfeasible

v1.0.0

This version adds a solver keyword argument to solve_mpc to select the backend QP solver.

It follows from the API update in qpsolvers v2.0 where the default solver was deprecated and the solver argument became mandatory.

Added

• New mpc_interface alternative in the README

Changed

• ⚠️ API update: solve_qp now takes a mandatory solver keyword argument

Fixed

• Unit tests for the new solver keyword argument

v0.7.0

This minor version adds a sparse keyword argument, for use with sparse QP solvers like OSQP.

Added

• sparse keyword argument to use with sparse QP solvers
• Usage and examples documentation sections

Changed

• Only export Problem, Solution and solve_mpc module-wide

Fixed

• Edge case where inputs don't affect the first inequality constraints

v0.6.0

Here comes the first working version of this module, following an initial import from pymanoid and refactoring to a functional API.

This module defines a one-stop shop solve_mpc(problem: Problem) -> Solution function. The Problem type defines the linear model predictive control problem (system, constraints, initial state and cost function to optimize) while the Solution holds the resulting state and input trajectories.

Added

• Initial import from pymanoid's mpc.py.
• Humanoid stepping example
• Triple integrator example

Changed

• Extend to time-varying state and input transition matrices