https://github.com/upkie/mpc_balancer
Model predictive control to balance Upkie wheeled bipeds
https://github.com/upkie/mpc_balancer
legged-robots locomotion model-predictive-control python robotics wheeled-biped
Last synced: about 2 months ago
JSON representation
Model predictive control to balance Upkie wheeled bipeds
- Host: GitHub
- URL: https://github.com/upkie/mpc_balancer
- Owner: upkie
- License: apache-2.0
- Created: 2024-02-23T15:24:54.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-11-04T06:55:43.000Z (7 months ago)
- Last Synced: 2024-11-04T07:30:05.852Z (7 months ago)
- Topics: legged-robots, locomotion, model-predictive-control, python, robotics, wheeled-biped
- Language: Python
- Homepage:
- Size: 68.4 KB
- Stars: 9
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# MPC balancer
[](https://github.com/upkie/upkie/tree/v7.0.0)
Make an Upkie to stand upright by closed-loop model predictive control.
## Installation
We recommend using Anaconda to install the agent and all its dependencies in a clean environment:
```console
conda env create -f environment.yaml
conda activate mpc_balancer
```
Alternatively, you should be able to install the packages listed in the environment file from PyPI.
## Usage
To run in simulation, clone the [upkie](https://github.com/upkie/upkie) repository and run:
```console
./start_simulation.sh
```
Activate your conda environment and run the agent by:
```console
python run_agent.py
```
## Solvers
This agent only works with QP solvers that support warm starting. At present we only support one solver:
| Solver | Algorithm | License | Warm-start |
| ------ | --------- | ------- |------------|
| [ProxQP](https://github.com/Simple-Robotics/proxsuite) | Augmented Lagrangian | BSD-2-Clause | ✔️ |
You can take a peek at the [ProxQP balancer](https://github.com/stephane-caron/proxqp_balancer) (research code) for more solvers.
## Export dependencies to your Upkie
This agent can export a pixi environment to your Upkie using `pixi-pack`. If you don't have pixi yet, you will need to [install it](https://pixi.sh/latest/#installation) first. Then, to pack an environment from your computer, run:
```bash
make pack_env
```
This will create an `environment.tar` archive in the current directory. You can upload it to your Upkie by `make upload` and unpack it from the agent's remote directory by:
```bash
your_user@your_upkie:~/mpc_balancer$ make unpack_env
```
If `pixi-pack` is not installed on your Upkie, you can download the `pixi-pack-aarch64-unknown-linux-gnu` binary from the [pixi-pack release page](https://github.com/Quantco/pixi-pack/releases). Finally, activate the environment and run the agent:
```bash
conda activate mpc_balancer
python mpc_balancer/run.py
```
## See also
- [PPO balancer](https://github.com/upkie/ppo_balancer): an MLP agent trained for the same task by reinforcement learning.
- [ProxQP balancer](https://github.com/stephane-caron/proxqp_balancer): prototype for this agent used in the code for the [ProxQP paper](https://inria.hal.science/hal-04198663v2). Currently supports more QP solvers.