bkolligs / pose-estimation-prototyping

Prototyping algorithms for pose estimation of the MoonRanger Lunar Rover. Flight code was based off these tests.

Pose Estimation Prototyping

This repository exists to prototype the code written for the pose estimation system for the MoonRanger Lunar Rover .

Modules

The main components tested here are as follows:

1. Sun Sensor: Tests for verifying the math that converts alpha and beta angles to a unit vector representation of a sun ray.
2. EKF Simulation: Python bindings for the current implementation of the orientation EKF so that we can test, verify and visualize the results of the math.
3. Translation Estimation: Tests that verify we can use wheel velocities to estimate 3D position for a certain timestep.

EKF Simulation Setup

The main dependencies for this project include a robotics package I wrote that contains helpful visualization tools called PyRobo, NumPy, PyBind11, and Matplotlib.

The easiest way to install all of these dependencies is to use a Conda environment with the following command:

conda env create -f environment.yml

Otherwise you will need to use pip:

pip install pybind11[global] 
pip install git+https://github.com/bkolligs/pyrobo.git@main

Then we can install all of the C++ code by performing the following commands in the top level directory:

pose-estimation-prototyping/

mkdir build
cd build
cmake ..
make
sudo make install

These commands build a shared library and install it in the site-packages directory corresponding to the current python executable, which should be in theproto-pose conda environment if you used conda. Otherwise it will correspond to your python3 executable, so any files should be run with that version of python. For example, python3 ekf_verify.py. This creates a python package called moonranger that can be accessed like this:

import moonranger as m

ekf = m.OrientationEKF()

# get the ekf's current state
print(ekf.get_state())