Cannon

This project is a ROS implementation of the cannonball simulation provided by NASA Johnson Space Center as part of the tutorial for the Trick Simulation Environment.

It determines the trajectory and time of impact of a cannon ball that is fired with an initial speed and initial angle, assuming a constant acceleration of gravity (g), and no aerodynamic forces.

Two versions of the simulation are provided: an analytic solution and a numeric integration solution.

Prerequisites

• Visual Studio code
• Docker
• Foxglove (optional)

Installation

    git clone [email protected]:citros-garden/cannon.git
    cd ~/cannon
    code .

and open the repository inside a container using VScode's reopen in container option.

Build

    colcon build
    source install/local_setup.bash

Run the analytic solution

    ros2 launch scheduler cannon_analytic.launch.py

Run the numeric integration solution

    ros2 launch scheduler cannon_numeric.launch.py

Running either of the two simulations will result in the logger output being written to the console.

Implementation Overview

The project is made out of three ROS nodes - cannon_analytic, cannon_numeric and scheduler. The scheduler node is responsible for driving the simulation by publishing a scheduler topic at a given rate (say, 100Hz). The cannon nodes subscribe to this topic, and upon receiving it perform a single calculation step. The rate (dt) is a ROS parameter for the scheduler node, which means you may change its value in the config/params.yaml file, without the need to recompile. The two cannon nodes also have params.yaml files of their own, in which you can set the initial speed and angle, and also the time/integration delta (dt).

Additionally, the scheduler node subscribes to a debug topic, which, together with the provided Foxglove layout, facilitates a play/pause/step/resume functionality.

The output of the simulation, i.e. the topic containing the calculated results, is called cannon/state (in both analytic and numeric versions). It is a list of float of size 4, the layout being:

    [position_x, position_y, velocity_x, velocity_y]

The simulation will halt when position_y reaches zero (i.e. impact).

Foxglove

To view a graphical representation of the simulation, you can open Foxglove and load the CITROS_Cannon.json layout file, or create your own layout.

It is recommended to start the simulation in a paused state, and then, once your foxglove layout is ready, resume it via the Play/Pause button.

To do that, in the __init__ member function of the scheduler node (in scheduler.py), change the line

    self.debug_mode = False

to

    self.debug_mode = True

You will need to build (and source) again.

Output example:

run

docker run -it -e CITROS_ENVIRONMENT=CLUSTER -e JOB_COMPLETION_INDEX=0 -e CITROS_DOMAIN=dev1.citros.io --mount type=bind,source="$(pwd)"/.citros/auth,target=/root/.citros/auth,readonly cannon citros run -b a519b502-e275-4ccc-90d7-9a2f6163c8d8 -i JOB_COMPLETION_INDEX -n "name" -m "message" -v -d

Production Docker

docker build -t cannon .

# *** when building from MAC M1 chip add FROM --platform=linux/amd64 ***
docker buildx build --platform linux/amd64 -t cannon .   


citros docker-login

# dev
docker tag cannon registry.local:32000/citros/lulav/cannon
docker push registry.local:32000/citros/lulav/cannon

# production
docker tag cannon us-central1-docker.pkg.dev/citros/lulav/cannon
docker push us-central1-docker.pkg.dev/citros/lulav/cannon