Open-source software for cyber-physical systems (CPS) often lacks robust testing involving robotic platforms, resulting in critical er- rors that remain undetected. This is especially challenging when multiple modules of CPS software are developed by various open- source contributors. To address this gap, we propose Automated CPS Testing (ACT) that performs automated, continuous testing of open-source software with its robotic platforms, integrated with the open-source infrastructure such as GitHub.
This figure represents the illustrates the proposed approach, for a testing framework for CPS with physical robotic hardware.
The current iteration of ACT is targetted towards the Pololu 3pi+ 2040 Robot which runs Lingua Franca programs. This robot is currently used in the embedded systems lab of the Lingua Franca ecosystem. We have testing methods for the
- On-board LED
- Motors
- Bump sensor and display
- IMU and display
The testing method, clones the repository into the self-hosted runner, where it installs the pre-requisite packages. The next step is compiling the lf files, flashing the binary to the robot, monitor the behavior of the robot and then upload the test data (graphs, csv) to this repo as artifacts.
The script "env_setup.py" needs to be run manually since many of the commands require user permission. The next step would be to host the runner in your local machine and it will await the trigger to start testing.
See docs/writing-tests.md for how the label-gated
workflow suite is put together, and step-by-step instructions for adding a
new test variant, a new hardware module, or wiring up a new repo for
cross-repo testing (act-lf-led, act-lf-motor, act-lf-imu, act-lf-bump
labels and friends).
- Adding test cases for Bump sensor and display, IMU paired with display.
- Test cases where hardware is offline. Corner cases where certain tests fail or hardware becomes unavailable.