Robust Barrier Functions with Gaussian Processes

Setting Up MATLAB Nodes

1. To run, you need two MATLAB nodes opened.
2. On each of the two nodes, clone this repo: git clone https://github.com/yemam3/robustCBFs_with_GPs.git
3. One node will run the experiment (robotarium node)
4. Other node will handle fitting the GPs (gp node)

Inter-Node Communication (both nodes)

MQTT

1. Clone MQTT MATLAB interface: git clone https://github.com/gnotomista/mqtt_matlab_interface.git
2. The readme will contain information on how to setup mqtt using a broker (e.g. emqtt, mosquitto)
3. Set the COMM_MODE to MQTT in the init.m file
4. Specify the IP and PORT to be used for the communication (e.g. localhost, 1883)
5. Make sure to add mqtt_matlab_interface to the path (edit init file)

File Sharing (not recommended)

1. If both nodes can share files (e.g. on same computer), we can just communicate using writing/reading files
2. Set the COMM_MODE to FileSharing in the init.m file

Experiment Setting

Running on the Robotarium

1. Set IS_SIM to 0 in the init.m file

Running a Simulation

1. Clone the Robotarium MATLAB simulator: git clone https://github.com/robotarium/robotarium-matlab-simulator.git
2. Set IS_SIM to 1 in the init.m file (this will add fake noise to the data)

To Run Experiment

1. Run main_robotarium.m on the robotarium node (runs robots and collects data)
2. Run main_gpr.m on the gp node (this node receives the data, runs fitrgp and returns models)