Requirements
- Navigate autonomously indoors: localize within the working environment, and detect and avoid static and dynamic obstacles.
- Build a map of a new environment on the first run, then localize in it. An RPLidar interfaced with an NVIDIA Jetson Nano runs the ROS Gmapping package, which is based on the FastSLAM algorithm.
- Control the speed of the robot with PID controllers and wheel encoders on the low-level controller (STM32).
- Recognize the customer's emotions with computer vision.
System design


Testing the subsystems

Simulation
After the course, the team planned to develop Robusta further for a coffee shop in Egypt.

