Obstacle detection for mobile robots using TensofFlow Keras and micro-controller

Just to share a project where I use TensorFlow Keras to create and train a convolutional neural network for analyzing gray scale video images, and estimate the position of objects on the floor. This model is then ported to an ARM Cortex M7 micro-controller board with a low-resolution camera. The system can be used in small mobile robots as detailed in the blog below. [Note: the controller I am using, Microchips SAMS70J20 is not supported by TFLite, so I need to create my own workflow)