Robot escorting with online adjustment

The same escorting problem seen in a previous video (, but using a neural network based on behavioral psychology presented in Calvo et al. (2003). The robots do not know how to act initially, but learn from their mistakes. They may transmit their knowledge to a neighboring robot during navigation. This is the result of my ongoing Master’s project.

Calvo et al. (2003): Calvo, Rodrigo, and Mauricio Figueiredo. “Reinforcement learning for hierarchical and modular neural network in autonomous robot navigation.” Proceedings of the International Joint Conference on Neural Networks, vol. 2, pp. 1340-1345, 2003.