Self-Sustainable Robotic Environment Discovery for Energy Harvesting Internet of Things

This paper considers autonomous environment discovery for energy harvesting Internet of things (IoT) applications. A self-sustainable mobile micro-robot explores an unknown environment to collect data under energy constraints. The data are fed into a machine-learning model to construct an energy harvesting map of the environment, which will then be used by IoT devices for self-sustainable operations in the same environment. The objective is to develop an efficient robotic exploration algorithm that helps construct as accurate an energy harvesting map as possible. The challenge lies in the unknown environment and the sometimes conflicting goals of data collection and energy management during the robotic exploration. We propose an efficient exploration algorithm in combination with the support vector regression (SVR) model. Simulation validates the effectiveness of the proposed method and shows exemplary traces of robotic exploration to illustrate the core ideas.