Idham Ananta Timur, S.T., M. Kom
Department of Computer Science and Electronics,
Faculty of Mathematics and Natural Science,
Universitas Gadjah Mada, Indonesia.
He is an active member of Computer System and Network Research Labs. His main research interest is Intelligent Environments which refer to physical spaces in which pervasive computing technology are woven and used to achieve specific goals for the user, the environment or both. At the moment, he and his team are working on the idea of how autonomous robots collaboratively interact within an intelligent environment.
Tutorial Title: Toward Autonomous Robot for Disaster Mitigation Support using Deep Reinforcement Learning
Disaster usually leads to uncertainty. Natural disaster, like earthquake, tsunami, landslides, volcanoes, floods usually change the landscape on earth. The rescue missions afterward, depending on the destruction level, will involve searching for survivors, deploying logistics and mapping the area to for a better mitigation management. However, doing all those actions might need more expensive vehicles, gasolines, and manpower. In this area, a smart autonomous robot might help.
However, developing a smart autonomous robot is also a challenging problem. Training an autonomous car to follow the lane in a known and well-defined road is already difficult let alone exploring unexpected, unseen aftermath of disaster landscape with so many kinds of unidentified objects. This tutorial will introduce the ideas and demonstrate how an end-to-end deep learning and deep reinforcement learning might be able to help developing autonomous vehicle and discuss the issues surrounding such approaches to tackle the aforementioned challenge.