Cloud-Based Perception and Control of Sensor Nets and Robot Swarms
Technical Report,30 Sep 2013,31 Dec 2015
TRUSTEES OF INDIANA UNIVERSITY BLOOMINGTON United States
Pagination or Media Count:
This project Cloud-Based Perception and Control of Sensor Nets and Robot Swarms was performed by an interdisciplinary team at Indiana University with expertise in cloud and parallel computing, computer vision and robotics. It investigated the use of Cloud Computing as a key technology for Internet of Things IoT and DDDAS applications. The project developed an open source framework called IoT Cloud to connect IoT devices to cloud services and used it to investigated three algorithms controlling robots from the cloud. Our three major applications were a parallel particle filtering based SLAM algorithm a deep learning based drone control algorithm and a robot swarm algorithm for n-body collision avoidance. These applications had significant time complexity needing the additional computer power offered by the cloud and we parallelized applications so that they could respond quickly to the edge devices. This project produced six major papers and a report. It also contributed to the STREAM2015 workshop and its final report. The final report summarizes highlights here of published work and give some comments for follow on activities.