Definition : Internet of Things (IoT)
The Internet of Things is broad term covering several research aspects related to interconnecting a wide scale of physical (embedded) electronic devices. By exchanging data (typically acquired with sensors) the network is more context-aware as it aggregates knowledge on - and integrates in - the physical world. Areas such as smart grids, domotics and more recently smart cities are concrete realizations of IoT.
IoT realizations require several aspects to be tightly coupled:
- Connectivity: large-scale, long-range, low-latency network technologies are required to physically connect all devices
- Data processing: the collected data needs to be cleaned, analysed and processed. The end goal is to learn from the environment and be able to predict the context of the devices, allowing them to make correct and intelligent decisions.
- Security: the data must be shared in a secure manner. Over the years, the increasing exchange (and centralization) of data has raised serious issues related to data ownership, security and privacy. Rather than giving up on data exchange, these challenges have to be tackled.
The expertise of SUMO Lab in IoT currently spans following research areas:
- Design optimization of (micro)electronic devices for enhanced connectivity. This aspect is tightly coupled with the research on Surrogate Modeling
[Design of Experiments (DOE), Efficient Global Optimization, Multi-objective (Bayesian) Optimization, Connectivity]
- Data collection with Unmanned (Aerial) Vehicles (UAV). Similar to Experimental Design for computer experiments, UAVs can be used to collect data. Because these devices are resource constrained (mostly by battery) intelligent collection strategies and predictive modeling provide answers to problems faster
[Design Of Experiments (DOE), Response Surface Modeling (RSM), Design and Analysis of Computer Experiments (DACE), Metamodeling, Data-driven information processing]
- Cloud modeling services
[Cloud computing, Response Surface Modeling, Multi-objective Optimization]
We aim development of an autopilot which sequentially re-evaluates the flightpath of the drone in order to maximize the information gain obtained by the UAV. This requires sufficient exploration of the search areas, as well as analysis of the obtained sensor data. In addition it also requires a stable connection to a ground station. This project is a cooperation with the MOSAIC Research group and several students, using the Raspberry Pi 2 as flight computer.