CPS and IoT interact with the physical world by tightly integrating sensing, actuation, computation, and physical objects. As a key enabling technology for many critical domains such as environment, smart cities, transportation, and smart grids, CPS and IoT have been recognized as a top national research priority in several US presidential reports. Many CPS and IoT systems must process large volume of data from dynamic environments under tight energy and real-time constraints. We have developed new systems and computing paradigms for such data-intensive CPS/IoT applications. My group has developed new systems and computing paradigms for such data-intensive CPS applications.
Real-time 4D volcano tomography. We developed new wireless sensor systems for real-time and long-lived volcano monitoring, and deployed 20+ nodes on two live volcanoes in Ecuador and Chile. These systems can analyze seismic signals and compute real-time, full-scale three-dimensional tomography of the volcano conduit system within the network in a distributed manner.
Robotic sensor systems for aquatic monitoring. We developed an aquatic robot system that integrates an off-the-shelf smartphone and a gliding robotic fish for debris monitoring, fish tracking, and harmful algae blooms (HABs) detection. We also developed new aquatic diffusion process profiling and spatiotemporal aquatic field reconstruction algorithms.
Data center thermal/energy management. We developed a novel cyber-physical approach for thermal management in data centers, which integrate enables more efficient thermal/power management. It predicts the temperature evolution of a data center in real time, and then finds the temperature setpoints, cold air supply rates, and the speeds of server internal fans to minimize the expected total energy consumption of cooling and circulation system.
Networked Vehicular Systems: Driver Sensing and Wireless Networking. We propose a novel cross-layer wireless design approach that leverages signatures of upper-layer vehicular protocols to improve PHY and link layer performance. We also develop a system based on smartwatches and smartphones, which detects the driver's unsafe driving behaviors in a real-time manner.
CPS platforms and performance control frameworks. We develop a smartphone-based platform for data-intensive cyber-physical applications, which has been adopted in several large-scale volcano-monitoring sensor systems and aquatic robotic systems. We also proposed a unified sensing fidelity and real-time performance control framework for cyber-physical systems whose operations are subject to uncertainties from physical environments.