Decision-making Theory on Intelligent Robot Sensor Networks
Smarter Unmanned Vehicle Technology: The Foundation of Basic Theories Spurring Advancement in Applied Technology
“Unmanned vehicle technology is getting smarter every day. The self-driving car has become a reality and the auto-flight of drones has become a hot topic among global citizens, heralding the era of ‘Dream express’. However, fundamental basic research should be conducted in order to facilitate the development of the applied technologies that utilize unmanned vehicles as a sensor platform. Professor Choi, Han-Lim and his team vigorously conducted a basic theoretical study seeking more accurate prediction of the ‘Large-scale environmental system.' The team established ‘Decision-making theory on intelligent robot sensor networks’, laying the groundwork for the development of applied technology in the future.”
Multiple unmanned robots with sensors predicting weather conditions and other large-scale environment systems
The era of the ‘human robot’, an advanced version of existing unmanned vehicle technology, is just around the corner. The technological theory for unmanned aerial robots that could measure environmental changes with their sensors was established by Prof. Choi’s team. This theory can be applied to the development of a wide array of applied technologies. These robot sensor platforms could measure physical entities of interest such as pressure, temperature, vorticity, and species concentration beyond the reach of current sensing systems. It can also help us handle environmental accidents such as oil spills in the ocean.
Prof. Choi at KIR established the ‘Decision-making theory on intelligent robot sensor networks’, by expanding the study on the network routes of multiple unmanned robot sensors, a theme that Prof. Choi was interested in when writing his doctoral thesis at MIT. This theory is about how effectively improve the prediction accuracy of current environmental prediction systems (e.g., numerical weather prediction, regional torrential rains, or yellow dust in the spring) that typically rely on the satellites. The problem is that the weather variable information changing in a vertical direction can be hard to obtain since the satellite data is information obtained from high altitude. Furthermore, the computing speed of the super computer is not fast enough to handle uncertain and subtle phenomena in a precise manner. This technological gap has been filled with information obtained from weather radars on the ground or the Radiosonde, i.e., a weather balloon. Yet, those methods can only be used on the ground, not in the ocean. In this regard, the advancement in unmanned vehicle technology can give us a clue to resolve such limitations.
Prof. Choi and his team established a basic theory for obtaining the maximum amount of information on the system throughout the sensor platform routes by optimizing the moving routes of the sensor platform, as well as a theory for required algorithm development. In a paper published in Automatica, the team formalized the problem of seeking the measurement path in which the mutual information between the measured values and the interested variables can be maximized when the changes in environmental variables by time can be drawn in a linear system, and suggested a solution to that problem as well. In a paper published in IEEE Transactions on Control Systems Technology, a method in which the optimal routes are calculated under the distributed surroundings with multiple sensor platforms employed was suggested based on game theory.
This study was conducted in line with the project, ‘Fluid SLAM and Robotic Reconstruction of Local Atmospheric Phenomena’ that the team jointly conducted with Dr. Sai Ravela at the Department of Earth, Atmosphere, and Planetary Sciences, MIT with the support of the Air Force Office of Scientific Research (AFOSR) in the United States. The theoretical basis that Prof. Choi and his team laid out is expected to speed up the advancement of the relevant applied technologies, including computational engineering for higher accuracy of the model and technologies for increased autonomy of the unmanned vehicles or more effective sensors.
Prof. Choi, Han-Lim
2015 Annual Report