HEROES: Control design for optimal estimation using heterogeneous sensors
Controlul senzorilor mobili eterogeni pentru estimare optima
Grant of the Romanian National Authority for Scientific Research, CNCS - UEFISCDI, project number PN-III-P1-1.1-TE-2019-1956
Summary and objectives
Today's sensors are capable of not only sensing, but also communicating, self-scheduling, data processing, and moving from one place to another. An important type of such moving sensors are multi-robot teams, in which a group of robotic agents work cooperatively e.g. to position themselves such that the surveillance of a domain of interest is maximized. Such robot sensing teams are used in diverse domains: precision agriculture, environmental monitoring, search-and-rescue operations, etc. However, most existing results focus on identical agents that have significantly simplified dynamics, whereas in fact the dynamics of mobile robots contain complex nonlinearities. Moreover, the domain of interest is usually represented via a static density function, which does not provide enough flexibility.
In this project we therefore develop methods and algorithms for obtaining optimal coverage of a dynamically changing area of interest using heterogeneous mobile robots. We start with a single-robot scenario, when the robot is equipped with various sensors and should be optimally positioned to perform a given task. The results are extended first to a multi-robot team that must optimally cover or track an area with a known map, and finally for the case when the map has to be estimated. We will build on and further develop state-of-the art methods for nonlinear control and estimation. To efficiently address the nonlinear dynamics and at the same time keep it in a natural form, Takagi-Sugeno models in descriptor form will be used. The results will facilitate a more efficient, cost-effective, dependable and reliable operation of mobile robotic systems and significantly contribute to the reduction of the socio-economical costs currently involved.
The main objective of this project is to develop methods and algorithms for obtaining optimal coverage of a dynamically changing area of interest using heterogeneous robots. We start with the positioning problem of a single robot equipped with limited sensors, then extend it to multiple robots, and finally, we consider the general problem of optimal coverage of an unknown, dynamically changing map. The final goal is to test and validate the methods in complex simulation scenarios.
First we will developed methods for optimal estimation for control, such that we take into account the uncertainty model of the sensors and are able to the estimate the states of a mobile target. Second, distributed algorithms will be derived. Finally, we develop methods for tracking points of interest whose dynamics are estimated based on the current measurements. We will use a TS descriptor representation of the dynamic models and develop design methods applicable for such systems. The methods and algorithms will be tested and validated in numerical examples and simulated scenarios.
Keywords: Takagi-Sugeno fuzzy models, estimation and control, distributed systems.
The publications resulting from this research can be found here.
ContactComments, suggestions, and questions concerning the research or the website are welcome. Please contact preferably Zsófia Lendek, or otherwise any of the other members of the project (see their respective websites for contact information).