Distributed Estimation and Active Sensing w/ Mobile Robot Networks
- Faraz Mirzaei, Anastasios Mourikis, Nikolas Trawny, Ke Zhou, Xun (Sam) Zhou, and Stergios Roumeliotis (PI)
This research effort will focus on developing detailed models that quantify the effect on the accuracy of distributed estimation tasks of important factors such as the size of the robot team, type and precision of sensors, frequency of observations, and availability of communication and processing resources. Examples of cases that will be considered are cooperative localization, mapping, tracking, and detection with robots navigating in 3D under realistic sensing, communication, and processing limitations. This analysis will culminate in a concrete set of rules and methods that will be used to direct the design of robotic groups capable of achieving their mission objectives as described by user-imposed requirements on the expected precision and time to complete their task. The direct impact of this work will be significant cost savings during the design phase of a robotic team and throughout the robots' operation. Furthermore, this performance evaluation effort will empower robotics engineers with the ability to extrapolate from current design paradigms and reason for their selections based on formal analysis. The analytical results from this work will also be used as the basis for determining the optimal motion and communication strategies that maximize the efficiency of distributed sensing and estimation tasks. These algorithms will advance the state of the art in robot coordination for information acquisition, communication, and management by providing adaptability to changing conditions and increasing the reliability of mobile robot networks.
- Performance Characterization of Cooperative Localization and SLAM
- Performance Characterization of Cooperative Localization and Tracking
- Multi-robot SLAM with Unknown Initial Correspondence
- Determining the Robot-to-Robot Relative Pose using Range only Measurements
- Optimal Sensing Strategies for Mobile Robot Formations
- Optimal Motion Strategies for Distributed Range only Tracking
- Adaptive Sensing for Plume Tracking