Autonomy Talks - Takashi Tanaka: Minimum-Information Kalman-Bucy Filtering
Autonomy Talks - 28/06/2022
Speaker: Prof. Takashi Tanaka, University of Texas at Austin
Title: Minimum-Information Kalman-Bucy Filtering and Fundamental Limitation of Continuous-Time Data Compression
Abstract: Motivated by a practical scenario where a continuous-time source signal is encoded, compressed, and transmitted to a remote user where the signal is reproduced in a real-time manner (e.g., streaming of neuromorphic camera data), we study the fundamental trade-off between the encoding data rate and the best achievable data quality (distortion). After briefly reviewing the “causal” rate-distortion theory in discrete-time, in this talk, we consider the problem of estimating a continuous-time Gauss-Markov source process observed through a vector Gaussian channel with an adjustable channel gain matrix. For a given (generally time-varying) channel gain matrix, we provide formulas to compute (i) the mean-square estimation error attainable by the classical Kalman-Bucy filter, and (ii) the mutual information between the source process and its Kalman-Bucy estimate. We then formulate a novel “optimal channel gain control problem” where the objective is to control the channel gain matrix strategically to minimize the weighted sum of these two performance metrics. To develop insights into the optimal solution, we first consider the problem of controlling a time-varying channel gain over a finite time interval. A necessary optimality condition is derived based on Pontryagin's minimum principle. For a scalar system, we show that the optimal channel gain is a piece-wise constant signal with at most two discontinuities. We also consider the problem of designing the optimal time-invariant gain to minimize the average cost over an infinite time horizon. A novel semidefinite programming (SDP) heuristic is proposed to compute the optimal solution.
If you want to know more, check out our official webpage: https://idsc.ethz.ch/research-frazzoli/autonomy-talks.html
Видео Autonomy Talks - Takashi Tanaka: Minimum-Information Kalman-Bucy Filtering канала Autonomy Talks
Speaker: Prof. Takashi Tanaka, University of Texas at Austin
Title: Minimum-Information Kalman-Bucy Filtering and Fundamental Limitation of Continuous-Time Data Compression
Abstract: Motivated by a practical scenario where a continuous-time source signal is encoded, compressed, and transmitted to a remote user where the signal is reproduced in a real-time manner (e.g., streaming of neuromorphic camera data), we study the fundamental trade-off between the encoding data rate and the best achievable data quality (distortion). After briefly reviewing the “causal” rate-distortion theory in discrete-time, in this talk, we consider the problem of estimating a continuous-time Gauss-Markov source process observed through a vector Gaussian channel with an adjustable channel gain matrix. For a given (generally time-varying) channel gain matrix, we provide formulas to compute (i) the mean-square estimation error attainable by the classical Kalman-Bucy filter, and (ii) the mutual information between the source process and its Kalman-Bucy estimate. We then formulate a novel “optimal channel gain control problem” where the objective is to control the channel gain matrix strategically to minimize the weighted sum of these two performance metrics. To develop insights into the optimal solution, we first consider the problem of controlling a time-varying channel gain over a finite time interval. A necessary optimality condition is derived based on Pontryagin's minimum principle. For a scalar system, we show that the optimal channel gain is a piece-wise constant signal with at most two discontinuities. We also consider the problem of designing the optimal time-invariant gain to minimize the average cost over an infinite time horizon. A novel semidefinite programming (SDP) heuristic is proposed to compute the optimal solution.
If you want to know more, check out our official webpage: https://idsc.ethz.ch/research-frazzoli/autonomy-talks.html
Видео Autonomy Talks - Takashi Tanaka: Minimum-Information Kalman-Bucy Filtering канала Autonomy Talks
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![Autonomy Talks - David Fridovich-Keil: Learning Mixed Strategies in Lifted Trajectory Games](https://i.ytimg.com/vi/27kXwewzIlw/default.jpg)
![Autonomy Talks - Kaidi Yang: Operation of AV-Enabled Future Mobility Systems](https://i.ytimg.com/vi/4r5OB2PiFdo/default.jpg)
![Autonomy Talks - Kevin D. Smith: Semicontracting Dynamical Systems](https://i.ytimg.com/vi/MfqIK3tomXk/default.jpg)
![Autonomy Talks - Neda Masoud: Cybersecurity in Connected and Automated Vehicle Systems](https://i.ytimg.com/vi/y7qd4jmiJ-k/default.jpg)
![Autonom Talks - Arash Ajoudani: Learning and control for interactive robots](https://i.ytimg.com/vi/jm6r_QEcxNQ/default.jpg)
![Autonomy Talks - Erdem Bıyık: Learning how to Route Autonomous Vehicles on Shared Roads](https://i.ytimg.com/vi/5S6hI82jfM8/default.jpg)
![Autonomy Talks - Stephanie Gil: Resilient Coordination in Networked Multi-Robot Teams](https://i.ytimg.com/vi/sJ8SSxzVMlg/default.jpg)
![Autonomy Talks - Nicolò Pagan: Game Theoretical Inference of Human Behavior in Social Networks](https://i.ytimg.com/vi/JwPvpV2WF3o/default.jpg)
![Autonomy Talks - Sanjiban Choudhury: The Virtues of Laziness in Model-based RL](https://i.ytimg.com/vi/qygIp9B7A44/default.jpg)
![Autonomy Talks - Nicolas Lanzetti: Dynamic Programming in Probability Spaces via Optimal Transport](https://i.ytimg.com/vi/FXvsDCt4pjQ/default.jpg)
![Autonomy Talks - Abolfazl Lavaei: Automated Verification and Control of Large-Scale Stochastic CPSs](https://i.ytimg.com/vi/6QLewaMHi4Y/default.jpg)
![Autonomy Talks - Spring Berman: Scalable Control of Robotic Swarms in Uncertain Environments](https://i.ytimg.com/vi/xpx34vAWLKY/default.jpg)
![Autonomy Talks - Laura Ferranti: An Optimization-based Framework for Local Robot Motion Planning](https://i.ytimg.com/vi/wa1KXguAu1M/default.jpg)
![Autonomy Talks - Yingzhao Lian: New Perspectives on Advanced Building Control](https://i.ytimg.com/vi/d6VAj9-w85A/default.jpg)
![Autonomy Talks - Jonathan Gammell: Measuring and Planning Motion in Robotics](https://i.ytimg.com/vi/w4MXkUwIqAA/default.jpg)
![Autonomy Talks - Stelian Coros: Physics-based Modeling and the Quest for Intelligent Robots](https://i.ytimg.com/vi/Gn98nibsahI/default.jpg)
![Autonomy Talks - Stephen James: Sample-Efficient Robot Learning](https://i.ytimg.com/vi/hi9QdFZVb3I/default.jpg)
![Autonomy Talks - Tichakorn Wongpiromsarn: Formal Methods for Control Synthesis of Autonomous Systems](https://i.ytimg.com/vi/duRSRJIvN0U/default.jpg)
![Autonomy Talks - Margarita Grinvald: Dense Object-Level Mapping](https://i.ytimg.com/vi/KAU4ELnxHKY/default.jpg)
![Autonomy Talks - Wilko Schwarting: Learning & Control for Interactions in Mixed Environments](https://i.ytimg.com/vi/GLoEe8A8AG0/default.jpg)