MURI: Nonparametric Bayesian Models to Represent Knowledge Uncertainty for Decentralized Planning

ONR_logo.jpg

N000141110688

ONR_logo_small.jpg

September 1, 2010 – August 31, 2015

$7M

ONR

Topic: Knowledge Representation
and Reasoning for Decentralized Autonomy

PI:

J. How (MIT),

Co-PI:

S. Ferrari, L. Carin (Duke),
J. Leonard (MIT), N. Roy (MIT), T. Darrell (UCB),
M. Jordan (UCB), M. Wainright (UCB),
J. Fisher (MIT), A. Willsky (MIT)

Project Description

This research enables radically new capabilities to deploy intelligent decentralized knowledge learning and planning algorithms for teams of heterogeneous autonomous static and mobile agents. The research plan is based on the key insight that Bayesian nonparametric models (BNPM) provide a powerful framework for reasoning about objects and relations in settings in which these objects and relations are not predefined. This feature is particularly attractive for missions such as long term persistent surveillance for which it is virtually impossible to specify the size of the model and the number of variables a priori.

Research Goals

  • Decentralized inference and model learning using Bayesian nonparametric models
  • Decentralized sensor planning and control under uncertainty
  • Information sharing and consensus under uncertainty

Peer-Reviewed Publications

  1. H. Wei, P. Zhu, M. Liu, J. How, S. Ferrari, “Automatic Pan-tilt Camera Control for Learning Dirichlet Process Gaussian Process (DPGP) Mixture Models of Multiple Moving Targets,” IEEE Transactions on Automatic Control, Vol. 64, No. 1, pp. 159 – 173, 2019. [PDF]
  2. H. Wei, W. Lu, P. Zhu, S. Ferrari, M. Liu, R. H. Klein, S. Omidshafiei, J. How, “Information value in nonparametric Dirichlet-process Gaussian-process (DPGP) mixture models,” Automatica, Vol. 74, pp. 360-368, 2016. [PDF]
  3. H. Wei and S. Ferrari, “A geometric transversals approach to sensor motion planning for tracking maneuvering targets,” IEEE Transactions on Automatic Control, Vol. 60, No. 10, pp. 2773-2778, 2015. [PDF]
  4. H. Wei, W. Lu, P. Zhu, S. Ferrari, R. H. Klein, S. Omidshafiei, and J. How, “Camera control for learning nonlinear target dynamics via bayesian non-parametric dirichlet-process gaussian-process (dp-gp) models,” IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Chicago, IL, 2014. [PDF]
  5. H. Wei, W. Lu, P. Zhu, G. Huang, J. Leonard, and S. Ferrari, “Optimized visibility motion planning for target tracking and localization,” IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Chicago, IL, 2014. [PDF]
  6. W. Lu, G. Zhang, and S. Ferrari, “An information potential approach to integrated sensor path planning and control,” IEEE Transactions on Robotics, Vol. 30, No. 4, pp. 919-934, 2014. [PDF]
  7. A.C. Bellini, W. Lu, R. Naldi, and S. Ferrari, “Information driven path planning and control for collaborative aerial robotic sensors using artificial potential functions,” American Control Conference (ACC), Portland, OR, 2014. [PDF]
  8. H. Wei and S. Ferrari, “A geometric transversals approach to analyzing the probability of track detection for maneuvering targets,” IEEE Transactions on Computers, Vol. 63, No. 11, pp. 2633-2646, 2013. [PDF]
  9. W. Lu and S. Ferrari, “An approximate dynamic programming approach for model-free control of switched systems,” IEEE 52nd Annual Conference on Decision and Control (CDC), Florence, Italy, 2013, pp. 3837–3844. [PDF]
  10. P. Cruz, R. Fierro, W. Lu, and S. Ferrari, “Maintaining robust connectivity in heterogeneous robotic networks,” in SPIE Defense, Security, and Sensing, pp. 87410N–87410N, International Society for Optics and Photonics, 2013. [PDF]
  11. G. Foderaro, S. Ferrari, and M. Zavlanos, “A decentralized kernel density estimation approach to distributed robot path planning,” Proceedings of Neural Information Processing Systems Conference (NIPS), Lake Tahoe, NV, 2012. [PDF]
  12. H. Wei, W. Lu, and S. Ferrari, “An information value function for nonparametric gaussian processes,” Proceedings of Neural Information Processing Systems Conference (NIPS), Lake Tahoe, NV, 2012. [PDF]

Presentations

  1. “Sensor Planning for Multiple Targets Tracking,” Cornell Robotics In Society Seminar, Cornell, NY, December 2015. [PDF]
  2. “Optimized Visibility Motion Planning for Target Tracking and Localization,” IROS, Chicago, IL, September 2014. [PDF]
  3. “Gaussian Processes Performance Bounds for Decentralized Control with Intermittent Communications,” ONR MURI Review Meeting, Boston, MA, September 2014. [PDF]
  4. “Information Driven Path Planning and Control for Collaborative Aerial Robotic Sensor Using Artificial Potential Functions,” ACC, Portland, OR, June 2014. [PDF]
  5. “Decentralized Stochastic Planning for Nonparametric Bayesian Models”, ONR MURI Review Meeting, Boston, MA, January 2014. [PDF]
  6. “An Approximate Dynamic Programming Approach for Model-free Control of Switched Systems,” CDC, Florence, Italy, December 2013. [PDF]
  7. “Decentralized Stochastic Planning via Approximate Dynamic Programming,” ONR MURI Review Meeting, Boston, MA, October 2012. [PDF]
  8. “An Information Value Function For Nonparametric Gaussian Processes,” ONR MURI Review Meeting, Boston, MA, April 2012. [PDF]