Distributed Hybrid Information and Plan Consensus HIPC for Semi-autonomous UAV Teams
Abstract
The research objective of this project was to increase the capabilities of decentralized task allocation algorithms. Progress has been made on several different fronts, including: a) chance-constrained task allocation, b) task allocation with tasks defined as Markov Decision Processes, c) guaranteeing network connectivity during mission execution, d) allowing the use of non-submodular score functions during the decentralized allocation, and e) decreasing the convergence time by utilizing all of the information available in the network, f) hardware results that demonstrate the difficulty of planning in communication contested environments and the utility of using LICA algorithms, and g) a tutorial on the basics of decentralized task allocation for a general audience is currently in revision for Control Systems Magazine. Combining these results has significantly improved the state of the art capabilities of decentralized task allocation and work continues to refine these approaches.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 18, 2015
- Accession Number
- ADA627417
Entities
People
- Jonathan How
Organizations
- Massachusetts Institute of Technology