Design and implementation of an automated battery management platform

[摘要] This thesis describes the design and the implementation of the hardware platform for automated battery management with battery changing/charging capability for autonomous UAV missions with persistency requirement that extends the mission duration beyond the life of a single UAV battery. The platform is tested through a series of missions lasting at least 3 hours to prove it meets design requirements and to show its feasibility. This thesis also provides a method to modify existing scenarios to proactively plan for the battery maintenance so that the overall system performance is increased. The modifications made to the problem definition increased the state-space significantly, and means of solving a problem of that scale needed to be developed. To address this challenge, this thesis extends a previously developed approach called Incremental Feature Dependency Discovery (iFDD) by allowing to use caches from computer science literature to make conversion from basic features to extended features faster. By doing so, this method significantly reduces the computational complexity.

[发布机构] Massachusetts Institute of Technology