Electrical and Computer Engineering ETDs

Publication Date

6-26-2015

Abstract

Manned vehicles are maturing into robotics agents under shared control. Vehicle op- erators will be confronted with understanding the expected response from the robotic agent to their actions. I propose a warning system framework to continuously track human inputs, identify possible conflicts and provide contextual warning information to the operator to help avoid accidents. I consider a robotic agent which nominal operation can be encoded using the hybrid automata framework with human inputs. Trajectory prediction methods are used to identify possible conflicts, coded as the avoid set of the system. Using Dynamic Information Flow Tracking (DIFT), human inputs and their effect over time are accumulated and classified in a continuous range between spurious or legitimate inputs.

Keywords

DIFT, hybrid, control, warning

Sponsors

NSF grant CNS 1017602

Document Type

Thesis

Language

English

Degree Name

Electrical Engineering

Level of Degree

Masters

Department Name

Electrical and Computer Engineering

First Advisor

Fierro, Rafael

First Committee Member (Chair)

Caudell, Thomas

Second Committee Member

Naseri, Asal

Share

COinS