Modular Optimization for Robotic Explorers

摘要

Robotic explorers perform multiple tasks and have multiple constraints to optimize. Planning the best sequence of tasks requires information about how a given sequence will affect the constraints and goals of the explorer. Modules for all constraints and goals are created, each one of which evaluates a given plan from its point of view. The resulting values of the modules are then combined in a weighted sum, producing a score for each plan. The best plan is then selected. Robotic Explorers Robotic explorers are devices which investigate unknown environments. Typically, such robots are designed for places where humans either cannot go, or where sending humans would be too expensive or dangerous, such as space or remote earthly locations. As an extension of the human presence, robotic explorers should provide as much information as possible about the environment being explored, performing a wide variety of activities to act as a human's eyes and hands. Robots have progressed from simple...