The Cornell RoboCup Robot Soccer Team: 1999–2003

摘要

The main difference between the two leagues, contrary to what the names suggest, is the allowed sensing technologies. Teams in the Middle Size league are constrained to use local, on-board vision, while competitors in the Small Size league are permitted to use global vision systems (in addition to local vision systems, if they so desire). As a result, the Middle Size league is dominated by robot localization and perception, while the main challenges in the Small Size league are at the system and integration level. In addition, due to the enhanced situational awareness obtained with global vision, the Small Size league games are much faster and typically involve more team play and collaboration, such as passing. This chapter describes the Small Size league Cornell RoboCup team in the time period starting in 1999, the year it first competed, and ending in 2003, after five years of competition. In this time period, the Cornell team won the competition four times and placed third once. See Table 1 for a summary of the competition outcomes since inception or the RoboCup web site for a more detailed summary. The emphases of this chapter are on the parts of the system that distinguished the Cornell team from its competitors and led to its success in competition.