IIIT Hyderabad Publications
A Simple Yet Effective Failure Resilient Robotic Soccer Strategy
Author: Sabitha Sarvasiddhi
Report no: IIIT/TH/2016/27
Robots tend to fail very frequently in games like Soccer. Due to either collision with other players or objects, or extreme body actions or accident, or wear and tear of robot parts, the robot might either become unstable and even fail in the worst case. Also in soccer, fouls happen and players are sent off most commonly resulting in the variation of the number of players active in the game. The temporary or permanent unavailability of a robot player due to these situations disturbs the strategy of the team which in turn impacts the outcome of the game since the team strategy plays a key role in taking major decisions like formations of the robots, roles and actions assignment to robots to achieve the final target - win the game. So, it is essential for the game strategy of a soccer team to dynamically handle fail and fouls of the robots during the game. Whenever there are failures of the robots in the team, the remaining robots need to quickly analyze the state of the game, restrategize and come up with a new set of actions for the team. The aim of this thesis is to propose an elementary failure resilient strategy for an 11 robots vs. 11 robots soccer team that works towards scoring the goals considering the failure of the robots. Our strategy is elementary because it is based on the idea of redistributing the work among surviving robots to either defend or attack. This study has two major parts:(1) Investigation of different game strategies that are already implemented in vogue, (2) from the observation of these strategies, design a game strategy for an 11vs11 Robot Soccer Team that can effectively handle the fail and fouls of the robots during the game. Four different strategies each focusing on different perspectives of the game are taken and simulated for observation and analysis. In the failure resilient strategy, decisions on player’s formation, roles and actions are made dynamically considering the state of the robots (active or failed) to achieve the common objective of defending the goal line and scoring a goal. The strategy is designed in such a way that it is not affected by the dynamic change in the number of robots due to failure during the game. In this thesis, mechanisms followed to implement the four basic independent actions of the robot player necessary to play the game are also discussed.
Full thesis: pdf
Centre for Data Engineering
Copyright © 2009 - IIIT Hyderabad. All Rights Reserved.