中文摘要：

对机器人足球球员如何实现复杂任务中的行为学习理论、方法、技术和应用进行评述，指出其存在的局限性，以及在机器人足球领域的学习策略．机器人足球系统作为多智能体系统研究的测试床，许多研究者从不同的侧面对该项技术进行了研究并取得了一定的成果．对机器人足球系统的研究，目前包括足球机器人体系结构、多机器人的协作、动态环境下的推理和行动、传感器数据融合、复杂任务中的行为学习、对手建模等内容。

英文摘要：

In recent years , robotsoccer competition, has been developped quickly as a kind of adversary hightech action. As an important means of multi-agent system research, many researchers have studied it from different perspectives and much progress has been achieved. Today, a lot of researchers focus on the system construction of the robots, Multi-robot cooperation, the inference and the motion Under dynamic environment, Sensor data fusion , behavior learning in complex duty and prediction of opponent＇s model. The paper introduces the robotsoccers＇ learning theory, methods, technology and application from the view of behavior learning in complex duty, and point outs their limit, and indicates the learning strategy in the robotsoccer field.