The AI-Robot Challenge, a competition to assemble furniture using AI and robots, was held at COEX in Seoul, Korea in 2021. In this competition, the technology to recognize IKEA's Stefan chair using AI to create an assembly sequence for a robot to follow was dealt with. Our team used Faster-region-based convolutional neural networks (Faster-RCNN) and multi-object rectified attention network (MORAN) technologies for the recognition of assembly instructions and generated the assembly sequence by interpreting the instructions with a connection relationship tree. After the sequence has been generated, three cooperative robots equipped with multi-finger grippers and a transformable station have been utilized to execute the assembly task with novel postural control. As a result, among the 4 participating teams, we were able to win the competition together with the Korea Institute of Industrial Technology.