Robots, originally conceived as substitutes for drudgerous human labor, are artifacts representing a utilitarian worldview. This position has driven the vast majority of current design efforts in the Human-Robot Interaction (HRI) research field. This thesis work explores an alternative position and approach to the design of robots: One that prioritizes playfulness over practical purposes. People are inherently playful. Playfulness matters when engaging in all sorts of activities. However, this factor has seldom been considered in robot design. Extending on the well-established approach of Ludic Design in Human-Computer Interaction (HCI), a Character-Driven Robot Design approach is proposed to explore Ludic HRI. By first developing robot characters to drive the design process, it opens up a new design space for robots with selfness, living for themselves before living for humans. The quality of selfness serves as a mechanism to achieve Ludic Design on robots, drawing the attention away from the external functionality to the existence and enabling the fostering of playfulness. Four Research-through-Design (RtD) studies, SketchBot, NodBot, BubbleBot, and ConeBot, are presented respectively to demonstrate the conceptualization of the proposed design approach. The method of Annotated Portfolios, accompanied by sequential information of annotations, is then adapted to communicate the design knowledge generated through the congregation of the four robotic artifacts. Reflections are made on the cases to discuss the key elements of creating robot characters with selfness. Lessons learned on applying annotated portfolios, adopting Character-Driven Robot Design, and exploring Ludic HRI are also provided for other HRI researchers and designers to build upon. This thesis work shows that by incorporating playfulness from humanity, it allows the space to re-examine the normalized value towards robots. Further, it broadens the scope of which robots can and should exist in the world.