This paper presents a systematic formulation of the self model for embodied artificial intelligence (AI), aiming to provide the missing internal representation that enables an agent to understand its own body, capabilities, memories, and decision processes. Unlike existing approaches that address isolated aspects such as perception, prediction, or skill adaptation, we propose a unified computational framework that integrates body schema, forward and inverse models, perceptual memory mechanisms, and agency. This framework captures how an embodied agent represents its physical structure, predicts the consequences of its actions, selects policies, and accumulates experiences to form a coherent sense of self. We further introduce a six-level hierarchy (L0–L5) that characterizes the developmental stages of self model from non-self representation to full self-awareness, providing the first operational taxonomy for evaluating self-awareness in embodied AI systems. Then, a practical implementation is proposed and validated on embodied navigation and manipulation tasks. Experimental results demonstrate the effectiveness of the self model and its components, including self-perception, self-memory, self-prediction, and self-decision. Overall, this work proposes the concept and instantiation of the self model in embodied AI and discusses future directions of the self model. Related video demonstrations can be accessed from https://taoshida11.github.io/Selfmodel/.

Object navigation, whose goal is to let the agent to reach some places (or objects), has been a popular topic in embodied Artificial Intelligence (AI) researches. However, in our real-world applications, it is more practical to find the targets with particular goals, raising the new requirements of finding the places to achieve the particular functions. In this paper, we define a new task of affordance navigation, whose goal is to find possible places to accomplish the required functions, achieving some particular effects. We first introduce a new dataset for affordance navigation, collected by the proposed affordance algorithm. In order to avoid the high cost of labor, the groundtruth of each episode which is annotated with the interaction data provided by the AI2-THOR simulator. In addition, we also propose an affordance navigation framework, where an Object-to-Manipulation Graph (OMG) is constructed and optimized to emphasize the corresponding nodes (including object nodes and manipulation nodes). Finally, a navigation policy is implemented (trained by reinforcement learning) to guide the navigation to the target places. Experimental results on AI2-THOR simulator illustrate the effectiveness of the proposed approach, which achieves significant gains of 14.0% and 11.7% (on success rate and Success weighted by Path Length (SPL), respectively) over the baseline model.