In this project, we investigate methods for giving intelligent robots an understanding and a sense of physics about their environment. By this, the robots could obtain plausible 3D scene reconstructions and predict the outcome of interactions with the environment. Our research in this project in 2022-2024 expands upon our preliminary works in~[ ] which learn models of video using physics-based priors and DiffSDFSim~[ ] which proposes differentiable physics simulation including time of contact differentials to optimize shapes from collisions.
todo.