MuTT: A Multimodal Trajectory Transformer for Robot Skills
This addresses the challenge of automating robot skill parameter optimization in dynamic environments, offering a versatile solution that reduces the need for real-world executions.
The paper tackles the problem of manually configuring robot skill parameters by proposing MuTT, a multimodal transformer that predicts environment-aware executions, achieving superior performance across three experiments with two skill representations.
High-level robot skills represent an increasingly popular paradigm in robot programming. However, configuring the skills' parameters for a specific task remains a manual and time-consuming endeavor. Existing approaches for learning or optimizing these parameters often require numerous real-world executions or do not work in dynamic environments. To address these challenges, we propose MuTT, a novel encoder-decoder transformer architecture designed to predict environment-aware executions of robot skills by integrating vision, trajectory, and robot skill parameters. Notably, we pioneer the fusion of vision and trajectory, introducing a novel trajectory projection. Furthermore, we illustrate MuTT's efficacy as a predictor when combined with a model-based robot skill optimizer. This approach facilitates the optimization of robot skill parameters for the current environment, without the need for real-world executions during optimization. Designed for compatibility with any representation of robot skills, MuTT demonstrates its versatility across three comprehensive experiments, showcasing superior performance across two different skill representations.