Enhancing Context-Aware Human Motion Prediction for Efficient Robot Handovers

Enhancing Context-A ware Human Motion Prediction for Efficient Robot Handovers Gerard G omez-Izquierdo 1, Javier Laplaza 1, Alberto Sanfeliu 1 and Ana ıs Garrell 1 Abstract -- Accurate human motion prediction (HMP) is critical for seamless human-robot collaboration, particularly in handover tasks that require real-time adaptability. In this work, we enhance human motion forecasting for handover tasks by leveraging siMLPe [1], a lightweight yet powerful architecture, and introducing key improvements. Our approach, named IntentMotion incorporates intention-aware conditioning, task-specific loss functions, and a novel intention classifier, significantly improving motion prediction accuracy while maintaining efficiency. Experimental results demonstrate that our method reduces body loss error by over 50%, achieves 200 faster inference, and requires only 3% of the parameters compared to existing state-of-the-art HMP models. These advancements establish our framework as a highly efficient and scalable solution for real-time human-robot interaction. I. INTRODUCTION Human motion prediction (HMP) plays a crucial role in human-robot collaboration (HRC) by enabling robots to anticipate human movements and respond proactively. This capability is particularly important in handover tasks, where the seamless exchange of objects between humans and robots requires both accuracy and speed. The ability to predict human motion allows robots to preemptively adjust their trajectories, improving efficiency and ensuring safety. In this context, human intention--whether the motion is collaborative or non-collaborative--directly influences the prediction and subsequent robot response.