Hierarchical Instruction-aware Embodied Visual Tracking

User-Centric Embodied Visual Tracking (UC-EVT) presents a novel challenge for reinforcement learning-based models due to the substantial gap between high-level user instructions and low-level agent actions. While recent advancements in language models (e.g., LLMs, VLMs, VLAs) have improved instruction comprehension, these models face critical limitations in either inference speed (LLMs, VLMs) or generalizability (VLAs) for UC-EVT tasks. To address these challenges, we propose $\mathbf{\text{Hierarchical}}\mathbf{\text{Instruction-aware}}\mathbf{\text{Embodied}}\mathbf{\text{Visual}}\mathbf{\text{Tracking}}\mathbf{\text{(HIEVT)}}$ agent, which bridges instruction comprehension and action generation using $\mathit{\text{spatial}}\mathit{\text{goals}}$ as intermediaries. HIEVT first introduces $\mathit{\text{LLM-based}}\mathit{\text{Semantic-Spatial}}\mathit{\text{Goal}}\mathit{\text{Aligner}}$ to translate diverse human instructions into spatial goals that directly annotate the desired spatial position. Then the $\mathit{\text{RL-based}}\mathit{\text{Adaptive}}\mathit{\text{Goal-Aligned}}\mathit{\text{Policy}}$, a general offline policy, enables the tracker to position the target as specified by the spatial goal. To benchmark UC-EVT tasks, we collect over ten million trajectories for training and evaluate across one seen environment and nine unseen challenging environments. Extensive experiments and real-world deployments demonstrate the robustness and generalizability of HIEVT across diverse environments, varying target dynamics, and complex instruction combinations. The complete project is available at https://sites.google.com/view/hievt.