Hierarchical Task Taxonomy

• Hierarchical task taxonomy is a structured framework that decomposes complex tasks into multi-level goals, subgoals, and primitive actions.
• It employs directed acyclic graphs or trees to model explicit parent-child relationships, ensuring modular control and efficient reasoning.
• Commonly applied in robotics, AI, HCI, and transfer learning, it facilitates practical applications like automated planning and dynamic task execution.

A hierarchical task taxonomy is a structured framework that organizes complex tasks into multi-level representations, systematically decomposing high-level goals into intermediate subgoals and finally into primitive actions or atomic operations. Such taxonomies are foundational in robotics, artificial intelligence, human-computer interaction, transfer learning, and multi-label classification, as they enable modular control, efficient reasoning, and the application of domain knowledge through explicit parent–child relationships and semantic structures.

1. Formal Structures and Levels in Hierarchical Task Taxonomies

Hierarchical task taxonomies typically instantiate a directed acyclic graph (DAG) or tree, where nodes represent goals, subgoals, or actions, and edges codify logical, temporal, or semantic dependencies. In cognitive robotics, such as in "Cognitive Approach to Hierarchical Task Selection for Human-Robot Interaction in Dynamic Environments," the taxonomy comprises:

• Level 1: SkillNodes corresponding to high-level intentions (e.g., "TeaMaking", "SandwichMaking").
• Level 2: Control nodes such as THEN (imposing sequential execution), AND (mandating all children to execute, order-agn