Movement Primitive Categorization Balancing the Learnability and Adaptability

Given the rapid advancement of robotic technologies, robots will eventually enter our daily lives, performing complex long-horizon tasks. Complex long-horizon tasks, such as furniture assembly, typically contain dozens of subtasks with various scenes. Although complex, assembly is based on several reusable movements. Movement Primitive (MP) is a promising framework for learning reusable movements from demonstrations and adapting the learned movements to the test scenes. The critical step in employing MP methods is categorizing the unlabeled demonstrations into different MPs. However, current MP methods focus on individual MP learning using manually selected demonstrations, neglecting categorization. Manual categorization of demonstrations is easy to fall into suboptimal. If the demonstrations within the same category are too similar, the learned MP cannot be adapted to task scenes with various obstacles. Conversely, a significant distance between demonstrations leads to the MP’s failure in learning. To this end, we propose the following principle for MP categorization: balance the Learnability and Adaptability. Following this principle, we introduce an optimal transportation (OT)-based theoretical framework and a practical solution utilizing an auto-encoder network. We obtain the lower threshold of learnability by OT. Then we increase the adaptability of MP until it reaches the lower threshold of learnability. For complex long-horizon task learning, we propose a balanced MPs-based learning framework that contains four modules, termed BaMPs. BaMPs achieved success rates of 100% and 80%, respectively, in the 12-step and 20-step tasks. Note to Practitioners—This work aims to learn multiple MPs capable of adaptation to complex long-horizon tasks from unlabeled demonstrations. Such tasks frequently consist of numerous subtasks, each distinguished by unique start, via, and target points, alongside specific obstacles, thereby posing considerable adaptability challenges for MPs. A critical step in learning MPs from unlabeled demonstrations is the categorization of these unlabeled demonstrations. However, current MP methods primarily focus on individual MP learning using manually selected demonstrations, which easily leads to suboptimal outcomes. To this end, we propose a principle for MP categorization: balance learnability and adaptability. In line with this principle, we introduce an OT-based theoretical framework and a practical implementation leveraging an autoencoder network. Furthermore, we introduce a balanced MPs-based complex long-horizon tasks learning framework, termed BaMPs. BaMPs comprises four modules: (a) MP categorization; (b) MP learning; (c) MP selection; and (d) MP editing.