Learning and Executing Re-Usable Behaviour Trees From Natural Language Instruction

Domestic and service robots have the potential to transform industries such as health care small-scale manufacturing, well homes in which we live. However, due overwhelming variety of tasks these will be expected complete, providing generic out-of-the-box solutions that meet needs every possible user is clearly intractable. To address this problem, must therefore not only capable learning how complete novel at run-time, but also informed by user. In letter demonstrate behaviour trees, a established control architecture fields gaming robotics, can used conjunction with natural language instruction provide robust modular for instructing autonomous agents learn perform complex tasks. We show trees generated using our approach generalised scenarios, re-used future episodes create increasingly behaviours. validate work against an existing corpus instructions, application on both simulated robot solving toy two distinct real-world platforms which, respectively, block sorting scenario, patrol scenario.