Dynamic Task Sequence Allocation Based on Natural Language Instruction

In recent years, with the development of technology, robots have become more prominent in the public eye. Service robots are now being used in some smart hotels to decrease manual transportation costs. However, these service robots operate according to set instructions, and humans usually apply fixed operating instructions to control the robots. Whether in daily life or at work, humans are more likely to use natural language to communicate and give task instructions. Moreover, in the field of multi-robotics, most of the existing researches on multi-task assignment have been conducted in simulation environments on a list of pre-set tasks, and seldom update the task sequences in real time in conjunction with the real environmental information and time. In this paper, we propose a method for assigning dynamic navigation task sequences based on natural language instruction and verify its effectiveness in a real scenario.