Deep Learning has revolutionized our ability to solve complex problems such as Vision-and-Language Navigation (VLN). This task requires the agent to navigate to a goal purely based on visual sensory inputs given natural language instructions. However, prior works formulate the problem as a navigation graph with a discrete action space. In this work, we lift the agent off the navigation graph and propose a more complex VLN setting in continuous 3D reconstructed environments. Our proposed setting, Robo-VLN, more closely mimics the challenges of real world navigation. Robo-VLN tasks have longer trajectory lengths, continuous action spaces, and challenges such as obstacles. We provide a suite of baselines inspired by state-of-the-art works in discrete VLN and show that they are less effective at this task. We further propose that decomposing the task into specialized high- and low-level policies can more effectively tackle this task. With extensive experiments, we show that by using layered decision making, modularized training, and decoupling reasoning and imitation, our proposed Hierarchical Cross-Modal (HCM) agent outperforms existing baselines in all key metrics and sets a new benchmark for Robo-VLN.

Different from existing VLN environments, we propose a new continuous environment for VLN that more closely mirrors the challenges of the real world, Robo-VLN --- a continuous control formulation for Vision-and-Language Navigation. Compared to navigation graph based and discrete VLN settings, Robo-VLN provides longer horizon trajectories (4.5x average number of steps), more visual frames (3.5M visual frames), and a balanced high-level action distribution. Hence, making the problem more challenging and closer to the real-world. Try our Dataset!

We have released the PyTorch based implementation and dataset on the github page. Try our code!

                @inproceedings{irshad2021hierarchical,
                  title={Hierarchical Cross-Modal Agent for 
                        Robotics Vision-and-Language Navigation},
                  author={Muhammad Zubair Irshad and Chih-Yao Ma and Zsolt Kira},
                  booktitle={Proceedings of the IEEE International Conference 
                              on Robotics and Automation (ICRA)},
                  year={2021}
                  }