# A Data-Efficient Framework for Training and Sim-to-Real Transfer of Navigation Policies

arXiv: https://arxiv.org/abs/1810.04871 · HTML: https://mlsanity.com/paper/1810.04871/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Oct 2018.

## Built on by (most cited first, 10 of 10)

- [Transfer Learning in Robotics: An Upcoming Breakthrough? A Review of Promises and Challenges](https://mlsanity.com/paper/2311.18044/) (2023)
- [Autonomous Robotic Reinforcement Learning with Asynchronous Human Feedback](https://mlsanity.com/paper/2310.20608/) (2023)
- [Weakly Supervised Correspondence Learning](https://mlsanity.com/paper/2203.00904/) (2022)
- [Autonomous Driving Small-Scale Cars: A Survey of Recent Development](https://mlsanity.com/paper/2404.06229/) (2024)
- [Learning from Imperfect Demonstrations via Adversarial Confidence Transfer](https://mlsanity.com/paper/2202.02967/) (2022)
- [Harfang3D Dog-Fight Sandbox: A Reinforcement Learning Research Platform for the Customized Control Tasks of Fighter Aircrafts](https://mlsanity.com/paper/2210.07282/) (2022)
- [Predicting Human Perceptions of Robot Performance During Navigation Tasks](https://mlsanity.com/paper/2310.11590/) (2023)
- [Towards self-attention based visual navigation in the real world](https://mlsanity.com/paper/2209.07043/) (2022)
- [Cross-modality debiasing: using language to mitigate sub-population shifts in imaging](https://mlsanity.com/paper/2403.07888/) (2024)
- [On Assessing the Usefulness of Proxy Domains for Developing and Evaluating Embodied Agents](https://mlsanity.com/paper/2109.14516/) (2021)

## Abstract (arXiv)

Learning effective visuomotor policies for robots purely from data is challenging, but also appealing since a learning-based system should not require manual tuning or calibration. In the case of a robot operating in a real environment the training process can be costly, time-consuming, and even dangerous since failures are common at the start of training. For this reason, it is desirable to be able to leverage textitsimulation and textitoff-policy data to the extent possible to train the robot. In this work, we introduce a robust framework that plans in simulation and transfers well to the real environment. Our model incorporates a gradient-descent based planning module, which, given the initial image and goal image, encodes the images to a lower dimensional latent state and plans a trajectory to reach the goal. The model, consisting of the encoder and planner modules, is trained through a meta-learning strategy in simulation first. We subsequently perform adversarial domain transfer on the encoder by using a bank of unlabelled but random images from the simulation and real environments to enable the encoder to map images from the real and simulated environments to a similarly distributed latent representation. By fine tuning the entire model (encoder + planner) with far fewer real world expert demonstrations, we show successful planning performances in different navigation tasks.