We present Success weighted by Completion Time (SCT), a new metric for evaluating navigation performance for mobile robots. Several related works on navigation have used Success weighted by Path Length (SPL) as the primary method of evaluating the path an agent makes to a goal location, but SPL is limited in its ability to properly evaluate agents with complex dynamics. In contrast, SCT explicitly takes the agent's dynamics model into consideration, and aims to accurately capture how well the agent has approximated the fastest navigation behavior afforded by its dynamics. While several embodied navigation works use point-turn dynamics, we focus on unicycle-cart dynamics for our agent, which better exemplifies the dynamics model of popular mobile robotics platforms (e.g., LoCoBot, TurtleBot, Fetch, etc.). We also present RRT*-Unicycle, an algorithm for unicycle dynamics that estimates the fastest collision-free path and completion time from a starting pose to a goal location in an environment containing obstacles. We experiment with deep reinforcement learning and reward shaping to train and compare the navigation performance of agents with different dynamics models. In evaluating these agents, we show that in contrast to SPL, SCT is able to capture the advantages in navigation speed a unicycle model has over a simpler point-turn model of dynamics. Lastly, we show that we can successfully deploy our trained models and algorithms outside of simulation in the real world. We embody our agents in an real robot to navigate an apartment, and show that they can generalize in a zero-shot manner.
Simulation provides a safe and efficient way to generate useful data for learning complex robotic tasks. However, matching simulation and real-world dynamics can be quite challenging, especially for systems that have a large number of unobserved or unmeasurable parameters, which may lie in the robot dynamics itself or in the environment with which the robot interacts. We introduce a novel approach to tackle such a sim-to-real problem by developing policies capable of adapting to new environments, in a zero-shot manner. Key to our approach is an error-aware policy (EAP) that is explicitly made aware of the effect of unobservable factors during training. An EAP takes as input the predicted future state error in the target environment, which is provided by an error-prediction function, simultaneously trained with the EAP. We validate our approach on an assistive walking device trained to help the human user recover from external pushes. We show that a trained EAP for a hip-torque assistive device can be transferred to different human agents with unseen biomechanical characteristics. In addition, we show that our method can be applied to other standard RL control tasks.
"Looking for things" is a mundane but critical task we repeatedly carry on in our daily life. We introduce a method to develop a human character capable of searching for a randomly located target object in a detailed 3D scene using its locomotion capability and egocentric vision perception represented as RGBD images. By depriving the privileged 3D information from the human character, it is forced to move and look around simultaneously to account for the restricted sensing capability, resulting in natural navigation and search behaviors. Our method consists of two components: 1) a search control policy based on an abstract character model, and 2) an online replanning control module for synthesizing detailed kinematic motion based on the trajectories planned by the search policy. We demonstrate that the combined techniques enable the character to effectively find often occluded household items in indoor environments. The same search policy can be applied to different full-body characters without the need for retraining. We evaluate our method quantitatively by testing it on randomly generated scenarios. Our work is a first step toward creating intelligent virtual agents with humanlike behaviors driven by onboard sensors, paving the road toward future robotic applications.
Target-driven visual navigation is a challenging problem that requires a robot to find the goal using only visual inputs. Many researchers have demonstrated promising results using deep reinforcement learning (deep RL) on various robotic platforms, but typical end-to-end learning is known for its poor extrapolation capability to new scenarios. Therefore, learning a navigation policy for a new robot with a new sensor configuration or a new target still remains a challenging problem. In this paper, we introduce a learning algorithm that enables rapid adaptation to new sensor configurations or target objects with a few shots. We design a policy architecture with latent features between perception and inference networks and quickly adapt the perception network via meta-learning while freezing the inference network. Our experiments show that our algorithm adapts the learned navigation policy with only three shots for unseen situations with different sensor configurations or different target colors. We also analyze the proposed algorithm by investigating various hyperparameters.
Recent advances in deep reinforcement learning (deep RL) enable researchers to solve challenging control problems, from simulated environments to real-world robotic tasks. However, deep RL algorithms are known to be sensitive to the problem formulation, including observation spaces, action spaces, and reward functions. There exist numerous choices for observation spaces but they are often designed solely based on prior knowledge due to the lack of established principles. In this work, we conduct benchmark experiments to verify common design choices for observation spaces, such as Cartesian transformation, binary contact flags, a short history, or global positions. Then we propose a search algorithm to find the optimal observation spaces, which examines various candidate observation spaces and removes unnecessary observation channels with a Dropout-Permutation test. We demonstrate that our algorithm significantly improves learning speed compared to manually designed observation spaces. We also analyze the proposed algorithm by evaluating different hyperparameters.
Safety is an essential component for deploying reinforcement learning (RL) algorithms in real-world scenarios, and is critical during the learning process itself. A natural first approach toward safe RL is to manually specify constraints on the policy's behavior. However, just as learning has enabled progress in large-scale development of AI systems, learning safety specifications may also be necessary to ensure safety in messy open-world environments where manual safety specifications cannot scale. Akin to how humans learn incrementally starting in child-safe environments, we propose to learn how to be safe in one set of tasks and environments, and then use that learned intuition to constrain future behaviors when learning new, modified tasks. We empirically study this form of safety-constrained transfer learning in three challenging domains: simulated navigation, quadruped locomotion, and dexterous in-hand manipulation. In comparison to standard deep RL techniques and prior approaches to safe RL, we find that our method enables the learning of new tasks and in new environments with both substantially fewer safety incidents, such as falling or dropping an object, and faster, more stable learning. This suggests a path forward not only for safer RL systems, but also for more effective RL systems.
Reliable and stable locomotion has been one of the most fundamental challenges for legged robots. Deep reinforcement learning (deep RL) has emerged as a promising method for developing such control policies autonomously. In this paper, we develop a system for learning legged locomotion policies with deep RL in the real world with minimal human effort. The key difficulties for on-robot learning systems are automatic data collection and safety. We overcome these two challenges by developing a multi-task learning procedure, an automatic reset controller, and a safety-constrained RL framework. We tested our system on the task of learning to walk on three different terrains: flat ground, a soft mattress, and a doormat with crevices. Our system can automatically and efficiently learn locomotion skills on a Minitaur robot with little human intervention.
The ability to walk in new scenarios is a key milestone on the path toward real-world applications of legged robots. In this work, we introduce Meta Strategy Optimization, a meta-learning algorithm for training policies with latent variable inputs that can quickly adapt to new scenarios with a handful of trials in the target environment. The key idea behind MSO is to expose the same adaptation process, Strategy Optimization (SO), to both the training and testing phases. This allows MSO to effectively learn locomotion skills as well as a latent space that is suitable for fast adaptation. We evaluate our method on a real quadruped robot and demonstrate successful adaptation in various scenarios, including sim-to-real transfer, walking with a weakened motor, or climbing up a slope. Furthermore, we quantitatively analyze the generalization capability of the trained policy in simulated environments. Both real and simulated experiments show that our method outperforms previous methods in adaptation to novel tasks.
Imitation learning is a popular approach for training effective visual navigation policies. However, for legged robots collecting expert demonstrations is challenging as these robotic systems can be hard to control, move slowly, and cannot operate continuously for long periods of time. In this work, we propose a zero-shot imitation learning framework for training a visual navigation policy on a legged robot from only human demonstration (third-person perspective), allowing for high-quality navigation and cost-effective data collection. However, imitation learning from third-person perspective demonstrations raises unique challenges. First, these human demonstrations are captured from different camera perspectives, which we address via a feature disentanglement network~(FDN) that extracts perspective-agnostic state features. Second, as potential actions vary between systems, we reconstruct missing action labels by either building an inverse model of the robot's dynamics in the feature space and applying it to the demonstrations or developing an efficient Graphic User Interface (GUI) to label human demonstrations. To train a visual navigation policy we use a model-based imitation learning approach with the perspective-agnostic FDN and action-labeled demonstrations. We show that our framework can learn an effective policy for a legged robot, Laikago, from expert demonstrations in both simulated and real-world environments. Our approach is zero-shot as the robot never tries to navigate a given navigation path in the testing environment before the testing phase. We also justify our framework by performing an ablation study and comparing it with baselines.