arXiv subjects

Tingnan Zhang

Publications and source records attributed to Tingnan Zhang.

52 records · Page 3Linked to original sources

arXiv2022

PI-ARS: Accelerating Evolution-Learned Visual-Locomotion with Predictive Information Representations

Evolution Strategy (ES) algorithms have shown promising results in training complex robotic control policies due to their massive parallelism capability, simple implementation, effective parameter-space exploration, and fast training time. However, a key limitation of ES is its scalability to large capacity models, including modern neural network architectures. In this work, we develop Predictive Information Augmented Random Search (PI-ARS) to mitigate this limitation by leveraging recent advancements in representation learning to reduce the parameter search space for ES. Namely, PI-ARS combines a gradient-based representation learning technique, Predictive Information (PI), with a gradient-free ES algorithm, Augmented Random Search (ARS), to train policies that can process complex robot sensory inputs and handle highly nonlinear robot dynamics. We evaluate PI-ARS on a set of challenging visual-locomotion tasks where a quadruped robot needs to walk on uneven stepping stones, quincuncial piles, and moving platforms, as well as to complete an indoor navigation task. Across all tasks, PI-ARS demonstrates significantly better learning efficiency and performance compared to the ARS baseline. We further validate our algorithm by demonstrating that the learned policies can successfully transfer to a real quadruped robot, for example, achieving a 100% success rate on the real-world stepping stone environment, dramatically improving prior results achieving 40% success.

cs.RO↗

arXiv2022

Adversarial Motion Priors Make Good Substitutes for Complex Reward Functions

Training a high-dimensional simulated agent with an under-specified reward function often leads the agent to learn physically infeasible strategies that are ineffective when deployed in the real world. To mitigate these unnatural behaviors, reinforcement learning practitioners often utilize complex reward functions that encourage physically plausible behaviors. However, a tedious labor-intensive tuning process is often required to create hand-designed rewards which might not easily generalize across platforms and tasks. We propose substituting complex reward functions with "style rewards" learned from a dataset of motion capture demonstrations. A learned style reward can be combined with an arbitrary task reward to train policies that perform tasks using naturalistic strategies. These natural strategies can also facilitate transfer to the real world. We build upon Adversarial Motion Priors -- an approach from the computer graphics domain that encodes a style reward from a dataset of reference motions -- to demonstrate that an adversarial approach to training policies can produce behaviors that transfer to a real quadrupedal robot without requiring complex reward functions. We also demonstrate that an effective style reward can be learned from a few seconds of motion capture data gathered from a German Shepherd and leads to energy-efficient locomotion strategies with natural gait transitions.

cs.AI↗

arXiv2022

Safe Reinforcement Learning for Legged Locomotion

Designing control policies for legged locomotion is complex due to the under-actuated and non-continuous robot dynamics. Model-free reinforcement learning provides promising tools to tackle this challenge. However, a major bottleneck of applying model-free reinforcement learning in real world is safety. In this paper, we propose a safe reinforcement learning framework that switches between a safe recovery policy that prevents the robot from entering unsafe states, and a learner policy that is optimized to complete the task. The safe recovery policy takes over the control when the learner policy violates safety constraints, and hands over the control back when there are no future safety violations. We design the safe recovery policy so that it ensures safety of legged locomotion while minimally intervening in the learning process. Furthermore, we theoretically analyze the proposed framework and provide an upper bound on the task performance. We verify the proposed framework in four locomotion tasks on a simulated and real quadrupedal robot: efficient gait, catwalk, two-leg balance, and pacing. On average, our method achieves 48.6% fewer falls and comparable or better rewards than the baseline methods in simulation. When deployed it on real-world quadruped robot, our training pipeline enables 34% improvement in energy efficiency for the efficient gait, 40.9% narrower of the feet placement in the catwalk, and two times more jumping duration in the two-leg balance. Our method achieves less than five falls over the duration of 115 minutes of hardware time.

cs.RO↗

arXiv2021

Fast and Efficient Locomotion via Learned Gait Transitions

We focus on the problem of developing energy efficient controllers for quadrupedal robots. Animals can actively switch gaits at different speeds to lower their energy consumption. In this paper, we devise a hierarchical learning framework, in which distinctive locomotion gaits and natural gait transitions emerge automatically with a simple reward of energy minimization. We use evolutionary strategies (ES) to train a high-level gait policy that specifies gait patterns of each foot, while the low-level convex MPC controller optimizes the motor commands so that the robot can walk at a desired velocity using that gait pattern. We test our learning framework on a quadruped robot and demonstrate automatic gait transitions, from walking to trotting and to fly-trotting, as the robot increases its speed. We show that the learned hierarchical controller consumes much less energy across a wide range of locomotion speed than baseline controllers.

cs.RO↗

arXiv2021

Unlocking Pixels for Reinforcement Learning via Implicit Attention

There has recently been significant interest in training reinforcement learning (RL) agents in vision-based environments. This poses many challenges, such as high dimensionality and the potential for observational overfitting through spurious correlations. A promising approach to solve both of these problems is an attention bottleneck, which provides a simple and effective framework for learning high performing policies, even in the presence of distractions. However, due to poor scalability of attention architectures, these methods cannot be applied beyond low resolution visual inputs, using large patches (thus small attention matrices). In this paper we make use of new efficient attention algorithms, recently shown to be highly effective for Transformers, and demonstrate that these techniques can be successfully adopted for the RL setting. This allows our attention-based controllers to scale to larger visual inputs, and facilitate the use of smaller patches, even individual pixels, improving generalization. We show this on a range of tasks from the Distracting Control Suite to vision-based quadruped robots locomotion. We provide rigorous theoretical analysis of the proposed algorithm.

cs.LG↗

arXiv2021

SimGAN: Hybrid Simulator Identification for Domain Adaptation via Adversarial Reinforcement Learning

As learning-based approaches progress towards automating robot controllers design, transferring learned policies to new domains with different dynamics (e.g. sim-to-real transfer) still demands manual effort. This paper introduces SimGAN, a framework to tackle domain adaptation by identifying a hybrid physics simulator to match the simulated trajectories to the ones from the target domain, using a learned discriminative loss to address the limitations associated with manual loss design. Our hybrid simulator combines neural networks and traditional physics simulation to balance expressiveness and generalizability, and alleviates the need for a carefully selected parameter set in System ID. Once the hybrid simulator is identified via adversarial reinforcement learning, it can be used to refine policies for the target domain, without the need to interleave data collection and policy refinement. We show that our approach outperforms multiple strong baselines on six robotic locomotion tasks for domain adaptation.

cs.RO↗

arXiv2020

Model-based Reinforcement Learning for Decentralized Multiagent Rendezvous

Collaboration requires agents to align their goals on the fly. Underlying the human ability to align goals with other agents is their ability to predict the intentions of others and actively update their own plans. We propose hierarchical predictive planning (HPP), a model-based reinforcement learning method for decentralized multiagent rendezvous. Starting with pretrained, single-agent point to point navigation policies and using noisy, high-dimensional sensor inputs like lidar, we first learn via self-supervision motion predictions of all agents on the team. Next, HPP uses the prediction models to propose and evaluate navigation subgoals for completing the rendezvous task without explicit communication among agents. We evaluate HPP in a suite of unseen environments, with increasing complexity and numbers of obstacles. We show that HPP outperforms alternative reinforcement learning, path planning, and heuristic-based baselines on challenging, unseen environments. Experiments in the real world demonstrate successful transfer of the prediction models from sim to real world without any additional fine-tuning. Altogether, HPP removes the need for a centralized operator in multiagent systems by combining model-based RL and inference methods, enabling agents to dynamically align plans.

cs.MA↗

arXiv2020

Learning Agile Robotic Locomotion Skills by Imitating Animals

Reproducing the diverse and agile locomotion skills of animals has been a longstanding challenge in robotics. While manually-designed controllers have been able to emulate many complex behaviors, building such controllers involves a time-consuming and difficult development process, often requiring substantial expertise of the nuances of each skill. Reinforcement learning provides an appealing alternative for automating the manual effort involved in the development of controllers. However, designing learning objectives that elicit the desired behaviors from an agent can also require a great deal of skill-specific expertise. In this work, we present an imitation learning system that enables legged robots to learn agile locomotion skills by imitating real-world animals. We show that by leveraging reference motion data, a single learning-based approach is able to automatically synthesize controllers for a diverse repertoire behaviors for legged robots. By incorporating sample efficient domain adaptation techniques into the training process, our system is able to learn adaptive policies in simulation that can then be quickly adapted for real-world deployment. To demonstrate the effectiveness of our system, we train an 18-DoF quadruped robot to perform a variety of agile behaviors ranging from different locomotion gaits to dynamic hops and turns.

cs.RO↗

arXiv2020

Zero-shot Imitation Learning from Demonstrations for Legged Robot Visual Navigation

Imitation learning is a popular approach for training visual navigation policies. However, collecting expert demonstrations for legged robots is challenging as these robots can be hard to control, move slowly, and cannot operate continuously for a long time. Here, we propose a zero-shot imitation learning approach for training a visual navigation policy on legged robots from human (third-person perspective) demonstrations, enabling high-quality navigation and cost-effective data collection. However, imitation learning from third-person demonstrations raises unique challenges. First, these demonstrations are captured from different camera perspectives, which we address via a feature disentanglement network (FDN) that extracts perspective-invariant state features. Second, as transition dynamics vary across systems, we label missing actions by either building an inverse model of the robot's dynamics in the feature space and applying it to the human demonstrations or developing a Graphic User Interface(GUI) to label human demonstrations. To train a navigation policy we use a model-based imitation learning approach with FDN and labeled human demonstrations. We show that our framework can learn an effective policy for a legged robot, Laikago, from human demonstrations in both simulated and real-world environments. Our approach is zero-shot as the robot never navigates the same paths during training as those at testing time. We justify our framework by performing a comparative study.

cs.AI↗

arXiv2019

A resistive force model of legged locomotion on granular media

Compared to agile legged animals, wheeled and tracked vehicles often suffer large performance loss on granular surfaces like sand and gravel. Understanding the mechanics of legged locomotion on granular media can aid the development of legged robots with improved mobility on granular surfaces; however, no general force model yet exists for granular media to predict ground reaction forces during complex limb intrusions. Inspired by a recent study of sand-swimming, we develop a resistive force model in the vertical plane for legged locomotion on granular media. We divide an intruder of complex morphology and kinematics, e.g., a bio-inspired robot L-leg rotated through uniform granular media (loosely packed ~ 1 mm diameter poppy seeds), into small segments, and measure stresses as a function of depth, orientation, and direction of motion using a model leg segment. Summation of segmental forces over the intruder predicts the net forces on both an L-leg and a reversed L-leg rotated through granular media with better accuracy than using simple one-dimensional penetration and drag force models. A multi-body dynamic simulation using the resistive force model predicts the speeds of a small legged robot (15 cm, 150 g) moving on granular media using both L-legs and reversed L-legs.

physics.bio-ph↗

arXiv2019

Policies Modulating Trajectory Generators

We propose an architecture for learning complex controllable behaviors by having simple Policies Modulate Trajectory Generators (PMTG), a powerful combination that can provide both memory and prior knowledge to the controller. The result is a flexible architecture that is applicable to a class of problems with periodic motion for which one has an insight into the class of trajectories that might lead to a desired behavior. We illustrate the basics of our architecture using a synthetic control problem, then go on to learn speed-controlled locomotion for a quadrupedal robot by using Deep Reinforcement Learning and Evolutionary Strategies. We demonstrate that a simple linear policy, when paired with a parametric Trajectory Generator for quadrupedal gaits, can induce walking behaviors with controllable speed from 4-dimensional IMU observations alone, and can be learned in under 1000 rollouts. We also transfer these policies to a real robot and show locomotion with controllable forward velocity.

cs.RO↗

arXiv2019

Data Efficient Reinforcement Learning for Legged Robots

We present a model-based framework for robot locomotion that achieves walking based on only 4.5 minutes (45,000 control steps) of data collected on a quadruped robot. To accurately model the robot's dynamics over a long horizon, we introduce a loss function that tracks the model's prediction over multiple timesteps. We adapt model predictive control to account for planning latency, which allows the learned model to be used for real time control. Additionally, to ensure safe exploration during model learning, we embed prior knowledge of leg trajectories into the action space. The resulting system achieves fast and robust locomotion. Unlike model-free methods, which optimize for a particular task, our planner can use the same learned dynamics for various tasks, simply by changing the reward function. To the best of our knowledge, our approach is more than an order of magnitude more sample efficient than current model-free methods.

cs.LG↗

arXiv2019

Modeling of the interaction of rigid wheels with dry granular media

We analyze the capabilities of various recently developed techniques, namely Resistive Force Theory (RFT) and continuum plasticity implemented with the Material Point Method (MPM), in capturing dynamics of wheel--dry granular media interactions. We compare results to more conventionally accepted methods of modeling wheel locomotion. While RFT is an empirical force model for arbitrarily-shaped bodies moving through granular media, MPM-based continuum modeling allows the simulation of full granular flow and stress fields. RFT allows for rapid evaluation of interaction forces on arbitrary shaped intruders based on a local surface stress formulation depending on depth, orientation, and movement of surface elements. We perform forced-slip experiments for three different wheel types and three different granular materials, and results are compared with RFT, continuum modeling, and a traditional terramechanics semi-empirical method. Results show that for the range of inputs considered, RFT can be reliably used to predict rigid wheel granular media interactions with accuracy exceeding that of traditional terramechanics methodology in several circumstances. Results also indicate that plasticity-based continuum modeling provides an accurate tool for wheel-soil interaction while providing more information to study the physical processes giving rise to resistive stresses in granular media.

cond-mat.soft↗

arXiv2018

Sim-to-Real: Learning Agile Locomotion For Quadruped Robots

Designing agile locomotion for quadruped robots often requires extensive expertise and tedious manual tuning. In this paper, we present a system to automate this process by leveraging deep reinforcement learning techniques. Our system can learn quadruped locomotion from scratch using simple reward signals. In addition, users can provide an open loop reference to guide the learning process when more control over the learned gait is needed. The control policies are learned in a physics simulator and then deployed on real robots. In robotics, policies trained in simulation often do not transfer to the real world. We narrow this reality gap by improving the physics simulator and learning robust policies. We improve the simulation using system identification, developing an accurate actuator model and simulating latency. We learn robust controllers by randomizing the physical environments, adding perturbations and designing a compact observation space. We evaluate our system on two agile locomotion gaits: trotting and galloping. After learning in simulation, a quadruped robot can successfully perform both gaits in the real world.

cs.RO↗

arXiv2016

A review on locomotion robophysics: the study of movement at the intersection of robotics, soft matter and dynamical systems

In this review we argue for the creation of a physics of moving systems -- a locomotion "robophysics" -- which we define as the pursuit of the discovery of principles of self generated motion. Robophysics can provide an important intellectual complement to the discipline of robotics, largely the domain of researchers from engineering and computer science. The essential idea is that we must complement study of complex robots in complex situations with systematic study of simplified robophysical devices in controlled laboratory settings and simplified theoretical models. We must thus use the methods of physics to examine successful and failed locomotion in simplified (abstracted) devices using parameter space exploration, systematic control, and techniques from dynamical systems. Using examples from our and other's research, we will discuss how such robophysical studies have begun to aid engineers in the creation of devices that begin to achieve life-like locomotor abilities on and within complex environments, have inspired interesting physics questions in low dimensional dynamical systems, geometric mechanics and soft matter physics, and have been useful to develop models for biological locomotion in complex terrain. The rapidly decreasing cost of constructing sophisticated robot models with easy access to significant computational power bodes well for scientists and engineers to engage in a discipline which can readily integrate experiment, theory and computation.

cs.RO↗

arXiv2013

A Terradynamics of Legged Locomotion on Granular Media

The theories of aero- and hydrodynamics predict animal movement and device design in air and water through the computation of lift, drag, and thrust forces. Although models of terrestrial legged locomotion have focused on interactions with solid ground, many animals move on substrates that flow in response to intrusion. However, locomotor-ground interaction models on such flowable ground are often unavailable. We developed a force model for arbitrarily-shaped legs and bodies moving freely in granular media, and used this "terradynamics" to predict a small legged robot's locomotion on granular media using various leg shapes and stride frequencies. Our study reveals a complex but generic dependence of stresses in granular media on intruder depth, orientation, and movement direction and gives insight into the effects of leg morphology and kinematics on movement.

physics.bio-ph↗