The Surprising Effectiveness of Linear Models for Visual Foresight in Object Pile Manipulation

By H. Suh, Russ Tedrake. 2020

70 citations recorded by the source corpus.

In this paper, we tackle the problem of pushing piles of small objects into a desired target set using visual feedback. Unlike conventional single-object manipulation pipelines, which estimate the state of the system parametrized by pose, the underlying physical state of this system is difficult to observe from images. Thus, we take the approach of reasoning directly in the space of images, and acquire the dynamics of visual measurements in order to synthesize a visual-feedback policy. We present a simple controller using an image-space Lyapunov function, and evaluate the closed-loop performance using three different class of models for image prediction: deep-learning-based models for image-to-image translation, an object-centric model obtained from treating each pixel as a particle, and a switched-linear system where an action-dependent linear map is used. Through results in simulation and experiment, we show that for this task, a linear model works surprisingly well -- achieving better prediction error, downstream task performance, and generalization to new environments than the deep models we trained on the same amount of data. We believe these results provide an interesting example in the spectrum of models that are most useful for vision-based feedback in manipulation, considering both the quality of visual prediction, as well as compatibility with rigorous methods for control design and analysis.

Subfields: Biophysics, Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biophysics

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Implicit Behavioral Cloning (2021, background): “This is likely because the particles have symmetries163 in image space, but not when observed as a vector of poses.164 Simulated Bi-Manual Sweeping consists of two robot KUKA IIWA arms equipped with spatula-like end165 effectors.”
Transporter Networks: Rearranging the Visual World for Robotic Manipulation (2020, background): “manipulating-rope∗† 3 3 3 7 sweeping-piles∗† [12] 3 3 3 7”
Example-Driven Model-Based Reinforcement Learning for Solving Long-Horizon Visuomotor Tasks (2021, methodology): “…model-based approaches [26, 27], which explicitly learn the environment forward dynamics from images have also been employed for multi-task vision-based robotic manipulation tasks, either with planning algorithms [28, 29, 30, 31, 32, 33, 34, 35] or for optimizing a parametric policy [36].”
MT-Opt: Continuous Multi-Task Robotic Reinforcement Learning at Scale (2021, background): “…supervision to learn diverse skills, e.g. by scaling up data collection [55], sampling suitable tasks [68] or goals [51] to practice, learning a task embedding space amenable to sampling [30], or learning a dynamics model and using model-predictive control to achieve goals [23, 19, 42, 73].”
Offline Reinforcement Learning from Images with Latent Space Models (2020, methodology): “Visual foresight algorithms (Finn and Levine, 2017; Ebert et al., 2018; Suh and Tedrake, 2020; Yen-Chen et al., 2019; Chen et al., 2020) handle control from pixels in a fully ofﬂine setting, but do not explicitly tackle the distributional shift issue that arises; meanwhile, our method is designed…”
Keypoints into the Future: Self-Supervised Correspondence in Model-Based Reinforcement Learning (2020, methodology): “Methods can be broadly categorized into whether they predict the image-space dynamics of the entire image [6, 5, 4, 28, 23], or predict the dynamics of a low-dimensional latent state [27, 25, 1, 10].”