This paper presents a deep learning based model predictive control (MPC) algorithm for systems with unmatched and bounded state-action dependent uncertainties of unknown structure. We utilize a deep neural network (DNN) as an oracle in the underlying optimization problem of learning based MPC (LBMPC) to estimate unmatched uncertainties. Generally, non-parametric oracles such as DNN are considered difficult to employ with LBMPC due to the technical difficulties associated with estimation of their coefficients in real time. We employ a dual-timescale adaptation mechanism, where the weights of the last layer of the neural network are updated in real time while the inner layers are trained on a slower timescale using the training data collected online and selectively stored in a buffer. Our results are validated through a numerical experiment on the compression system model of jet engine. These results indicate that the proposed approach is implementable in real time and carries the theoretical guarantees of LBMPC.

This paper presents a deep learning based model predictive control algorithm for control affine nonlinear discrete time systems with matched and bounded state-dependent uncertainties of unknown structure. Since the structure of uncertainties is not known, a deep neural network (DNN) is employed to approximate the disturbances. In order to avoid any unwanted behavior during the learning phase, a tube based model predictive controller is employed, which ensures satisfaction of constraints and input-to-state stability of the closed-loop states.

Remain Well Clear, keeping the aircraft away from hazards by the appropriate separation distance, is an essential technology for the safe operation of uncrewed aerial vehicles in congested airspace. This work focuses on automating the horizontal separation of two aircraft and presents the obstacle avoidance problem as a 2D surrogate optimization task. By our design, the surrogate task is made more conservative to guarantee the execution of the solution in the primary domain. Using Reinforcement Learning (RL), we optimize the avoidance policy and model the dynamics, interactions, and decision-making. By recursively sampling the resulting policy and the surrogate transitions, the system translates the avoidance policy into a complete avoidance trajectory. Then, the solver publishes the trajectory as a set of waypoints for the airplane to follow using the Robot Operating System (ROS) interface. The proposed system generates a quick and achievable avoidance trajectory that satisfies the safety requirements. Evaluation of our system is completed in a high-fidelity simulation and full-scale airplane demonstration. Moreover, the paper concludes an enormous integration effort that has enabled a real-life demonstration of the RL-based system.

Node classification for graph-structured data aims to classify nodes whose labels are unknown. While studies on static graphs are prevalent, few studies have focused on dynamic graph node classification. Node classification on dynamic graphs is challenging for two reasons. First, the model needs to capture both structural and temporal information, particularly on dynamic graphs with a long history and require large receptive fields. Second, model scalability becomes a significant concern as the size of the dynamic graph increases. To address these problems, we propose the Time Augmented Dynamic Graph Neural Network (TADGNN) framework. TADGNN consists of two modules: 1) a time augmentation module that captures the temporal evolution of nodes across time structurally, creating a time-augmented spatio-temporal graph, and 2) an information propagation module that learns the dynamic representations for each node across time using the constructed time-augmented graph. We perform node classification experiments on four dynamic graph benchmarks. Experimental results demonstrate that TADGNN framework outperforms several static and dynamic state-of-the-art (SOTA) GNN models while demonstrating superior scalability. We also conduct theoretical and empirical analyses to validate the efficiency of the proposed method. Our code is available at https://sites.google.com/view/tadgnn.

Realistic long-horizon tasks like image-goal navigation involve exploratory and exploitative phases. Assigned with an image of the goal, an embodied agent must explore to discover the goal, i.e., search efficiently using learned priors. Once the goal is discovered, the agent must accurately calibrate the last-mile of navigation to the goal. As with any robust system, switches between exploratory goal discovery and exploitative last-mile navigation enable better recovery from errors. Following these intuitive guide rails, we propose SLING to improve the performance of existing image-goal navigation systems. Entirely complementing prior methods, we focus on last-mile navigation and leverage the underlying geometric structure of the problem with neural descriptors. With simple but effective switches, we can easily connect SLING with heuristic, reinforcement learning, and neural modular policies. On a standardized image-goal navigation benchmark (Hahn et al. 2021), we improve performance across policies, scenes, and episode complexity, raising the state-of-the-art from 45% to 55% success rate. Beyond photorealistic simulation, we conduct real-robot experiments in three physical scenes and find these improvements to transfer well to real environments.

For soft continuum arms, visual servoing is a popular control strategy that relies on visual feedback to close the control loop. However, robust visual servoing is challenging as it requires reliable feature extraction from the image, accurate control models and sensors to perceive the shape of the arm, both of which can be hard to implement in a soft robot. This letter circumvents these challenges by presenting a deep neural network-based method to perform smooth and robust 3D positioning tasks on a soft arm by visual servoing using a camera mounted at the distal end of the arm. A convolutional neural network is trained to predict the actuations required to achieve the desired pose in a structured environment. Integrated and modular approaches for estimating the actuations from the image are proposed and are experimentally compared. A proportional control law is implemented to reduce the error between the desired and current image as seen by the camera. The model together with the proportional feedback control makes the described approach robust to several variations such as new targets, lighting, loads, and diminution of the soft arm. Furthermore, the model lends itself to be transferred to a new environment with minimal effort.

We describe a system for visually guided autonomous navigation of under-canopy farm robots. Low-cost under-canopy robots can drive between crop rows under the plant canopy and accomplish tasks that are infeasible for over-the-canopy drones or larger agricultural equipment. However, autonomously navigating them under the canopy presents a number of challenges: unreliable GPS and LiDAR, high cost of sensing, challenging farm terrain, clutter due to leaves and weeds, and large variability in appearance over the season and across crop types. We address these challenges by building a modular system that leverages machine learning for robust and generalizable perception from monocular RGB images from low-cost cameras, and model predictive control for accurate control in challenging terrain. Our system, CropFollow, is able to autonomously drive 485 meters per intervention on average, outperforming a state-of-the-art LiDAR based system (286 meters per intervention) in extensive field testing spanning over 25 km.

Efficient and robust policy transfer remains a key challenge for reinforcement learning to become viable for real-wold robotics. Policy transfer through warm initialization, imitation, or interacting over a large set of agents with randomized instances, have been commonly applied to solve a variety of Reinforcement Learning tasks. However, this seems far from how skill transfer happens in the biological world: Humans and animals are able to quickly adapt the learned behaviors between similar tasks and learn new skills when presented with new situations. Here we seek to answer the question: Will learning to combine adaptation and exploration lead to a more efficient transfer of policies between domains? We introduce a principled mechanism that can "Adapt-to-Learn", that is adapt the source policy to learn to solve a target task with significant transition differences and uncertainties. We show that the presented method learns to seamlessly combine learning from adaptation and exploration and leads to a robust policy transfer algorithm with significantly reduced sample complexity in transferring skills between related tasks.

This paper presents high precision control and deep learning-based corn stand counting algorithms for a low-cost, ultra-compact 3D printed and autonomous field robot for agricultural operations. Currently, plant traits, such as emergence rate, biomass, vigor, and stand counting, are measured manually. This is highly labor-intensive and prone to errors. The robot, termed TerraSentia, is designed to automate the measurement of plant traits for efficient phenotyping as an alternative to manual measurements. In this paper, we formulate a Nonlinear Moving Horizon Estimator (NMHE) that identifies key terrain parameters using onboard robot sensors and a learning-based Nonlinear Model Predictive Control (NMPC) that ensures high precision path tracking in the presence of unknown wheel-terrain interaction. Moreover, we develop a machine vision algorithm designed to enable an ultra-compact ground robot to count corn stands by driving through the fields autonomously. The algorithm leverages a deep network to detect corn plants in images, and a visual tracking model to re-identify detected objects at different time steps. We collected data from 53 corn plots in various fields for corn plants around 14 days after emergence (stage V3 - V4). The robot predictions have agreed well with the ground truth with $C_{robot}=1.02 \times C_{human}-0.86$ and a correlation coefficient $R=0.96$. The mean relative error given by the algorithm is $-3.78\%$, and the standard deviation is $6.76\%$. These results indicate a first and significant step towards autonomous robot-based real-time phenotyping using low-cost, ultra-compact ground robots for corn and potentially other crops.

To achieve high-levels of autonomy, modern robots require the ability to detect and recover from anomalies and failures with minimal human supervision. Multi-modal sensor signals could provide more information for such anomaly detection tasks; however, the fusion of high-dimensional and heterogeneous sensor modalities remains a challenging problem. We propose a deep learning neural network: supervised variational autoencoder (SVAE), for failure identification in unstructured and uncertain environments. Our model leverages the representational power of VAE to extract robust features from high-dimensional inputs for supervised learning tasks. The training objective unifies the generative model and the discriminative model, thus making the learning a one-stage procedure. Our experiments on real field robot data demonstrate superior failure identification performance than baseline methods, and that our model learns interpretable representations. Videos of our results are available on our website: https://sites.google.com/illinois.edu/supervised-vae .