A Vision-Guided Robotic System for Grasping Harvested Tomato Trusses in Cluttered Environments

Currently, truss tomato weighing and packaging require significant manual work. The main obstacle to automation lies in the difficulty of developing a reliable robotic grasping system for already harvested trusses. We propose a method to grasp trusses that are stacked in a crate with considerable clutter, which is how they are commonly stored and transported after harvest. The method consists of a deep learning-based vision system to first identify the individual trusses in the crate and then determine a suitable grasping location on the stem. To this end, we have introduced a grasp pose ranking algorithm with online learning capabilities. After selecting the most promising grasp pose, the robot executes a pinch grasp without needing touch sensors or geometric models. Lab experiments with a robotic manipulator equipped with an eye-in-hand RGB-D camera showed a 100% clearance rate when tasked to pick all trusses from a pile. 93% of the trusses were successfully grasped on the first try, while the remaining 7% required more attempts.

Imitrob: Imitation Learning Dataset for Training and Evaluating 6D Object Pose Estimators

This paper introduces a dataset for training and evaluating methods for 6D pose estimation of hand-held tools in task demonstrations captured by a standard RGB camera. Despite the significant progress of 6D pose estimation methods, their performance is usually limited for heavily occluded objects, which is a common case in imitation learning where the object is typically partially occluded by the manipulating hand. Currently, there is a lack of datasets that would enable the development of robust 6D pose estimation methods for these conditions. To overcome this problem, we collect a new dataset (Imitrob) aimed at 6D pose estimation in imitation learning and other applications where a human holds a tool and performs a task. The dataset contains image sequences of three different tools and six manipulation tasks with two camera viewpoints, four human subjects, and left/right hand. Each image is accompanied by an accurate ground truth measurement of the 6D object pose, obtained by the HTC Vive motion tracking device. The use of the dataset is demonstrated by training and evaluating a recent 6D object pose estimation method (DOPE) in various setups. The dataset and code are publicly available at http://imitrob.ciirc.cvut.cz/imitrobdataset.php.

Geometry-Based Grasping of Vine Tomatoes

We propose a geometry-based grasping method for vine tomatoes. It relies on a computer-vision pipeline to identify the required geometric features of the tomatoes and of the truss stem. The grasping method then uses a geometric model of the robotic hand and the truss to determine a suitable grasping location on the stem. This approach allows for grasping tomato trusses without requiring delicate contact sensors or complex mechanistic models and under minimal risk of damaging the tomatoes. Lab experiments were conducted to validate the proposed methods, using an RGB-D camera and a low-cost robotic manipulator. The success rate was 83% to 92%, depending on the type of truss.

GEM: Glare or Gloom, I Can Still See You -- End-to-End Multimodal Object Detector

Deep neural networks designed for vision tasks are often prone to failure when they encounter environmental conditions not covered by the training data. Efficient fusion strategies for multi-sensor configurations can enhance the robustness of the detection algorithms by exploiting redundancy from different sensor streams. In this paper, we propose sensor-aware multi-modal fusion strategies for 2D object detection in harsh-lighting conditions. Our network learns to estimate the measurement reliability of each sensor modality in the form of scalar weights and masks, without prior knowledge of the sensor characteristics. The obtained weights are assigned to the extracted feature maps which are subsequently fused and passed to the transformer encoder-decoder network for object detection. This is critical in the case of asymmetric sensor failures and to prevent any tragic consequences. Through extensive experimentation, we show that the proposed strategies out-perform the existing state-of-the-art methods on the FLIR-Thermal dataset, improving the mAP up-to 25.2%. We also propose a new "r-blended" hybrid depth modality for RGB-D multi-modal detection tasks. Our proposed method also obtained promising results on the SUNRGB-D dataset.

DeepKoCo: Efficient latent planning with an invariant Koopman representation

This paper presents DeepKoCo, a novel model-based agent that learns a latent Koopman representation from images. This representation allows DeepKoCo to plan efficiently using linear control methods, such as linear model predictive control. Compared to traditional agents, DeepKoCo is invariant to task-irrelevant dynamics, thanks to the use of a tailored lossy autoencoder network that allows DeepKoCo to learn latent dynamics that reconstruct and predict only observed costs, rather than all observed dynamics. As our results show, DeepKoCo achieves a similar final performance as traditional model-free methods on complex control tasks, while being considerably more robust to distractor dynamics, making the proposed agent more amenable for real-life applications.

Teaching robots to imitate a human with no on-teacher sensors. What are the key challenges?

In this paper, we consider the problem of learning object manipulation tasks from human demonstration using RGB or RGB-D cameras. We highlight the key challenges in capturing sufficiently good data with no tracking devices - starting from sensor selection and accurate 6DoF pose estimation to natural language processing. In particular, we focus on two showcases: gluing task with a glue gun and simple block-stacking with variable blocks. Furthermore, we discuss how a linguistic description of the task could help to improve the accuracy of task description. We also present the whole architecture of our transfer of the imitated task to the simulated and real robot environment.

Reinforcement Learning Applied to an Electric Water Heater: From Theory to Practice

Electric water heaters have the ability to store energy in their water buffer without impacting the comfort of the end user. This feature makes them a prime candidate for residential demand response. However, the stochastic and nonlinear dynamics of electric water heaters, makes it challenging to harness their flexibility. Driven by this challenge, this paper formulates the underlying sequential decision-making problem as a Markov decision process and uses techniques from reinforcement learning. Specifically, we apply an auto-encoder network to find a compact feature representation of the sensor measurements, which helps to mitigate the curse of dimensionality. A wellknown batch reinforcement learning technique, fitted Q-iteration, is used to find a control policy, given this feature representation. In a simulation-based experiment using an electric water heater with 50 temperature sensors, the proposed method was able to achieve good policies much faster than when using the full state information. In a lab experiment, we apply fitted Q-iteration to an electric water heater with eight temperature sensors. Further reducing the state vector did not improve the results of fitted Q-iteration. The results of the lab experiment, spanning 40 days, indicate that compared to a thermostat controller, the presented approach was able to reduce the total cost of energy consumption of the electric water heater by 15%.

Sampling-based Approximations with Quantitative Performance for the Probabilistic Reach-Avoid Problem over General Markov Processes

This article deals with stochastic processes endowed with the Markov (memoryless) property and evolving over general (uncountable) state spaces. The models further depend on a non-deterministic quantity in the form of a control input, which can be selected to affect the probabilistic dynamics. We address the computation of maximal reach-avoid specifications, together with the synthesis of the corresponding optimal controllers. The reach-avoid specification deals with assessing the likelihood that any finite-horizon trajectory of the model enters a given goal set, while avoiding a given set of undesired states. This article newly provides an approximate computational scheme for the reach-avoid specification based on the Fitted Value Iteration algorithm, which hinges on random sample extractions, and gives a-priori computable formal probabilistic bounds on the error made by the approximation algorithm: as such, the output of the numerical scheme is quantitatively assessed and thus meaningful for safety-critical applications. Furthermore, we provide tighter probabilistic error bounds that are sample-based. The overall computational scheme is put in relationship with alternative approximation algorithms in the literature, and finally its performance is practically assessed over a benchmark case study.

Residential Demand Response Applications Using Batch Reinforcement Learning

Driven by recent advances in batch Reinforcement Learning (RL), this paper contributes to the application of batch RL to demand response. In contrast to conventional model-based approaches, batch RL techniques do not require a system identification step, which makes them more suitable for a large-scale implementation. This paper extends fitted Q-iteration, a standard batch RL technique, to the situation where a forecast of the exogenous data is provided. In general, batch RL techniques do not rely on expert knowledge on the system dynamics or the solution. However, if some expert knowledge is provided, it can be incorporated by using our novel policy adjustment method. Finally, we tackle the challenge of finding an open-loop schedule required to participate in the day-ahead market. We propose a model-free Monte-Carlo estimator method that uses a metric to construct artificial trajectories and we illustrate this method by finding the day-ahead schedule of a heat-pump thermostat. Our experiments show that batch RL techniques provide a valuable alternative to model-based controllers and that they can be used to construct both closed-loop and open-loop policies.

Reinforcement learning for port-Hamiltonian systems

Passivity-based control (PBC) for port-Hamiltonian systems provides an intuitive way of achieving stabilization by rendering a system passive with respect to a desired storage function. However, in most instances the control law is obtained without any performance considerations and it has to be calculated by solving a complex partial differential equation (PDE). In order to address these issues we introduce a reinforcement learning approach into the energy-balancing passivity-based control (EB-PBC) method, which is a form of PBC in which the closed-loop energy is equal to the difference between the stored and supplied energies. We propose a technique to parameterize EB-PBC that preserves the systems's PDE matching conditions, does not require the specification of a global desired Hamiltonian, includes performance criteria, and is robust to extra non-linearities such as control input saturation. The parameters of the control law are found using actor-critic reinforcement learning, enabling learning near-optimal control policies satisfying a desired closed-loop energy landscape. The advantages are that near-optimal controllers can be generated using standard energy shaping techniques and that the solutions learned can be interpreted in terms of energy shaping and damping injection, which makes it possible to numerically assess stability using passivity theory. From the reinforcement learning perspective, our proposal allows for the class of port-Hamiltonian systems to be incorporated in the actor-critic framework, speeding up the learning thanks to the resulting parameterization of the policy. The method has been successfully applied to the pendulum swing-up problem in simulations and real-life experiments.