Local-remote systems allow robots to execute complex tasks in hazardous environments such as space and nuclear power stations. However, establishing accurate positional mapping between local and remote devices can be difficult due to time delays that can compromise system performance and stability. Enhancing the synchronicity and stability of local-remote systems is vital for enabling robots to interact with environments at greater distances and under highly challenging network conditions, including time delays. We introduce an adaptive control method employing reinforcement learning to tackle the time-delayed control problem. By adjusting controller parameters in real-time, this adaptive controller compensates for stochastic delays and improves synchronicity between local and remote robotic manipulators. To improve the adaptive PD controller's performance, we devise a model-based reinforcement learning approach that effectively incorporates multi-step delays into the learning framework. Utilizing this proposed technique, the local-remote system's performance is stabilized for stochastic communication time-delays of up to 290ms. Our results demonstrate that the suggested model-based reinforcement learning method surpasses the Soft-Actor Critic and augmented state Soft-Actor Critic techniques. Access the code at: //github.com/CAV-Research-Lab/Predictive-Model-Delay-Correction
相關內容
We present an efficient framework for solving constrained global non-convex polynomial optimization problems. We prove the existence of an equivalent nonlinear reformulation of such problems that possesses essentially no spurious local minima. We show through numerical experiments that polynomial scaling in dimension and degree is achievable for computing the optimal value and location of previously intractable global constrained polynomial optimization problems in high dimension.
We proposed an extension of Akaike's relative power contribution that could be applied to data with correlations between noises. This method decomposes the power spectrum into a contribution of the terms caused by correlation between two noises, in addition to the contributions of the independent noises. Numerical examples confirm that some of the correlated noise has the effect of reducing the power spectrum.
In this work, we consider a real-time IoT monitoring system in which an energy harvesting sensor with a finite-size battery measures a physical process and transmits the status updates to an aggregator. The aggregator, equipped with caching capabilities, can serve the external requests of a destination network with either a stored update or a fresh update from the sensor. We assume the destination network acts as a gossiping network in which the update packets are forwarded among the nodes in a randomized setting. We utilize the Markov Decision Process framework to model and optimize the network's average Version Age of Information (AoI) and obtain the optimal policy at the aggregator. The structure of the optimal policy is analytically demonstrated and numerically verified. Numerical results highlight the effect of the system parameters on the average Version AoI. The simulations reveal the superior performance of the optimal policy compared to a set of baseline policies.
Time-optimal path planning in high winds for a turning-rate constrained UAV is a challenging problem to solve and is important for deployment and field operations. Previous works have used trochoidal path segments comprising straight and maximum-rate turn segments, as optimal extremal paths in uniform wind conditions. Current methods iterate over all candidate trochoidal trajectory types and select the one that is time-optimal; however, this exhaustive search can be computationally slow. In this paper, we introduce a method to decrease the computation time. This is achieved by reducing the number of candidate trochoidal trajectory types by framing the problem in the air-relative frame and bounding the solution within a subset of candidate trajectories. Our method reduces overall computation by 37.4% compared to pre-existing methods in Bang-Straight-Bang trajectories, freeing up computation for other onboard processes and can lead to significant total computational reductions when solving many trochoidal paths. When used within the framework of a global path planner, faster state expansions help find solutions faster or compute higher-quality paths. We also release our open-source codebase as a C++ package. The website and demo can be bound at //bradymoon.com/trochoids, codebase at //github.com/castacks/trochoids, and video at //youtu.be/qOU5gI7JshI .
Point cloud registration is challenging in the presence of heavy outlier correspondences. This paper focuses on addressing the robust correspondence-based registration problem with gravity prior that often arises in practice. The gravity directions are typically obtained by inertial measurement units (IMUs) and can reduce the degree of freedom (DOF) of rotation from 3 to 1. We propose a novel transformation decoupling strategy by leveraging screw theory. This strategy decomposes the original 4-DOF problem into three sub-problems with 1-DOF, 2-DOF, and 1-DOF, respectively, thereby enhancing the computation efficiency. Specifically, the first 1-DOF represents the translation along the rotation axis and we propose an interval stabbing-based method to solve it. The second 2-DOF represents the pole which is an auxiliary variable in screw theory and we utilize a branch-and-bound method to solve it. The last 1-DOF represents the rotation angle and we propose a global voting method for its estimation. The proposed method sequentially solves three consensus maximization sub-problems, leading to efficient and deterministic registration. In particular, it can even handle the correspondence-free registration problem due to its significant robustness. Extensive experiments on both synthetic and real-world datasets demonstrate that our method is more efficient and robust than state-of-the-art methods, even when dealing with outlier rates exceeding 99%.
Soft robotics aims to develop robots able to adapt their behavior across a wide range of unstructured and unknown environments. A critical challenge of soft robotic control is that nonlinear dynamics often result in complex behaviors hard to model and predict. Typically behaviors for mobile soft robots are discovered through empirical trial and error and hand-tuning. More recently, optimization algorithms such as Genetic Algorithms (GA) have been used to discover gaits, but these behaviors are often optimized for a single environment or terrain, and can be brittle to unplanned changes to terrain. In this paper we demonstrate how Quality Diversity Algorithms, which search of a range of high-performing behaviors, can produce repertoires of gaits that are robust to changing terrains. This robustness significantly out-performs that of gaits produced by a single objective optimization algorithm.
Grasping using an aerial robot can have many applications ranging from infrastructure inspection and maintenance to precise agriculture. However, aerial grasping is a challenging problem since the robot has to maintain an accurate position and orientation relative to the grasping object, while negotiating various forms of uncertainties (e.g., contact force from the object). To address such challenges, in this paper, we integrate a novel passive gripper design and advanced adaptive control methods to enable robust aerial grasping. The gripper is enabled by a pre-stressed band with two stable states (a flat shape and a curled shape). In this case, it can automatically initiate the grasping process upon contact with an object. The gripper also features a cable-driven system by a single DC motor to open the gripper without using cumbersome pneumatics. Since the gripper is passively triggered and initially has a straight shape, it can function without precisely aligning the gripper with the object (within an $80$ mm tolerance). Our adaptive control scheme eliminates the need for any a priori knowledge (nominal or upper bounds) of uncertainties. The closed-loop stability of the system is analyzed via Lyapunov-based method. Combining the gripper and the adaptive control, we conduct comparative real-time experimental results to demonstrate the effectiveness of the proposed integrated system for grasping. Our integrated approach can pave the way to enhance aerial grasping for different applications.
Knowledge graphs capture interlinked information between entities and they represent an attractive source of structured information that can be harnessed for recommender systems. However, existing recommender engines use knowledge graphs by manually designing features, do not allow for end-to-end training, or provide poor scalability. Here we propose Knowledge Graph Convolutional Networks (KGCN), an end-to-end trainable framework that harnesses item relationships captured by the knowledge graph to provide better recommendations. Conceptually, KGCN computes user-specific item embeddings by first applying a trainable function that identifies important knowledge graph relations for a given user and then transforming the knowledge graph into a user-specific weighted graph. Then, KGCN applies a graph convolutional neural network that computes an embedding of an item node by propagating and aggregating knowledge graph neighborhood information. Moreover, to provide better inductive bias KGCN uses label smoothness (LS), which provides regularization over edge weights and we prove that it is equivalent to label propagation scheme on a graph. Finally, We unify KGCN and LS regularization, and present a scalable minibatch implementation for KGCN-LS model. Experiments show that KGCN-LS outperforms strong baselines in four datasets. KGCN-LS also achieves great performance in sparse scenarios and is highly scalable with respect to the knowledge graph size.
Object detection typically assumes that training and test data are drawn from an identical distribution, which, however, does not always hold in practice. Such a distribution mismatch will lead to a significant performance drop. In this work, we aim to improve the cross-domain robustness of object detection. We tackle the domain shift on two levels: 1) the image-level shift, such as image style, illumination, etc, and 2) the instance-level shift, such as object appearance, size, etc. We build our approach based on the recent state-of-the-art Faster R-CNN model, and design two domain adaptation components, on image level and instance level, to reduce the domain discrepancy. The two domain adaptation components are based on H-divergence theory, and are implemented by learning a domain classifier in adversarial training manner. The domain classifiers on different levels are further reinforced with a consistency regularization to learn a domain-invariant region proposal network (RPN) in the Faster R-CNN model. We evaluate our newly proposed approach using multiple datasets including Cityscapes, KITTI, SIM10K, etc. The results demonstrate the effectiveness of our proposed approach for robust object detection in various domain shift scenarios.
Detecting carried objects is one of the requirements for developing systems to reason about activities involving people and objects. We present an approach to detect carried objects from a single video frame with a novel method that incorporates features from multiple scales. Initially, a foreground mask in a video frame is segmented into multi-scale superpixels. Then the human-like regions in the segmented area are identified by matching a set of extracted features from superpixels against learned features in a codebook. A carried object probability map is generated using the complement of the matching probabilities of superpixels to human-like regions and background information. A group of superpixels with high carried object probability and strong edge support is then merged to obtain the shape of the carried object. We applied our method to two challenging datasets, and results show that our method is competitive with or better than the state-of-the-art.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191