Search

We explore the impact of training with more diverse datasets, characterized by the number of unique samples, on the performance of self-supervised learning (SSL) under a fixed computational budget. Our findings consistently demonstrate that increasing pretraining data diversity enhances SSL performance, albeit only when the distribution distance to the downstream data is minimal. Notably, even with an exceptionally large pretraining data diversity achieved through methods like web crawling or diffusion-generated data, among other ways, the distribution shift remains a challenge. Our experiments are comprehensive with seven SSL methods using large-scale datasets such as ImageNet and YFCC100M amounting to over 200 GPU days. Code and trained models will be available at https://github.com/hammoudhasan/DiversitySSL ., Comment: Under review

In unstructured urban canals, regulation-aware interactions with other vessels are essential for collision avoidance and social compliance. In this paper, we propose a regulations aware motion planning framework for Autonomous Surface Vessels (ASVs) that accounts for dynamic and static obstacles. Our method builds upon local model predictive contouring control (LMPCC) to generate motion plans satisfying kino-dynamic and collision constraints in real-time while including regulation awareness. To incorporate regulations in the planning stage, we propose a cost function encouraging compliance with rules describing interactions with other vessels similar to COLlision avoidance REGulations at sea (COLREGs). These regulations are essential to make an ASV behave in a predictable and socially compliant manner with regard to other vessels. We compare the framework against baseline methods and show more effective regulation-compliance avoidance of moving obstacles with our motion planner. Additionally, we present experimental results in an outdoor environment, Comment: Accepted for presentation at ICRA 2022

To alleviate lower classification performance on rare classes in imbalanced datasets, a possible solution is to augment the underrepresented classes with synthetic samples. Domain adaptation can be incorporated in a classifier to decrease the domain discrepancy between real and synthetic samples. While domain adaptation is generally applied on completely synthetic source domains and real target domains, we explore how domain adaptation can be applied when only a single rare class is augmented with simulated samples. As a testbed, we use a camera trap animal dataset with a rare deer class, which is augmented with synthetic deer samples. We adapt existing domain adaptation methods to two new methods for the single rare class setting: DeerDANN, based on the Domain-Adversarial Neural Network (DANN), and DeerCORAL, based on deep correlation alignment (Deep CORAL) architectures. Experiments show that DeerDANN has the highest improvement in deer classification accuracy of 24.0% versus 22.4% improvement of DeerCORAL when compared to the baseline. Further, both methods require fewer than 10k synthetic samples, as used by the baseline, to achieve these higher accuracies. DeerCORAL requires the least number of synthetic samples (2k deer), followed by DeerDANN (8k deer)., Comment: 14 pages, 6 figures, to be published

We consider the problem of shaping the transient step response of nonlinear systems to satisfy a class of integral constraints. Such constraints are inherent in hybrid energy systems consisting of energy sources and storage elements. While typical transient specifications aim to minimize overshoot, this problem is unique in that it requires the presence of an appreciable overshoot to satisfy the foregoing constraints. The problem was previously studied in the context of stable linear systems. A combined integral and feedforward control, that requires minimal knowledge of the plant model, was shown to make the system amenable to meeting such constraints. This paper extends that work to nonlinear systems and proves the effectiveness of the same compensation structure under added conditions. Broadly, it is shown that the integral constraint is satisfied when this compensation structure is applied to nonlinear systems with stable open-loop step response and a positive DC gain. However, stability of the resulting closed-loop system mandates bounds on the controller gain., Comment: 14 pages, 4 figures

A rapid fracture test in four-point bending is proposed to assess hydrogen embrittlement (HE) susceptibility of high strength martensitic steels. The novelty of this technique is the rapid rate of loading, whereas conventional approaches require prolonged slow strain rate testing. The essential fractographic features required to identify the mechanisms of HE failure remain evident, despite the fast loading conditions. To demonstrate these attributes, two quenched and tempered steels at two different strength levels were tested, with and without pre-charging of hydrogen. Stress coupled hydrogen diffusion finite element analysis (FEA) was performed to calculate both stress and hydrogen concentration distributions. In addition to fractographic analysis, a mechanistic description rooted in hydrogen enhanced decohesion (HEDE) mechanism was used to corroborate the mechanical test data. The study shows that the approach is capable of quantifying HE susceptibility by being responsive to key factors affecting hydrogen induced fracture, thus developing further understanding on the HE of martensitic steels., Comment: 62 pages, 26 figures

This research addresses distributed proportional power sharing of inverter-based Distributed Generators (DGs) in microgrids under variations in maximum power capacity of DGs. A microgrid can include renewable energy resources such as wind turbines, solar panels, fuel cells, etc. The intermittent nature of such energy resources causes variations in their maximum power capacities. Since DGs in microgrids can be regarded as Multi-Agent-Systems (MASs), a consensus algorithm is designed to have the DGs generate their output power in proportion to their maximum capacities under capacity fluctuations. A change in power capacity of a DG triggers the consensus algorithm which uses a communication map at the cyber layer to estimate the corresponding change. During the transient time of reaching a consensus, the delivered power may not match the load power demand. To eliminate this mismatch, a control law is augmented that consists of a finite-time consensus algorithm embedded within the overarching power sharing consensus algorithm. The effectiveness of the distributed controller is assessed through simulation of a microgrid consisting of a realistic model of inverter-based DGs., Comment: 26 pages, 19 figures