The hybrid zero dynamics control concept for bipedal walking is extended to include a non-instantaneous double support phase. A symmetric robot that consists of five rigid body segments which are connected by four actuated revolute joints is considered. Periodic walking gaits with a constant average walking speed consists of alternating single (SSP) and double support phases (DSP). Hybrid zero dynamics control designs usually assume an instantaneous DSP, which is a severe limitation. The proposed controllers use continuous SSPs and DSPs. Transitions between both phases are modeled as instantaneous events, when the rear leg lifts off at the end of the DSP and the swing leg touches down at the end of the SSP. Due to the fact that the model during the DSP has more actuators (4) than degrees of freedom (3), the system is overactuated. In order to combine it with the underactuated SSP model and then formulate a periodic walking gait, we suggest three controller designs for different applications. One with the underactuated DSP, one with the fully actuated DSP, and one with the overactuated DSP. A numerical optimization is used to generate energy efficient gaits in an offline process. According to the optimization results, artificially creating an underactuated controller for the DSP results in the most efficient gaits. Adding control tasks utilizing the full actuation or overactuation during the DSP significantly improves the gait stability.

The Multi-Commodity One-to-One Pickup and Delivery Traveling Salesman Problem finds the optimal tour that transports a set of unique commodities from their pickup to delivery locations, while never exceeding the maximum payload capacity of the material handling agent. For this NP hard problem, this paper presents adaptations of the nearest neighbor and cheapest insertion heuristics to account for the constraints related to the precedence between the locations and the cargo capacity limitations. To test the effectiveness of the proposed algorithms, the well-known TSPLIB benchmark data-set is modified in a replicable manner to create precedence constraints, while varying the cargo capacity of the agent. It is seen that the adapted Nearest Neighbor heuristic outperforms the adapted Cheapest Insertion algorithm in the majority of the cases studied, while providing near instantaneous solutions.

We present an optimization-based framework for rearranging indoor furniture to accommodate human-robot co-activities better. The rearrangement aims to afford sufficient accessible space for robot activities without compromising everyday human activities. To retain human activities, our algorithm preserves the functional relations among furniture by integrating spatial and semantic co-occurrence extracted from SUNCG and ConceptNet, respectively. By defining the robot's accessible space by the amount of open space it can traverse and the number of objects it can reach, we formulate the rearrangement for human-robot co-activity as an optimization problem, solved by adaptive simulated annealing (ASA) and covariance matrix adaptation evolution strategy (CMA-ES). Our experiments on the SUNCG dataset quantitatively show that rearranged scenes provide an average of 14% more accessible space and 30% more objects to interact with. The quality of the rearranged scenes is qualitatively validated by a human study, indicating the efficacy of the proposed strategy.

The success of machine learning is fueled by the increasing availability of computing power and large training datasets. The training data is used to learn new models or update existing ones, assuming that it is sufficiently representative of the data that will be encountered at test time. This assumption is challenged by the threat of poisoning, an attack that manipulates the training data to compromise the model's performance at test time. Although poisoning has been acknowledged as a relevant threat in industry applications, and a variety of different attacks and defenses have been proposed so far, a complete systematization and critical review of the field is still missing. In this survey, we provide a comprehensive systematization of poisoning attacks and defenses in machine learning, reviewing more than 100 papers published in the field in the last 15 years. We start by categorizing the current threat models and attacks, and then organize existing defenses accordingly. While we focus mostly on computer-vision applications, we argue that our systematization also encompasses state-of-the-art attacks and defenses for other data modalities. Finally, we discuss existing resources for research in poisoning, and shed light on the current limitations and open research questions in this research field.

Tensors in the form of multilinear arrays are ubiquitous in data science applications. Captured real-world data, including video, hyperspectral images, and discretized physical systems, naturally occur as tensors and often come with attendant noise. Under the additive noise model and with the assumption that the underlying clean tensor has low rank, many denoising methods have been created that utilize tensor decomposition to effect denoising through low rank tensor approximation. However, all such decomposition methods require estimating the tensor rank, or related measures such as the tensor spectral and nuclear norms, all of which are NP-hard problems. In this work we leverage our previously developed framework of $\textit{tensor amplification}$, which provides good approximations of the spectral and nuclear tensor norms, to denoising synthetic tensors of various sizes, ranks, and noise levels, along with real-world tensors derived from physiological signals. We also introduce two new notions of tensor rank -- $\textit{stable slice rank}$ and $\textit{stable }$$X$$\textit{-rank}$ -- and new denoising methods based on their estimation. The experimental results show that in the low rank context, tensor-based amplification provides comparable denoising performance in high signal-to-noise ratio (SNR) settings and superior performance in noisy (i.e., low SNR) settings, while the stable $X$-rank method achieves superior denoising performance on the physiological signal data.

Rawlings et al. provides an answer on how to design controllers directly (2017), Findeisen et al. (2007), Lucia et al. (2016)) has from data, where the system is implicitly represented via become a popular control scheme thanks to the ability of the Hankel matrix of measured trajectory. In short, the efficiently handling constraints and performance criteria lemma states that all trajectories of an controllable LTI as well as the coherent implementation of multiple hierarchical system can be represented by a finite set of its past trajectories, layers. The MPC scheme formulated as a semidefinite given that the past trajectories are generated by optimization problem in form of Linear Matrix a sufficiently exciting inputs. This idea has been investigated Inequalities (LMIs) is often used for some typical classes by De Persis and Tesi (2020), where the stabilizing of systems such as linear parameter-varying systems or feedback gain is designed by formulating the problem as Lur'e type systems, see for example, Kothare et al. (1996), LMIs without considering performance and constraints Böhm et al. (2009) and Nguyen et al. (2018). The reason guarantees, and Berberich et al. (2021), Coulson et al. for this is the formulated optimization problem is convex (2019) where this idea is used to develop a data-driven and can be efficiently solved.

This work links optimization approaches from hierarchical least-squares programming to instantaneous prioritized whole-body robot control. Concretely, we formulate the hierarchical Newton's method which solves prioritized non-linear least-squares problems in a numerically stable fashion even in the presence of kinematic and algorithmic singularities of the approximated kinematic constraints. These results are then transferred to control problems which exhibit the additional variability of time. This is necessary in order to formulate acceleration based controllers and to incorporate the second order dynamics. However, we show that the Newton's method without complicated adaptations is not appropriate in the acceleration domain. We therefore formulate a velocity based controller which exhibits second order proportional derivative convergence characteristics. Our developments are verified in toy robot control scenarios as well as in complex robot experiments which stress the importance of prioritized control and its singularity resolution.

We initiate the study of smoothed analysis for the sequential probability assignment problem with contexts. We study information-theoretically optimal minmax rates as well as a framework for algorithmic reduction involving the maximum likelihood estimator oracle. Our approach establishes a general-purpose reduction from minimax rates for sequential probability assignment for smoothed adversaries to minimax rates for transductive learning. This leads to optimal (logarithmic) fast rates for parametric classes and classes with finite VC dimension. On the algorithmic front, we develop an algorithm that efficiently taps into the MLE oracle, for general classes of functions. We show that under general conditions this algorithmic approach yields sublinear regret.

Finding a realistic deformation that transforms one image into another, in case large deformations are required, is considered a key challenge in medical image analysis. Having a proper image registration approach to achieve this could unleash a number of applications requiring information to be transferred between images. Clinical adoption is currently hampered by many existing methods requiring extensive configuration effort before each use, or not being able to (realistically) capture large deformations. A recent multi-objective approach that uses the Multi-Objective Real-Valued Gene-pool Optimal Mixing Evolutionary Algorithm (MO-RV-GOMEA) and a dual-dynamic mesh transformation model has shown promise, exposing the trade-offs inherent to image registration problems and modeling large deformations in 2D. This work builds on this promise and introduces MOREA: the first evolutionary algorithm-based multi-objective approach to deformable registration of 3D images capable of tackling large deformations. MOREA includes a 3D biomechanical mesh model for physical plausibility and is fully GPU-accelerated. We compare MOREA to two state-of-the-art approaches on abdominal CT scans of 4 cervical cancer patients, with the latter two approaches configured for the best results per patient. Without requiring per-patient configuration, MOREA significantly outperforms these approaches on 3 of the 4 patients that represent the most difficult cases.

In this paper, we study a navigation problem where a mobile robot needs to locate a mmWave wireless signal. Using the directionality properties of the signal, we propose an estimation and path planning algorithm that can efficiently navigate in cluttered indoor environments. We formulate Extended Kalman filters for emitter location estimation in cases where the signal is received in line-of-sight or after reflections. We then propose to plan motion trajectories based on belief-space dynamics in order to minimize the uncertainty of the position estimates. The associated non-linear optimization problem is solved by a state-of-the-art constrained iLQR solver. In particular, we propose a method that can handle a large number of obstacles (~300) with reasonable computation times. We validate the approach in an extensive set of simulations. We show that our estimators can help increase navigation success rate and that planning to reduce estimation uncertainty can improve the overall task completion speed.