arXiv subjects

Vijay Kumar

Publications and source records attributed to Vijay Kumar.

At least 163 records · Page 9Linked to original sources

arXiv2021

Semi-Supervised Visual Representation Learning for Fashion Compatibility

We consider the problem of complementary fashion prediction. Existing approaches focus on learning an embedding space where fashion items from different categories that are visually compatible are closer to each other. However, creating such labeled outfits is intensive and also not feasible to generate all possible outfit combinations, especially with large fashion catalogs. In this work, we propose a semi-supervised learning approach where we leverage large unlabeled fashion corpus to create pseudo-positive and pseudo-negative outfits on the fly during training. For each labeled outfit in a training batch, we obtain a pseudo-outfit by matching each item in the labeled outfit with unlabeled items. Additionally, we introduce consistency regularization to ensure that representation of the original images and their transformations are consistent to implicitly incorporate colour and other important attributes through self-supervision. We conduct extensive experiments on Polyvore, Polyvore-D and our newly created large-scale Fashion Outfits datasets, and show that our approach with only a fraction of labeled examples performs on-par with completely supervised methods.

cs.CV↗

arXiv2021

Defending a Perimeter from a Ground Intruder Using an Aerial Defender: Theory and Practice

The perimeter defense game has received interest in recent years as a variant of the pursuit-evasion game. A number of previous works have solved this game to obtain the optimal strategies for defender and intruder, but the derived theory considers the players as point particles with first-order assumptions. In this work, we aim to apply the theory derived from the perimeter defense problem to robots with realistic models of actuation and sensing and observe performance discrepancy in relaxing the first-order assumptions. In particular, we focus on the hemisphere perimeter defense problem where a ground intruder tries to reach the base of a hemisphere while an aerial defender constrained to move on the hemisphere aims to capture the intruder. The transition from theory to practice is detailed, and the designed system is simulated in Gazebo. Two metrics for parametric analysis and comparative study are proposed to evaluate the performance discrepancy.

cs.RO↗

arXiv2021

Multi-Robot Coverage and Exploration using Spatial Graph Neural Networks

The multi-robot coverage problem is an essential building block for systems that perform tasks like inspection or search and rescue. We discretize the coverage problem to induce a spatial graph of locations and represent robots as nodes in the graph. Then, we train a Graph Neural Network controller that leverages the spatial equivariance of the task to imitate an expert open-loop routing solution. This approach generalizes well to much larger maps and larger teams that are intractable for the expert. In particular, the model generalizes effectively to a simulation of ten quadrotors and dozens of buildings. We also demonstrate the GNN controller can surpass planning-based approaches in an exploration task.

cs.RO↗

arXiv2021

Learning Connectivity for Data Distribution in Robot Teams

Many algorithms for control of multi-robot teams operate under the assumption that low-latency, global state information necessary to coordinate agent actions can readily be disseminated among the team. However, in harsh environments with no existing communication infrastructure, robots must form ad-hoc networks, forcing the team to operate in a distributed fashion. To overcome this challenge, we propose a task-agnostic, decentralized, low-latency method for data distribution in ad-hoc networks using Graph Neural Networks (GNN). Our approach enables multi-agent algorithms based on global state information to function by ensuring it is available at each robot. To do this, agents glean information about the topology of the network from packet transmissions and feed it to a GNN running locally which instructs the agent when and where to transmit the latest state information. We train the distributed GNN communication policies via reinforcement learning using the average Age of Information as the reward function and show that it improves training stability compared to task-specific reward functions. Our approach performs favorably compared to industry-standard methods for data distribution such as random flooding and round robin. We also show that the trained policies generalize to larger teams of both static and mobile agents.

cs.RO↗

arXiv2021

Large Scale Model Predictive Control with Neural Networks and Primal Active Sets

This work presents an explicit-implicit procedure to compute a model predictive control (MPC) law with guarantees on recursive feasibility and asymptotic stability. The approach combines an offline-trained fully-connected neural network with an online primal active set solver. The neural network provides a control input initialization while the primal active set method ensures recursive feasibility and asymptotic stability. The neural network is trained with a primal-dual loss function, aiming to generate control sequences that are primal feasible and meet a desired level of suboptimality. Since the neural network alone does not guarantee constraint satisfaction, its output is used to warm start the primal active set method online. We demonstrate that this approach scales to large problems with thousands of optimization variables, which are challenging for current approaches. Our method achieves a 2x reduction in online inference time compared to the best method in a benchmark suite of different solver and initialization strategies.

cs.LG↗

arXiv2021

Importance Sampling based Exploration in Q Learning

Approximate Dynamic Programming (ADP) is a methodology to solve multi-stage stochastic optimization problems in multi-dimensional discrete or continuous spaces. ADP approximates the optimal value function by adaptively sampling both action and state space. It provides a tractable approach to very large problems, but can suffer from the exploration-exploitation dilemma. We propose a novel approach for selecting actions using importance sampling weighted by the value function approximation in continuous decision spaces to address this dilemma. An advantage of this approach is it balances exploration and exploitation without any tuning parameters when sampling actions compared to other exploration approaches such as Epsilon Greedy, instead relying only on the approximate value function. We compare the proposed algorithm with other exploration strategies in continuous action space in the context of a multi-stage generation expansion planning problem under uncertainty.

math.OC↗

arXiv2021

Adaptive and Risk-Aware Target Tracking with Heterogeneous Robot Teams

We consider a scenario where a team of robots with heterogeneous sensors must track a set of hostile targets which induce sensory failures on the robots. In particular, the likelihood of failures depends on the proximity between the targets and the robots. We propose a control framework that implicitly addresses the competing objectives of performance maximization and sensor preservation (which impacts the future performance of the team). Our framework consists of a predictive component -- which accounts for the risk of being detected by the target, and a reactive component -- which maximizes the performance of the team regardless of the failures that have already occurred. Based on a measure of the abundance of sensors in the team, our framework can generate aggressive and risk-averse robot configurations to track the targets. Crucially, the heterogeneous sensing capabilities of the robots are explicitly considered in each step, allowing for a more expressive risk-performance trade-off. Simulated experiments with induced sensor failures demonstrate the efficacy of the proposed approach.

cs.RO↗

arXiv2021

Perimeter-defense Game on Arbitrary Convex Shapes

This paper studies a variant of multi-player reach-avoid game played between intruders and defenders. The intruder team tries to score by sending as many intruders as possible to the target area, while the defender team tries to minimize this score by intercepting them. Specifically, we consider the case where the defenders are constrained to move on the perimeter of the target area. Since it is challenging to directly solve the multi-player game due to the high dimensionality of the joint state space, we leverage the solutions to smaller scale problems. First, we solve the one vs. one game, for which existing works either rely on numerical approaches or make simplifying assumptions (e.g., circular perimeter, or equal speed). This paper accommodates target areas with any arbitrary convex shapes and provides analytical solution which lends itself to a useful geometric interpretation. We also provide a detailed discussion on the optimality of the derived strategies. Secondly, we solve the two vs. one game to introduce a cooperative pincer maneuver, where a pair of defenders team up to capture an intruder that cannot be captured by either one of the defender individually. Finally, we introduce how the aforementioned building blocks are used in three different assignment-based defense strategies.

eess.SY↗

arXiv2021

Dispersion-Minimizing Motion Primitives for Search-Based Motion Planning

Search-based planning with motion primitives is a powerful motion planning technique that can provide dynamic feasibility, optimality, and real-time computation times on size, weight, and power-constrained platforms in unstructured environments. However, optimal design of the motion planning graph, while crucial to the performance of the planner, has not been a main focus of prior work. This paper proposes to address this by introducing a method of choosing vertices and edges in a motion primitive graph that is grounded in sampling theory and leads to theoretical guarantees on planner completeness. By minimizing dispersion of the graph vertices in the metric space induced by trajectory cost, we optimally cover the space of feasible trajectories with our motion primitive graph. In comparison with baseline motion primitives defined by uniform input space sampling, our motion primitive graphs have lower dispersion, find a plan with fewer iterations of the graph search, and have only one parameter to tune.

cs.RO↗

arXiv2021

Fair Robust Assignment using Redundancy

We study the consideration of fairness in redundant assignment for multi-agent task allocation. It has recently been shown that redundant assignment of agents to tasks provides robustness to uncertainty in task performance. However, the question of how to fairly assign these redundant resources across tasks remains unaddressed. In this paper, we present a novel problem formulation for fair redundant task allocation, which we cast as the optimization of worst-case task costs under a cardinality constraint. Solving this problem optimally is NP-hard. We exploit properties of supermodularity to propose a polynomial-time, near-optimal solution. In supermodular redundant assignment, the use of additional agents always improves task costs. Therefore, we provide a solution set that is $α$ times larger than the cardinality constraint. This constraint relaxation enables our approach to achieve a super-optimal cost by using a sub-optimal assignment size. We derive the sub-optimality bound on this cardinality relaxation, $α$. Additionally, we demonstrate that our algorithm performs near-optimally without the cardinality relaxation. We show simulations of redundant assignments of robots to goal nodes on transport networks with uncertain travel times. Empirically, our algorithm outperforms benchmarks, scales to large problems, and provides improvements in both fairness and average utility.

cs.RO↗

arXiv2021

Large Scale Distributed Collaborative Unlabeled Motion Planning with Graph Policy Gradients

In this paper, we present a learning method to solve the unlabelled motion problem with motion constraints and space constraints in 2D space for a large number of robots. To solve the problem of arbitrary dynamics and constraints we propose formulating the problem as a multi-agent problem. We are able to demonstrate the scalability of our methods for a large number of robots by employing a graph neural network (GNN) to parameterize policies for the robots. The GNN reduces the dimensionality of the problem by learning filters that aggregate information among robots locally, similar to how a convolutional neural network is able to learn local features in an image. Additionally, by employing a GNN we are also able to overcome the computational overhead of training policies for a large number of robots by first training graph filters for a small number of robots followed by zero-shot policy transfer to a larger number of robots. We demonstrate the effectiveness of our framework through various simulations.

cs.RO↗

arXiv2021

Belief Space Planning for Mobile Robots with Range Sensors using iLQG

In this work, we use iterative Linear Quadratic Gaussian (iLQG) to plan motions for a mobile robot with range sensors in belief space. We address two limitations that prevent applications of iLQG to the considered robotic system. First, iLQG assumes a differentiable measurement model, which is not true for range sensors. We show that iLQG only requires the differentiability of the belief dynamics. We propose to use a derivative-free filter to approximate the belief dynamics, which does not require explicit differentiability of the measurement model. Second, informative measurements from a range sensor are sparse. Uninformative measurements produce trivial gradient information, which prevent iLQG optimization from converging to a local minimum. We densify the informative measurements by introducing additional parameters in the measurement model. The parameters are iteratively updated in the optimization to ensure convergence to the true measurement model of a range sensor. We show the effectiveness of the proposed modifications through an ablation study. We also apply the proposed method in simulations of large scale real world environments, which show superior performance comparing to the state-of-the-art methods that either assume the separation principle or maximum likelihood measurements.

cs.RO↗

arXiv2021

ROS-NetSim: A Framework for the Integration of Robotic and Network Simulators

Multi-agent systems play an important role in modern robotics. Due to the nature of these systems, coordination among agents via communication is frequently necessary. Indeed, Perception-Action-Communication (PAC) loops, or Perception-Action loops closed over a communication channel, are a critical component of multi-robot systems. However, we lack appropriate tools for simulating PAC loops. To that end, in this paper, we introduce ROS-NetSim, a ROS package that acts as an interface between robotic and network simulators. With ROS-NetSim, we can attain high-fidelity representations of both robotic and network interactions by accurately simulating the PAC loop. Our proposed approach is lightweight, modular and adaptive. Furthermore, it can be used with many available network and physics simulators by making use of our proposed interface. In summary, ROS-NetSim is (i) Transparent to the ROS target application, (ii) Agnostic to the specific network and physics simulator being used, and (iii) Tunable in fidelity and complexity. As part of our contribution, we have made available an open-source implementation of ROS-NetSim to the community.

cs.RO↗

arXiv2021

Resilient Task Allocation in Heterogeneous Multi-Robot Systems

For a multi-robot system equipped with heterogeneous capabilities, this paper presents a mechanism to allocate robots to tasks in a resilient manner when anomalous environmental conditions such as weather events or adversarial attacks affect the performance of robots within the tasks. Our primary objective is to ensure that each task is assigned the requisite level of resources, measured as the aggregated capabilities of the robots allocated to the task. By keeping track of task performance deviations under external perturbations, our framework quantifies the extent to which robot capabilities (e.g., visual sensing or aerial mobility) are affected by environmental conditions. This enables an optimization-based framework to flexibly reallocate robots to tasks based on the most degraded capabilities within each task. In the face of resource limitations and adverse environmental conditions, our algorithm minimally relaxes the resource constraints corresponding to some tasks, thus exhibiting a graceful degradation of performance. Simulated experiments in a multi-robot coverage and target tracking scenario demonstrate the efficacy of the proposed approach.

cs.MA↗

arXiv2021

Neurosymbolic Transformers for Multi-Agent Communication

We study the problem of inferring communication structures that can solve cooperative multi-agent planning problems while minimizing the amount of communication. We quantify the amount of communication as the maximum degree of the communication graph; this metric captures settings where agents have limited bandwidth. Minimizing communication is challenging due to the combinatorial nature of both the decision space and the objective; for instance, we cannot solve this problem by training neural networks using gradient descent. We propose a novel algorithm that synthesizes a control policy that combines a programmatic communication policy used to generate the communication graph with a transformer policy network used to choose actions. Our algorithm first trains the transformer policy, which implicitly generates a "soft" communication graph; then, it synthesizes a programmatic communication policy that "hardens" this graph, forming a neurosymbolic transformer. Our experiments demonstrate how our approach can synthesize policies that generate low-degree communication graphs while maintaining near-optimal performance.

cs.MA↗

arXiv2020

Perimeter-defense Game between Aerial Defender and Ground Intruder

We study a variant of pursuit-evasion game in the context of perimeter defense. In this problem, the intruder aims to reach the base plane of a hemisphere without being captured by the defender, while the defender tries to capture the intruder. The perimeter-defense game was previously studied under the assumption that the defender moves on a circle. We extend the problem to the case where the defender moves on a hemisphere. To solve this problem, we analyze the strategies based on the breaching point at which the intruder tries to reach the target and predict the goal position, defined as optimal breaching point, that is achieved by the optimal strategies on both players. We provide the barrier that divides the state space into defender-winning and intruder-winning regions and prove that the optimal strategies for both players are to move towards the optimal breaching point. Simulation results are presented to demonstrate that the optimality of the game is given as a Nash equilibrium.

cs.RO↗

arXiv2020

Graph Neural Networks for Motion Planning

This paper investigates the feasibility of using Graph Neural Networks (GNNs) for classical motion planning problems. We propose guiding both continuous and discrete planning algorithms using GNNs' ability to robustly encode the topology of the planning space using a property called permutation invariance. We present two techniques, GNNs over dense fixed graphs for low-dimensional problems and sampling-based GNNs for high-dimensional problems. We examine the ability of a GNN to tackle planning problems such as identifying critical nodes or learning the sampling distribution in Rapidly-exploring Random Trees (RRT). Experiments with critical sampling, a pendulum and a six DoF robot arm show GNNs improve on traditional analytic methods as well as learning approaches using fully-connected or convolutional neural networks.

cs.RO↗

arXiv2020

Kernel Mean Embedding of Distributions: A Review and Beyond

A Hilbert space embedding of a distribution---in short, a kernel mean embedding---has recently emerged as a powerful tool for machine learning and inference. The basic idea behind this framework is to map distributions into a reproducing kernel Hilbert space (RKHS) in which the whole arsenal of kernel methods can be extended to probability measures. It can be viewed as a generalization of the original "feature map" common to support vector machines (SVMs) and other kernel methods. While initially closely associated with the latter, it has meanwhile found application in fields ranging from kernel machines and probabilistic modeling to statistical inference, causal discovery, and deep learning. The goal of this survey is to give a comprehensive review of existing work and recent advances in this research area, and to discuss the most challenging issues and open problems that could lead to new research directions. The survey begins with a brief introduction to the RKHS and positive definite kernels which forms the backbone of this survey, followed by a thorough discussion of the Hilbert space embedding of marginal distributions, theoretical guarantees, and a review of its applications. The embedding of distributions enables us to apply RKHS methods to probability measures which prompts a wide range of applications such as kernel two-sample testing, independent testing, and learning on distributional data. Next, we discuss the Hilbert space embedding for conditional distributions, give theoretical insights, and review some applications. The conditional mean embedding enables us to perform sum, product, and Bayes' rules---which are ubiquitous in graphical model, probabilistic inference, and reinforcement learning---in a non-parametric way. We then discuss relationships between this framework and other related areas. Lastly, we give some suggestions on future research directions.

stat.ML↗