arXiv subjects

Paramsothy Jayakumar

Publications and source records attributed to Paramsothy Jayakumar.

16 recordsLinked to original sources

arXiv2026

High-Speed, All-Terrain Autonomy: Ensuring Safety at the Limits of Mobility

A novel local trajectory planner, capable of controlling an autonomous off-road vehicle on rugged terrain at high-speed is presented. Autonomous vehicles are currently unable to safely operate off-road at high-speed, as current approaches either fail to predict and mitigate rollovers induced by rough terrain or are not real-time feasible. To address this challenge, a novel model predictive control (MPC) formulation is developed for local trajectory planning. A new dynamics model for off-road vehicles on rough, non-planar terrain is derived and used for prediction. Extreme mobility, including tire liftoff without rollover, is safely enabled through a new energy-based constraint. The formulation is analytically shown to mitigate rollover types ignored by many state-of-the-art methods, and real-time feasibility is achieved through parallelized GPGPU computation. The planner's ability to provide safe, extreme trajectories is studied through both simulated trials and full-scale physical experiments. The results demonstrate fewer rollovers and more successes compared to a state-of-the-art baseline across several challenging scenarios that push the vehicle to its mobility limits.

cs.RO↗

arXiv2025

Variational quantum and neural quantum states algorithms for the linear complementarity problem

Variational quantum algorithms (VQAs) are promising hybrid quantum-classical methods designed to leverage the computational advantages of quantum computing while mitigating the limitations of current noisy intermediate-scale quantum (NISQ) hardware. Although VQAs have been demonstrated as proofs of concept, their practical utility in solving real-world problems -- and whether quantum-inspired classical algorithms can match their performance -- remains an open question. We present a novel application of the variational quantum linear solver (VQLS) and its classical neural quantum states-based counterpart, the variational neural linear solver (VNLS), as key components within a minimum map Newton solver for a complementarity-based rigid body contact model. We demonstrate using the VNLS that our solver accurately simulates the dynamics of rigid spherical bodies during collision events. These results suggest that quantum and quantum-inspired linear algebra algorithms can serve as viable alternatives to standard linear algebra solvers for modeling certain physical systems.

cs.CE↗

arXiv2024

An application of continuous-variable gate synthesis to quantum simulation of classical dynamics

Although quantum computing holds promise to accelerate a wide range of computational tasks, the quantum simulation of quantum dynamics as originally envisaged by Feynman remains the most promising candidate for achieving quantum advantage. A less explored possibility with comparably far-reaching technological applicability is the quantum simulation of classical nonlinear dynamics. Attempts to develop digital quantum algorithms based on the Koopman von Neumann formalism have met with challenges because of the necessary projection step from an infinite-dimensional Hilbert space to the finite-dimensional subspace described by a collection of qubits. This finitization produces numerical artifacts that limit solutions to very short time horizons. In this paper we review continuous-variable quantum computing (CVQC), which naturally avoids such obstacles, and a CVQC algorithm for KvN simulation of classical nonlinear dynamics is advocated. In particular, we present explicit gate synthesis for product-formula Hamiltonian simulation of anharmonic vibrational dynamics.

quant-ph↗

arXiv2023

ConvBKI: Real-Time Probabilistic Semantic Mapping Network with Quantifiable Uncertainty

In this paper, we develop a modular neural network for real-time {\color{black}(> 10 Hz)} semantic mapping in uncertain environments, which explicitly updates per-voxel probabilistic distributions within a neural network layer. Our approach combines the reliability of classical probabilistic algorithms with the performance and efficiency of modern neural networks. Although robotic perception is often divided between modern differentiable methods and classical explicit methods, a union of both is necessary for real-time and trustworthy performance. We introduce a novel Convolutional Bayesian Kernel Inference (ConvBKI) layer which incorporates semantic segmentation predictions online into a 3D map through a depthwise convolution layer by leveraging conjugate priors. We compare ConvBKI against state-of-the-art deep learning approaches and probabilistic algorithms for mapping to evaluate reliability and performance. We also create a Robot Operating System (ROS) package of ConvBKI and test it on real-world perceptually challenging off-road driving data.

cs.RO↗

arXiv2022

Tensor-Train Compression of Discrete Element Method Simulation Data

We propose a framework for discrete scientific data compression based on the tensor-train (TT) decomposition. Our approach is tailored to handle unstructured output data from discrete element method (DEM) simulations, demonstrating its effectiveness in compressing both raw (e.g. particle position and velocity) and derived (e.g. stress and strain) datasets. We show that geometry-driven "tensorization" coupled with the TT decomposition (known as quantized TT) yields a hierarchical compression scheme, achieving high compression ratios for key variables in these DEM datasets.

math.NA↗

arXiv2022

Convolutional Bayesian Kernel Inference for 3D Semantic Mapping

Robotic perception is currently at a cross-roads between modern methods, which operate in an efficient latent space, and classical methods, which are mathematically founded and provide interpretable, trustworthy results. In this paper, we introduce a Convolutional Bayesian Kernel Inference (ConvBKI) layer which learns to perform explicit Bayesian inference within a depthwise separable convolution layer to maximize efficency while maintaining reliability simultaneously. We apply our layer to the task of real-time 3D semantic mapping, where we learn semantic-geometric probability distributions for LiDAR sensor information and incorporate semantic predictions into a global map. We evaluate our network against state-of-the-art semantic mapping algorithms on the KITTI data set, demonstrating improved latency with comparable semantic label inference results.

cs.RO↗

arXiv2022

Dynamic Semantic Occupancy Mapping using 3D Scene Flow and Closed-Form Bayesian Inference

This paper reports on a dynamic semantic mapping framework that incorporates 3D scene flow measurements into a closed-form Bayesian inference model. Existence of dynamic objects in the environment can cause artifacts and traces in current mapping algorithms, leading to an inconsistent map posterior. We leverage state-of-the-art semantic segmentation and 3D flow estimation using deep learning to provide measurements for map inference. We develop a Bayesian model that propagates the scene with flow and infers a 3D continuous (i.e., can be queried at arbitrary resolution) semantic occupancy map outperforming its static counterpart. Extensive experiments using publicly available data sets show that the proposed framework improves over its predecessors and input measurements from deep neural networks consistently.

cs.RO↗

arXiv2022

MotionSC: Data Set and Network for Real-Time Semantic Mapping in Dynamic Environments

This work addresses a gap in semantic scene completion (SSC) data by creating a novel outdoor data set with accurate and complete dynamic scenes. Our data set is formed from randomly sampled views of the world at each time step, which supervises generalizability to complete scenes without occlusions or traces. We create SSC baselines from state-of-the-art open source networks and construct a benchmark real-time dense local semantic mapping algorithm, MotionSC, by leveraging recent 3D deep learning architectures to enhance SSC with temporal information. Our network shows that the proposed data set can quantify and supervise accurate scene completion in the presence of dynamic objects, which can lead to the development of improved dynamic mapping algorithms. All software is available at https://github.com/UMich-CURLY/3DMapping.

cs.CV↗

arXiv2022

Continuous-variable optimization with neural network quantum states

Inspired by proposals for continuous-variable quantum approximate optimization (CV-QAOA), we investigate the utility of continuous-variable neural network quantum states (CV-NQS) for performing continuous optimization, focusing on the ground state optimization of the classical antiferromagnetic rotor model. Numerical experiments conducted using variational Monte Carlo with CV-NQS indicate that although the non-local algorithm succeeds in finding ground states competitive with the local gradient search methods, the proposal suffers from unfavorable scaling. A number of proposed extensions are put forward which may help alleviate the scaling difficulty.

quant-ph↗

arXiv2020

NLOptControl: A modeling language for solving optimal control problems

Current direct-collocation-based optimal control software is either easy to use or fast, but not both. This is a major limitation for users that are trying to formulate complex optimal control problems (OCPs) for use in on-line applications. This paper introduces NLOptControl, an open-source modeling language that allows users to both easily formulate and quickly solve nonlinear OCPs using direct-collocation methods. To achieve these attributes, NLOptControl (1) is written in an efficient, dynamically-typed computing language called Julia, (2) extends an optimization modeling language called JuMP to provide a natural algebraic syntax for modeling nonlinear OCPs; and (3) uses reverse automatic differentiation with the acrylic-coloring method to exploit sparsity in the Hessian matrix. This work explores the novel design features of NLOptControl and compares its syntax and speed to those of PROPT. The syntax comparisons shows that NLOptControl models OCPs more concisely than PROPT. The speeds of various collocation methods within PROPT and NLOptControl are benchmarked over a range of collocation points using performance profiles; overall, NLOptControl's single, two, and four interval pseudospectral methods are roughly $14$, $26$, and $36$ times faster than PROPT's, respectively. NLOptControl is well-suited to improve existing off-line and on-line control systems and to engender new ones.

cs.MS↗

arXiv2020

A Workload Adaptive Haptic Shared Control Scheme for Semi-Autonomous Driving

Haptic shared control is used to manage the control authority allocation between a human and an autonomous agent in semi-autonomous driving. Existing haptic shared control schemes, however, do not take full consideration of the human agent. To fill this research gap, this study presents a haptic shared control scheme that adapts to a human operator's workload, eyes on road and input torque in real-time. We conducted human-in-the-loop experiments with 24 participants. In the experiment, a human operator and an autonomy module for navigation shared the control of a simulated notional High Mobility Multipurpose Wheeled Vehicle (HMMWV) at a fixed speed. At the same time, the human operator performed a target detection task for surveillance. The autonomy could be either adaptive or non-adaptive to the above-mentioned human factors. Results indicate that the adaptive haptic control scheme resulted in significantly lower workload, higher trust in autonomy, better driving task performance and smaller control effort.

cs.HC↗

arXiv2020

An Active Learning Framework for Constructing High-fidelity Mobility Maps

A mobility map, which provides maximum achievable speed on a given terrain, is essential for path planning of autonomous ground vehicles in off-road settings. While physics-based simulations play a central role in creating next-generation, high-fidelity mobility maps, they are cumbersome and expensive. For instance, a typical simulation can take weeks to run on a supercomputer and each map requires thousands of such simulations. Recent work at the U.S. Army CCDC Ground Vehicle Systems Center has shown that trained machine learning classifiers can greatly improve the efficiency of this process. However, deciding which simulations to run in order to train the classifier efficiently is still an open problem. According to PAC learning theory, data that can be separated by a classifier is expected to require $\mathcal{O}(1/ε)$ randomly selected points (simulations) to train the classifier with error less than $ε$. In this paper, building on existing algorithms, we introduce an active learning paradigm that substantially reduces the number of simulations needed to train a machine learning classifier without sacrificing accuracy. Experimental results suggest that our sampling algorithm can train a neural network, with higher accuracy, using less than half the number of simulations when compared to random sampling.

cs.LG↗

arXiv2020

Neural network based terramechanics modeling and estimation for deformable terrains

In this work, a neural network based terramechanics model and terrain estimator are presented with an outlook for optimal control applications such as model predictive control. Recognizing the limitations of the state-of-the-art terramechanics models in terms of operating conditions, computational cost, and continuous differentiability for gradient-based optimization, an efficient and twice continuously differentiable terramechanics model is developed using neural networks for dynamic operations on deformable terrains. It is demonstrated that the neural network terramechanics model is able to predict the lateral tire forces accurately and efficiently compared to the Soil Contact Model as a state-of-the-art model. Furthermore, the neural network terramechanics model is implemented within a terrain estimator and it is shown that using this model the estimator converges within around 2% of the true terrain parameter. Finally, with model predictive control applications in mind, which typically rely on bicycle models for their predictions, it is demonstrated that utilizing the estimated terrain parameter can reduce prediction errors of a bicycle model by orders of magnitude. The result is an efficient, dynamic, twice continuously differentiable terramechanics model and estimator that has inherent advantages for implementation in model predictive control as compared to previously established models.

eess.SP↗

arXiv2020

Real-time trajectory planning for automated vehicle safety and performance in dynamic environments

Safe trajectory planning for high-performance automated vehicles in an environment with both static and moving obstacles is a challenging problem. Part of the challenge is developing a formulation that can be solved in real-time while including the following set of specifications: minimum time-to-goal, a dynamic vehicle model, minimum control effort, both static and moving obstacle avoidance, simultaneous optimization of speed and steering, and a short execution horizon. This paper presents a nonlinear model predictive control-based trajectory planning formulation, tailored for a large, high-speed unmanned ground vehicle, that includes the above set of specifications. This paper also evaluates NLOptControl's ability to solve this formulation in real-time in conjunction with the KNITRO nonlinear programming problem solver; NLOptControl is our open-source, direct-collocation based, optimal control problem solver. This formulation is tested with various sets of the specifications. In particular, a parametric study relating execution horizon and obstacle speed, indicates that the moving obstacle avoidance specification is not needed for safety when the planner has a small execution horizon ($\leq0.375\;s$) and the obstacles are moving slowly ($\leq2.11\frac{m}{s}$). However, a moving obstacle avoidance specification is needed when the obstacles are moving faster, and this specification improves the overall safety by a factor of $6.73$ ($p=2.2\times10^{-16}$) without, in most cases, increasing the solve-times. Overall, the results indicate that (1) safe trajectory planners for high-performance automated vehicles should include the entire set of specifications mentioned above, unless a static or low-speed environment permits a less comprehensive planner; and (2) NLOptControl can solve the formulation in real-time.

eess.SY↗

arXiv2019

Online terrain estimation for autonomous vehicles on deformable terrains

In this work, a terrain estimation framework is developed for autonomous vehicles operating on deformable terrains. Previous work in this area usually relies on steady state tire operation, linearized classical terramechanics models, or on computationally expensive algorithms that are not suitable for real-time estimation. To address these shortcomings, this work develops a reduced-order nonlinear terramechanics model as a surrogate of the Soil Contact Model (SCM) through extending a state-of-the-art Bekker model to account for additional dynamic effects. It is shown that this reduced-order surrogate model is able to accurately replicate the forces predicted by the SCM while reducing the computation cost by an order of magnitude. This surrogate model is then utilized in a unscented Kalman filter to estimate the sinkage exponent. Simulations suggest this parameter can be estimated within 4% of its true value for clay and sandy loam terrains. It is also shown that utilizing this estimated parameter can reduce the prediction errors of the future vehicle states by orders of magnitude, which could assist with achieving more robust model-predictive autonomous navigation strategies.

eess.SY↗

arXiv2018

Tensor Train accelerated solvers for nonsmooth rigid body dynamics

In the last two decades, increased need for high-fidelity simulations of the time evolution and propagation of forces in granular media has spurred renewed interest in discrete element method (DEM) modeling of frictional contact. Force penalty methods, while economic and accessible, introduce artificial stiffness, requiring small time steps to retain numerical stability. Optimization-based methods, which enforce contacts geometrically through complementarity constraints, allow the use of larger time steps at the expense of solving a nonlinear complementarity problem (NCP) each time step. We review the latest efforts to produce solvers for this NCP, focusing on its relaxation to a cone complementarity problem (CCP) and solution via an equivalent quadratic optimization problem with conic constraints. We distinguish between linearly convergent first order methods and second order methods, which gain quadratic convergence and more robust performance at the expense of the solution of large sparse linear systems. We propose a novel acceleration for the solution of Newton step linear systems in second order methods using low-rank compression based fast direct solvers. We use the Quantized Tensor Train (QTT) decomposition to produce efficient approximate representations of the system matrix and its inverse. This provides a robust framework to accelerate its solution in a direct or a preconditioned iterative method. In a number of numerical tests, we demonstrate that this approach displays sublinear scaling of precomputation costs, may be efficiently updated across Newton iterations as well as across time steps, and leads to a fast, optimal complexity solution of the Newton step. This allows our method to gain an order of magnitude speedups over state-of-the-art preconditioning techniques for moderate to large-scale systems, mitigating the computational bottleneck of second order methods.

math.NA↗