arXiv subjects

Nakul Garg

Publications and source records attributed to Nakul Garg.

6 recordsLinked to original sources

arXiv2026

GRADE: Single-Frame Generative Radar Depth Estimation Under Visual Degradation

Dense 3D depth perception fails under smoke, fog, and darkness because optical sensors cannot penetrate airborne particulates. mmWave radar remains usable and measures range accurately under these conditions, but its small aperture limits angular resolution. We present GRADE, which grounds a pretrained generative prior in single-frame radar geometry to estimate high-fidelity metric depth. GRADE first maps raw 4D radar spectra to coarse metric depth. A latent diffusion backbone then recovers structural detail while conditioning every denoising step on this estimate. A pixel-space adapter uses residual camera cues when available and is trained across clear, smoke-degraded, and occluded inputs so the full output approaches the radar-conditioned path as visibility degrades. Trained and evaluated on ~95K frames across 12 buildings with real smoke, GRADE achieves an MAE of 0.303 m in clear scenes and 0.313 m under smoke, outperforming existing baselines. Code and datasets are available at https://phi-lab-rice.github.io/GRADE.

cs.CV↗

arXiv2025

RadarSFD: Single-Frame Diffusion with Pretrained Priors for Radar Point Clouds

Millimeter-wave radar provides robust perception in fog, smoke, dust, and low light, making it attractive for size-, weight-, and power-constrained robotic platforms. Existing radar imaging methods typically rely on synthetic aperture or multi-frame aggregation to improve resolution, which is impractical for small aerial, inspection, or wearable systems. We present RadarSFD, a conditional latent diffusion framework that reconstructs dense LiDAR-like point clouds from a single radar frame without motion or SAR. Our approach transfers geometric priors from a pretrained monocular depth estimator into the diffusion backbone, anchors them to radar inputs via channel-wise latent concatenation, and regularizes outputs with a dual-space objective combining latent and pixel-space losses. On the RadarHD benchmark, RadarSFD achieves state-of-the-art performance against baseline models. Qualitative results show recovery of fine walls and narrow gaps, and experiments across new environments confirm strong generalization. Ablation studies highlight the importance of pretrained initialization, radar BEV conditioning, and the dual-space loss. Together, these results establish a practical single-frame, no-SAR mmWave radar pipeline for dense point cloud perception in compact robotic systems.

cs.RO↗

arXiv2025

DeepSync: A Learning Framework for Pervasive Localization using Code Synchronization on Compressed Cellular Spectrum

Pervasive localization is essential for continuous tracking applications, yet existing solutions face challenges in balancing power consumption and accuracy. GPS, while precise, is impractical for continuous tracking of micro-assets due to high power requirements. Recent advances in non-linear compressed spectrum sensing offer low-power alternatives, but existing implementations achieve only coarse positioning through Received Signal Strength Indicator (RSSI) measurements. We present DeepSync, a deep learning framework that enables precise localization using compressed cellular spectrum. Our key technical insight lies in formulating sub-sample timing estimation as a template matching problem, solved through a novel architecture combining temporal CNN encoders for multi-frame processing with cross-attention mechanisms. The system processes non-linear inter-modulated spectrum through hierarchical feature extraction, achieving robust performance at SNR levels below -10dB -- a regime where conventional timing estimation fails. By integrating real cellular infrastructure data with physics-based ray-tracing simulations, DeepSync achieves 2.128-meter median accuracy while consuming significantly less power than conventional systems. Real-world evaluations demonstrate 10x improvement over existing compressed spectrum approaches, establishing a new paradigm for ultra-low-power localization.

eess.SY↗

arXiv2025

Spatial Audio Processing with Large Language Model on Wearable Devices

Integrating spatial context into large language models (LLMs) has the potential to revolutionize human-computer interaction, particularly in wearable devices. In this work, we present a novel system architecture that incorporates spatial speech understanding into LLMs, enabling contextually aware and adaptive applications for wearable technologies. Our approach leverages microstructure-based spatial sensing to extract precise Direction of Arrival (DoA) information using a monaural microphone. To address the lack of existing dataset for microstructure-assisted speech recordings, we synthetically create a dataset called OmniTalk by using the LibriSpeech dataset. This spatial information is fused with linguistic embeddings from OpenAI's Whisper model, allowing each modality to learn complementary contextual representations. The fused embeddings are aligned with the input space of LLaMA-3.2 3B model and fine-tuned with lightweight adaptation technique LoRA to optimize for on-device processing. SING supports spatially-aware automatic speech recognition (ASR), achieving a mean error of $25.72^\circ$-a substantial improvement compared to the 88.52$^\circ$ median error in existing work-with a word error rate (WER) of 5.3. SING also supports soundscaping, for example, inference how many people were talking and their directions, with up to 5 people and a median DoA error of 16$^\circ$. Our system demonstrates superior performance in spatial speech understanding while addressing the challenges of power efficiency, privacy, and hardware constraints, paving the way for advanced applications in augmented reality, accessibility, and immersive experiences.

cs.SD↗

arXiv2024

IMUOptimize: A Data-Driven Approach to Optimal IMU Placement for Human Pose Estimation with Transformer Architecture

This paper presents a novel approach for predicting human poses using IMU data, diverging from previous studies such as DIP-IMU, IMUPoser, and TransPose, which use up to 6 IMUs in conjunction with bidirectional RNNs. We introduce two main innovations: a data-driven strategy for optimal IMU placement and a transformer-based model architecture for time series analysis. Our findings indicate that our approach not only outperforms traditional 6 IMU-based biRNN models but also that the transformer architecture significantly enhances pose reconstruction from data obtained from 24 IMU locations, with equivalent performance to biRNNs when using only 6 IMUs. The enhanced accuracy provided by our optimally chosen locations, when coupled with the parallelizability and performance of transformers, provides significant improvements to the field of IMU-based pose estimation.

cs.LG↗

arXiv2023

Fast Localization and Tracking in City-Scale UWB Networks

Localization of networked nodes is an essential problem in emerging applications, including first-responder navigation, automated manufacturing lines, vehicular and drone navigation, asset navigation and tracking, Internet of Things and 5G communication networks. In this paper, we present Locate3D, a novel system for peer-to-peer node localization and orientation estimation in large networks. Unlike traditional range-only methods, Locate3D introduces angle-of-arrival (AoA) data as an added network topology constraint. The system solves three key challenges: it uses angles to reduce the number of measurements required by 4x and jointly use range and angle data for location estimation. We develop a spanning-tree approach for fast location updates, and to ensure the output graphs are rigid and uniquely realizable, even in occluded or weakly connected areas. Locate3D cuts down latency by up to 75% without compromising accuracy, surpassing standard range-only solutions. It has a 10.2 meters median localization error for large-scale networks (30,000 nodes, 15 anchors spread across 14km square) and 0.5 meters for small-scale networks (10 nodes).

cs.NI↗