arXiv subjects

Kang G. Shin

Publications and source records attributed to Kang G. Shin.

At least 19 recordsLinked to original sources

arXiv2026

Context-Aware Intelligent Vehicles

Intelligent vehicles increasingly support adaptive applications beyond driving themselves, ranging from context-aware ADAS and automated driving to in-cabin monitoring and fleet management, all under tight requirements on accuracy, latency, cost, and reliability. Meeting these requirements is challenging because vehicles operate in complex, uncertain, and rapidly changing environments while running on resource-constrained computing platforms. This paper argues that context-situational factors that give meaning to sensor signals and constrain decisions-should be treated as a first-class principle for next-generation vehicle systems, and operationalized as a unified, shared state for learning, risk assessment, and closed-loop control across the software stack. We systematically review state-of-the- art (SOTA) context-aware methods spanning (i) environment understanding, (ii) planning and control, (iii) safety and security, and (iv) connected vehicles. Based on a trend analysis of context-aware design, we identify four key technical challenges in building a general contextual engine for future intelligent vehicles: multi-modal context fusion, temporal context modeling, handling rare events, and collaborative context sharing. We hope this survey will motivate the development of robust and efficient context-aware vehicle applications.

cs.RO↗

arXiv2026

Adversarial Calibration Attack on Autonomous Vehicles

Autonomous vehicles (AVs) rely on accurate camera-LiDAR calibration for multimodal sensor fusion. In practice, calibration can drift due to vibration, temperature variation, or minor sensor displacement, motivating online calibration algorithms that detect and correct misalignment at runtime while allowing the vehicle to continue operating without a factory visit. Existing AV attacks largely assume correct calibration. We instead identify online sensor calibration as a new attack plane. A corrupted calibration update can persist across subsequent fusion operations, causing system-wide errors that propagate from perception to planning and control. We present Adversarial Calibration Attack (ACA), the first physical attack against camera-LiDAR online calibration. Using a single adversarial poster, ACA first spoofs the miscalibration detector to trigger the calibration process and then steers the calibration estimator toward an incorrect transformation. A unified optimization jointly designs the poster's geometry and texture for both objectives. We evaluate ACA across benchmark datasets, simulation, and physical experiments. On benchmark datasets such as KITTI and nuScenes, ACA induces up to 33.9 degrees mean rotational calibration error, thereby severely degrading object detection. In the CARLA simulator, the attack causes a collision when the corrupted calibration is accepted in vulnerable scenarios crafted by the attacker. On a real Husky robot, a printed adversarial poster successfully reproduces the calibration error. These results demonstrate that online calibration is a practical and safety-critical attack surface for AVs.

cs.RO↗

arXiv2026

MM-BEV: Enhancing Timeliness by Computing Where and When it Matters

Multimodal bird's-eye-view (BEV) perception combines LiDAR depth accuracy with dense camera semantics, but its high computational cost and imperfect sensing conditions make real-time deployment challenging. Existing methods largely compress individual detectors and overlook three opportunities: structured sparsity within camera and LiDAR inputs, timing misalignment between modalities, and the fact that many detected objects do not affect the planner's immediate action. We present MM-BEV, a real-time multimodal BEV system guided by a simple principle: compute where and when it matters. MM-BEV divides perception into mandatory work for safety-critical objects within braking distance of the ego vehicle and with short time-to-collision (TTC), and optional work for less urgent regions. It prioritizes mandatory work and reduces or sheds optional work under tight compute budgets. MM-BEV integrates four mechanisms: (1) a criticality-ranked temporal ROI selector based on motion-extrapolated detections from prior frames; (2) sparse, ROI-aware feature extraction using shared-shape camera crops at context-adaptive resolution and ROI-aware LiDAR voxelization; (3) a latency-aware coordinator that adapts LiDAR sweeps, image resolution, and keyframes according to scene dynamics and TTC; and (4) an asynchronous scheduler that decouples sensing from inference and skips stale frames. On nuScenes, MM-BEV reduces inference latency by 1.96x and end-to-end latency by 2.93x, with no loss in geometry-critical recall and only a 0.2 percentage-point drop in safety-critical recall. On a Clearpath Husky A300 equipped with an Ouster-128 LiDAR, BEV cameras, and a Jetson AGX Orin, MM-BEV further reduces mean latency by 2.11x, demonstrating its potential for real-world autonomous systems.

cs.RO↗

arXiv2026

StateScribe: Towards Accessible Change Awareness Across Real-World Revisits

Real-world environments evolve continuously, yet blind and low-vision (BLV) individuals often have limited access to understanding how they change over time. Unexpected or relocated objects, layout modifications, and content updates (e.g., price changes) can introduce safety risks and cognitive burden. While existing visual assistive technologies can describe immediate surroundings, they operate as one-off interactions and lack mechanisms to surface meaningful changes across revisits. Informed by a survey of 33 BLV individuals, we develop StateScribe, a system that supports accessible awareness of real-world changes across revisits. StateScribe employs a dual-layer memory architecture that integrates episodic scene memory and object-centric temporal memory to enable scalable and structured change tracking. It provides both live descriptions of the current scene, and descriptions of what has changed, when and where it occurred across revisits, such as "The shop on your right has a "CLOSED" sign; it was open at this time last week.'' Our evaluation shows that StateScribe maintains high accuracy (F1-score=83.1%) across 11 revisits, while remaining low-latency (mean<1.54s) and memory-efficient (<54MB) across 110 revisits. A user study with nine BLV participants demonstrates that StateScribe improves change awareness across revisits in three real-world locations. Finally, we discuss implications for long-term AI-assisted companions that support broader change observation using multimodal sensing, extend beyond changes to other memory capabilities, and adapt to individual users, intents, and contexts.

cs.HC↗

arXiv2026

ES-Fuzz: Adaptive Modeling of MMIO Data Chunks for Firmware Fuzzing

Fuzzing has been widely used for testing embedded-system firmware in a fully rehosted environment without real peripherals, native system supports, or access to the source code and specifications. Some fuzzers emulate the MMIO behavior of missing peripherals based on the firmware binaries to boost code coverage. They emulate each individual MMIO read in the firmware with a fixed model. We find this ineffective when multiple MMIO reads collectively retrieve a data chunk, which greatly impedes the coverage growth. We propose ES-Fuzz to overcome this coverage bottleneck by adaptively modeling MMIO data chunks. ES-Fuzz runs alongside a given fuzzer and starts a new run when the fuzzer's coverage stagnates. In each run, it analyzes a high-coverage test case to infer new MMIO data-chunk models that unlock additional execution paths. We have implemented ES-Fuzz on Fuzzware and evaluated it on 24 popular firmware binaries. ES-Fuzz boosts Fuzzware's coverage by up to 68% in 11 -- and triggers additional bugs in 5 -- of them without degrading the coverage in the remainder. Its models describe a wide range of MMIO data chunks and the firmware's use of each across various contexts.

cs.CR↗

arXiv2026

Feeling the Space: Egomotion-Aware Video Representation for Efficient and Accurate 3D Scene Understanding

Recent Multimodal Large Language Models (MLLMs) have shown high potential for spatial reasoning within 3D scenes. However, they typically rely on computationally expensive 3D representations like point clouds or reconstructed Bird's-Eye View (BEV) maps, or lack physical grounding to resolve ambiguities in scale and size. This paper significantly enhances MLLMs with egomotion modality data, captured by Inertial Measurement Units (IMUs) concurrently with the video. In particular, we propose a novel framework, called Motion-MLLM, introducing two key components: (1) a cascaded motion-visual keyframe filtering module that leverages both IMU data and visual features to efficiently select a sparse yet representative set of keyframes, and (2) an asymmetric cross-modal fusion module where motion tokens serve as intermediaries that channel egomotion cues and cross-frame visual context into the visual representation. By grounding visual content in physical egomotion trajectories, Motion-MLLM can reason about absolute scale and spatial relationships across the scene. Our extensive evaluation shows that Motion-MLLM makes significant improvements in various tasks related to 3D scene understanding and spatial reasoning. Compared to state-of-the-art (SOTA) methods based on video frames and explicit 3D data, Motion-MLLM achieves competitive accuracy while running $1.30\times$ and $1.61\times$ faster, respectively.

cs.CV↗

arXiv2026

Enhancing Predictability of Multi-Tenant DNN Inference for Autonomous Vehicles' Perception

Autonomous vehicles (AVs) rely on sensors and deep neural networks (DNNs) to perceive their surrounding environment and make maneuver decisions in real time. However, achieving real-time DNN inference in the AV's perception pipeline is challenging due to the large gap between the computation requirement and the AV's limited resources. Most, if not all, of existing studies focus on optimizing the DNN inference time to achieve faster perception by compressing the DNN model with pruning and quantization. In contrast, we present a Predictable Perception system with DNNs (PP-DNN) that reduce the amount of image data to be processed while maintaining the same level of accuracy for multi-tenant DNNs by dynamically selecting critical frames and regions of interest (ROIs). PP-DNN is based on our key insight that critical frames and ROIs for AVs vary with the AV's surrounding environment. However, it is challenging to identify and use critical frames and ROIs in multi-tenant DNNs for predictable inference. Given image-frame streams, PP-DNN leverages an ROI generator to identify critical frames and ROIs based on the similarities of consecutive frames and traffic scenarios. PP-DNN then leverages a FLOPs predictor to predict multiply-accumulate operations (MACs) from the dynamic critical frames and ROIs. The ROI scheduler coordinates the processing of critical frames and ROIs with multiple DNN models. Finally, we design a detection predictor for the perception of non-critical frames. We have implemented PP-DNN in an ROS-based AV pipeline and evaluated it with the BDD100K and the nuScenes dataset. PP-DNN is observed to significantly enhance perception predictability, increasing the number of fusion frames by up to 7.3x, reducing the fusion delay by >2.6x and fusion-delay variations by >2.3x, improving detection completeness by 75.4% and the cost-effectiveness by up to 98% over the baseline.

cs.CV↗

arXiv2025

Power-Efficient Autonomous Mobile Robots

This paper presents pNav, a novel power-management system that significantly enhances the power/energy-efficiency of Autonomous Mobile Robots (AMRs) by jointly optimizing their physical/mechanical and cyber subsystems. By profiling AMRs' power consumption, we identify three challenges in achieving CPS (cyber-physical system) power-efficiency that involve both cyber (C) and physical (P) subsystems: (1) variabilities of system power consumption breakdown, (2) environment-aware navigation locality, and (3) coordination of C and P subsystems. pNav takes a multi-faceted approach to achieve power-efficiency of AMRs. First, it integrates millisecond-level power consumption prediction for both C and P subsystems. Second, it includes novel real-time modeling and monitoring of spatial and temporal navigation localities for AMRs. Third, it supports dynamic coordination of AMR software (navigation, detection) and hardware (motors, DVFS driver) configurations. pNav is prototyped using the Robot Operating System (ROS) Navigation Stack, 2D LiDAR, and camera. Our in-depth evaluation with a real robot and Gazebo environments demonstrates a >96% accuracy in predicting power consumption and a 38.1% reduction in power consumption without compromising navigation accuracy and safety.

cs.RO↗

arXiv2025

Ads that Talk Back: Implications and Perceptions of Injecting Personalized Advertising into LLM Chatbots

Recent advances in large language models (LLMs) have enabled the creation of highly effective chatbots. However, the compute costs of widely deploying LLMs have raised questions about profitability. Companies have proposed exploring ad-based revenue streams for monetizing LLMs, which could serve as the new de facto platform for advertising. This paper investigates the implications of personalizing LLM advertisements to individual users via a between-subjects experiment with 179 participants. We developed a chatbot that embeds personalized product advertisements within LLM responses, inspired by similar forays by AI companies. The evaluation of our benchmarks showed that ad injection only slightly impacted LLM performance, particularly response desirability. Results revealed that participants struggled to detect ads, and even preferred LLM responses with hidden advertisements. Rather than clicking on our advertising disclosure, participants tried changing their advertising settings using natural language queries. We created an advertising dataset and an open-source LLM, Phi-4-Ads, fine-tuned to serve ads and flexibly adapt to user preferences.

cs.HC↗

arXiv2025

Navigating Cookie Consent Violations Across the Globe

Online services provide users with cookie banners to accept/reject the cookies placed on their web browsers. Despite the increased adoption of cookie banners, little has been done to ensure that cookie consent is compliant with privacy laws around the globe. Prior studies have found that cookies are often placed on browsers even after their explicit rejection by users. These inconsistencies in cookie banner behavior circumvent users' consent preferences and are known as cookie consent violations. To address this important problem, we propose an end-to-end system, called ConsentChk, that detects and analyzes cookie banner behavior. ConsentChk uses a formal model to systematically detect and categorize cookie consent violations. We investigate eight English-speaking regions across the world, and analyze cookie banner behavior across 1,793 globally-popular websites. Cookie behavior, cookie consent violation rates, and cookie banner implementations are found to be highly dependent on region. Our evaluation reveals that consent management platforms (CMPs) and website developers likely tailor cookie banner configurations based on their (often incorrect) interpretations of regional privacy laws. We discuss various root causes behind these cookie consent violations. The resulting implementations produce misleading cookie banners, indicating the prevalence of inconsistently implemented and enforced cookie consent between various regions.

cs.CR↗

arXiv2025

Keypoints as Dynamic Centroids for Unified Human Pose and Segmentation

The dynamic movement of the human body presents a fundamental challenge for human pose estimation and body segmentation. State-of-the-art approaches primarily rely on combining keypoint heatmaps with segmentation masks but often struggle in scenarios involving overlapping joints or rapidly changing poses during instance-level segmentation. To address these limitations, we propose Keypoints as Dynamic Centroid (KDC), a new centroid-based representation for unified human pose estimation and instance-level segmentation. KDC adopts a bottom-up paradigm to generate keypoint heatmaps for both easily distinguishable and complex keypoints and improves keypoint detection and confidence scores by introducing KeyCentroids using a keypoint disk. It leverages high-confidence keypoints as dynamic centroids in the embedding space to generate MaskCentroids, allowing for swift clustering of pixels to specific human instances during rapid body movements in live environments. Our experimental evaluations on the CrowdPose, OCHuman, and COCO benchmarks demonstrate KDC's effectiveness and generalizability in challenging scenarios in terms of both accuracy and runtime performance. The implementation is available at: https://sites.google.com/view/niazahmad/projects/kdc.

cs.CV↗

arXiv2024

Steward: Natural Language Web Automation

Recently, large language models (LLMs) have demonstrated exceptional capabilities in serving as the foundation for AI assistants. One emerging application of LLMs, navigating through websites and interacting with UI elements across various web pages, remains somewhat underexplored. We introduce Steward, a novel LLM-powered web automation tool designed to serve as a cost-effective, scalable, end-to-end solution for automating web interactions. Traditional browser automation frameworks like Selenium, Puppeteer, and Playwright are not scalable for extensive web interaction tasks, such as studying recommendation algorithms on platforms like YouTube and Twitter. These frameworks require manual coding of interactions, limiting their utility in large-scale or dynamic contexts. Steward addresses these limitations by integrating LLM capabilities with browser automation, allowing for natural language-driven interaction with websites. Steward operates by receiving natural language instructions and reactively planning and executing a sequence of actions on websites, looping until completion, making it a practical tool for developers and researchers to use. It achieves high efficiency, completing actions in 8.52 to 10.14 seconds at a cost of $0.028 per action or an average of $0.18 per task, which is further reduced to 4.8 seconds and $0.022 through a caching mechanism. It runs tasks on real websites with a 40% completion success rate. We discuss various design and implementation challenges, including state representation, action sequence selection, system responsiveness, detecting task completion, and caching implementation.

cs.AI↗

arXiv2024

Analyzing Privacy Implications of Data Collection in Android Automotive OS

Modern vehicles have become sophisticated computation and sensor systems, as evidenced by advanced driver assistance systems, in-car infotainment, and autonomous driving capabilities. They collect and process vast amounts of data through various embedded subsystems. One significant player in this landscape is Android Automotive OS (AAOS), which has been integrated into over 100M vehicles and has become a dominant force in the in-vehicle infotainment market. With this extensive data collection, privacy has become increasingly crucial. The volume of data gathered by these systems raises questions about how this information is stored, used, and protected, making privacy a critical issue for manufacturers and consumers. However, very little has been done on vehicle data privacy. This paper focuses on the privacy implications of AAOS, examining the exact nature and scope of data collection and the corresponding privacy policies from the original equipment manufacturers (OEMs). We develop a novel automotive privacy analysis tool called PriDrive which employs three methodological approaches: network traffic inspection, and both static and dynamic analyses of Android images using rooted emulators from various OEMs. These methodologies are followed by an assessment of whether the collected data types were properly disclosed in OEMs and 3rd party apps' privacy policies (to identify any discrepancies or violations). Our evaluation on three different OEM platforms reveals that vehicle speed is collected at a sampling rate of roughly 25 Hz. Other properties such as model info, climate & AC, and seat data are collected in a batch 30 seconds into vehicle startup. In addition, several vehicle property types were collected without disclosure in their respective privacy policies. For example, OEM A's policies only covers 110 vehicle properties or 13.02% of the properties found in our static analysis.

cs.CR↗

arXiv2024

Achieving the Safety and Security of the End-to-End AV Pipeline

In the current landscape of autonomous vehicle (AV) safety and security research, there are multiple isolated problems being tackled by the community at large. Due to the lack of common evaluation criteria, several important research questions are at odds with one another. For instance, while much research has been conducted on physical attacks deceiving AV perception systems, there is often inadequate investigations on working defenses and on the downstream effects of safe vehicle control. This paper provides a thorough description of the current state of AV safety and security research. We provide individual sections for the primary research questions that concern this research area, including AV surveillance, sensor system reliability, security of the AV stack, algorithmic robustness, and safe environment interaction. We wrap up the paper with a discussion of the issues that concern the interactions of these separate problems. At the conclusion of each section, we propose future research questions that still lack conclusive answers. This position article will serve as an entry point to novice and veteran researchers seeking to partake in this research domain.

cs.RO↗

arXiv2024

Analysis and Prevention of MCAS-Induced Crashes

Semi-autonomous (SA) systems face the challenge of determining which source to prioritize for control, whether it's from the human operator or the autonomous controller, especially when they conflict with each other. While one may design an SA system to default to accepting control from one or the other, such design choices can have catastrophic consequences in safety-critical settings. For instance, the sensors an autonomous controller relies upon may provide incorrect information about the environment due to tampering or natural fault. On the other hand, the human operator may also provide erroneous input. To better understand the consequences and resolution of this safety-critical design choice, we investigate a specific application of an SA system that failed due to a static assignment of control authority: the well-publicized Boeing 737-MAX Maneuvering Characteristics Augmentation System (MCAS) that caused the crashes of Lion Air Flight 610 and Ethiopian Airlines Flight 302. First, using a representative simulation, we analyze and demonstrate the ease by which the original MCAS design could fail. Our analysis reveals the most robust public analysis of aircraft recoverability under MCAS faults, offering bounds for those scenarios beyond the original crashes. We also analyze Boeing's updated MCAS and show how it falls short of its intended goals and continues to rely upon on a fault-prone static assignment of control priority. Using these insights, we present Semi-Autonomous MCAS (SA-MCAS), a new MCAS that both meets the intended goals of MCAS and avoids the failure cases that plague both MCAS designs. We demonstrate SA-MCAS's ability to make safer and timely control decisions of the aircraft, even when the human and autonomous operators provide conflicting control inputs.

eess.SY↗

arXiv2023

End-to-End Asynchronous Traffic Scheduling in Converged 5G and Time-Sensitive Networks

As required by Industry 4.0, companies will move towards flexible and individual manufacturing. To succeed in this transition, convergence of 5G and time-sensitive networks (TSN) is the most promising technology and has thus attracted considerable interest from industry and standardization groups. However, the delay and jitter of end-to-end (e2e) transmission will get exacerbated if the transmission opportunities are missed in TSN due to the 5G transmission jitter and the clock skew between the two network systems. To mitigate this phenomenon, we propose a novel asynchronous access mechanism (AAM) that isolates the jitter only in the 5G system and ensures zero transmission jitter in TSN. We then exploit AAM to develop an e2e asynchronous traffic scheduling model for coordinated allocation of resources for 5G and TSN to provide e2e transmission delay guarantees for time-critical flows. The results of our extensive simulation of AAM on OMNET++ corroborate the superior performance of AAM and the scheduling model.

cs.NI↗

arXiv2023

A Switch Architecture for Time-Triggered Transmission with Best-Effort Delivery

In Time-Triggered (TT) or time-sensitive networks, the transmission of a TT frame is required to be scheduled at a precise time instant for industrial distributed real-time control systems. Other (or {\em best-effort} (BE)) frames are forwarded in a BE manner. Under this scheduling strategy, the transmission of a TT frame must wait until its scheduled instant even if it could have been transmitted sooner. On the other hand, BE frames are transmitted whenever possible but may miss deadlines or may even be dropped due to congestion. As a result, TT transmission and BE delivery are incompatible with each other. To remedy this incompatibility, we propose a synergistic switch architecture (SWA) for TT transmission with BE delivery to dynamically improve the end-to-end (e2e) latency of TT frames by opportunistically exploiting BE delivery. Given a TT frame, the SWA generates and transmits a cloned copy with BE delivery. The first frame arriving at the receiver device is delivered with a configured jitter and the other copy ignored. So, the SWA achieves shorter latency and controllable jitter, the best of both worlds. We have implemented SWA using FPGAs in an industry-strength TT switches and used four test scenarios to demonstrate SWA's improvements of e2e latency and controllable jitter over the state-of-the-art TT transmission scheme.

cs.NI↗

arXiv2023

Eye-Shield: Real-Time Protection of Mobile Device Screen Information from Shoulder Surfing

People use mobile devices ubiquitously for computing, communication, storage, web browsing, and more. As a result, the information accessed and stored within mobile devices, such as financial and health information, text messages, and emails, can often be sensitive. Despite this, people frequently use their mobile devices in public areas, becoming susceptible to a simple yet effective attack, shoulder surfing. Shoulder surfing occurs when a person near a mobile user peeks at the user's mobile device, potentially acquiring passcodes, PINs, browsing behavior, or other personal information. We propose Eye-Shield, a solution to prevent shoulder surfers from accessing or stealing sensitive on-screen information. Eye-Shield is designed to protect all types of on-screen information in real time, without any serious impediment to users' interactions with their mobile devices. Eye-Shield generates images that appear readable at close distances, but appear blurry or pixelated at farther distances and wider angles. It is capable of protecting on-screen information from shoulder surfers, operating in real time, and being minimally intrusive to the intended users. Eye-Shield protects images and text from shoulder surfers by reducing recognition rates to 24.24% and 15.91%. Our implementations of Eye-Shield, with frame rates of 24 FPS for Android and 43 FPS for iOS, effectively work on screen resolutions as high as 1440x3088. Eye-Shield also incurs acceptable memory usage, CPU utilization, and energy overhead. Finally, our MTurk and in-person user studies indicate that Eye-Shield protects on-screen information without a large usability cost for privacy-conscious users.

cs.CR↗