ASAbhijeet Solanki

Abhijeet Solanki

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Abhijeet Solanki has a total of 16 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Vulnerability Assessment of Adversarial Image Attacks in a Miniaturized Autonomous Vehicle Testbed without Relying on Way-Point Information

    by Tahmid Hasan Sakib, Abhijeet Solanki, Wesam Al Amiri, Syed Rafay Hasan, Syed Ali Asad Rizvi, Terry N Guo

    This paper presents a real-time perception and control framework to assess adversarial image attacks on an autonomous vehicle testbed. The system integrates 2D LiDAR with a YOLOv5n-based detector in a multithreaded architecture to execute a staged multi-stop navigation routine with turns and visual alignment before each stop. Baseline experiments show reliable operation without requiring way-point information, with consistent stopping accuracy and path completion across repeated runs. Instead of using traditional digital overlays, we evaluated our testbed on adversarial patterns on a physical traffic sign. To assess vulnerability against adversarial images, specific adversarial variants of traffic signs were tested. This cause marked reductions in detection range and even missed recognitions. To restore performance without changing sensing, fusion, or control, we apply adversarial learning to fine-tune YOLOv5n. The retrained model recovers detection reliability across all variants and returns the vehicle to normal closed-loop behavior without relying on waypoint information.

  • Open Access

    RAID: Real-Time XAI-Guided Adversarial Attacks for Edge-Deployed Neural Networks

    by Abhijeet Solanki, Faiq Khalid, Drew Phelps, Syed Rafay Hasan, Terry Nan Guo, Uvais Qidwai

    Interpretability in deep learning is essential for understanding and improving the reliability of model behavior, particularly on edge devices deployed in safetycritical environments. Grad-CAM highlights class-relevant regions in input images and is widely used for model interpretability. However, its computational overhead limits real-time use. We address this by training a lightweight network to approximate Grad-CAM masks, which are then exploited to launch a novel, targeted sticker-pasting attack that manipulates model attention. To enable real-time adversarial testing, we train a Fully Convolutional Network (FCN-ResNet50) to segment and predict Grad-CAM hotspot regions on the fly, allowing seamless attack deployment in real-time. Our pipeline is deployed on NVIDIA Jetson Orin, where it achieves efficient segmentation and real-time attack generation with an average latency of 0.09s per attack. Experimental results across multiple stateof-the-art classifiers, including VGG16, ResNet50, ViT-B16, and EfficientNet, demonstrate significant drops in Top-1 accuracy (e.g., from 96.5% to 16.0% on ResNet50) under our targeted attacks, outperforming random sticker and patch attacks in effectiveness. Furthermore, our attack framework shows strong transferability across diverse model architectures, highlighting a broader vulnerability in saliency-guided interpretation. These findings raise concerns about the malicious use of XAI tools and emphasize the need for robust defense mechanisms in real-time, resource-constrained environments.

  • Open Access

    Input-Slice Attacks and Defense against Pseudo-Model-Parallelism-Based Collaborative DNN

    by Abhijeet Solanki, Parth Patel, Faiq Khalid, Syed Rafay Hasan, Uvais Qidwa

    Collaborative deep neural network (CDNN) inference across distributed edge devices provides scalability and resilience in mission-critical scenarios. Prior research has concentrated on adversarial attacks targeting intermediate feature maps within pseudo-model-parallelism Convolutional Neural Networks (ps-CDNNs), where feature-level disturbances are relatively localized and statistically detectable. In this paper, we introduce and examine a new type of input-level slice attack, in which adversaries manipulate input image segments before entering the network pipeline. We show that such attacks are much harder to detect with traditional statistical or filter-based defenses because of their subtler effects on aggregated feature distributions. To combat this, we propose an ensemble of lightweight machine learning detectors that utilize statistical signatures from feature maps, achieving an F1 score of up to 87% while keeping a low inference overhead. Our comparative analysis across attack surfaces (input vs. intermediate) and model architectures (VGG16, ResNet18) indicates that input-level perturbations not only bypass classic defenses but also more effectively degrade accuracy in multislice settings. These findings reveal a previously underexplored vulnerability in collaborative inference frameworks and emphasize the need for adaptive, model-agnostic detection strategies at the system edge to secure artificial intelligence (AI) models.

  • Open Access

    Blinded by the Beam: A Unified Real-Time Defense Against Laser-Based Attacks on Navigational Perception of Autonomous Vehicles

    by Abhijeet Solanki, Wesam Al Amiri, Terry N Guo, Syed Rafay Hasan

    Autonomous vehicles (AVs) rely on multi-sensor perception, particularly camera and LiDAR systems, for critical tasks such as object detection and scene understanding. However, recent studies have revealed that camera-based perception is especially vulnerable to laser-based adversarial attacks, including the Adversarial Laser Beam (AdvLB) and Adversarial Laser Spot (AdvLS). These attacks are impractical for real-time deployment as they are originally designed with iterative and dynamic configurations. In this work, we introduce and analyze a static-parameter attack strategy targeting the camera, which eliminates the need for iterative tuning while consistently degrading perception accuracy. This simplicity enables realtime feasibility, posing a practical threat to deployed AV systems. We characterize the visual artifacts introduced by such attacks using first-order statistical metrics, image filtering techniques, and machine learning-based anomaly detection. Our results show that even static laser interference can impair object recognition and disrupt camera-LiDAR sensor fusion. Experimental evaluations demonstrate that basic filters such as sharpening and Laplacian, combined with Hue, Saturation, and Value (HSV)-based contour extraction, effectively detect AdvLB artifacts. In contrast, the more subtle AdvLS perturbations require geometry-aware, adaptive detection methods. To address both, we propose and implement a lightweight, unified detection framework that integrates statistical and geometric cues. The system achieves over 90% detection accuracy with near real-time performance on the Jetson Orin Nano device.

  • Open Access

    Survey of Navigational Perception Sensors’ Security in Autonomous Vehicles

    by Abhijeet Solanki, Wesam Al Amiri, Marim Mahmoud, Blaine Swieder, Syed Rafay Hasan, Terry N. Guo

    The adoption of autonomous vehicles (AVs) significantly increased in recent years, playing a crucial role in enhancing transportation safety and efficiency, expanding mobility options, and reducing user costs. However, as AVs become more connected and automated, they also become more susceptible to malicious attacks, making it imperative to prioritize their security. Specifically, AVs are vulnerable to a range of cyber-threats, leading to a growing body of literature on potential cyberattacks targeting AVs. A key focus of existing research is understanding these threats and proposing mitigation strategies, particularly those aimed at attacks on navigational perception sensors, which are critical to AV navigation and decision-making. AVs rely on an array of navigational perception sensors, including LiDAR, cameras, RADAR, global positioning system (GPS) sensors, and ultrasonic sensors, to ensure secure and safe operation. However, adversarial manipulation of these sensors can compromise an AV’s navigational perception system, potentially leading to catastrophic consequences. While existing surveys provide valuable insights on AV cybersecurity challenges, several gaps remain unexplored, such as hardware security issues in AVs, analyzing sensor fusion attacks and evaluating practical defense implementations. To address these gaps, this paper makes a novel contribution by offering a comprehensive review of sensor fusion vulnerabilities that have not been sufficiently addressed in previous literature, specifically focusing on the compounded effects of sensor manipulation in AVs. In addition, this paper introduces a new classification of sensor vulnerabilities, highlighting practical defense mechanisms based on real-world simulation and testbed environments, a key contribution not explored in earlier works. We also introduce a detailed classification of sensor vulnerabilities, categorize the attacks that exploit these vulnerabilities, and critically assess existing countermeasures and defense mechanisms. Our study provides unique insights into the practical implementation of these countermeasures through simulations and testbeds, which have been underrepresented in prior research. This approach aims to support the development of more robust security solutions for AVs.