Owais Ahmed

In response to the growing challenges of manual labor and efficiency in warehouse operations, Amazon has embarked on a significant transformation by incorporating robotics to assist with various tasks. While a substantial number of robots have been successfully deployed for tasks such as item transportation within warehouses, the complex process of object picking from shelves remains a significant challenge. This project addresses the issue by developing an innovative robotic system capable of autonomously fulfilling a simulated order by efficiently selecting specific items from shelves. A distinguishing feature of the proposed robotic system is its capacity to navigate the challenge of uncertain object positions within each bin of the shelf. The system is engineered to autonomously adapt its approach, employing strategies that enable it to efficiently locate and retrieve the desired items, even in the absence of pre-established knowledge about their placements. Crucially, the robotic system operates with a paramount emphasis on autonomy. The intricate interplay of algorithms, control mechanisms, and sensor fusion empowers the robot to execute the entire object picking task without human intervention. This unfaltering commitment to autonomy is a pivotal step towards revolutionizing warehouse operations, potentially paving the way for unprecedented levels of efficiency and productivity. This project serves as a testament to the intersection of robotics, computer vision, and artificial intelligence in tackling a complex challenge within the realm of modern logistics. The envisioned robotic system represents a significant advancement in autonomous object-picking technology, holding the promise of transforming conventional warehousing practices. As the fusion of cutting-edge technology and logistical innovation unfolds, the outcomes of this endeavor have the potential to redefine the future of warehouse operations and automation within the industry.

This paper presents a real-time retail marketing system designed to intelligently detect product interaction and display contextual advertisements using low-cost hardware. Deployed across multiple retail locations, the system integrates Apriltag tracking, YOLOv8-based object segmentation, and Arduino-controlled lighting to detect when a noodle pack is picked from a shelf and trigger the corresponding video advertisement. Robustness to object occlusion and deformation is achieved via a frame-voting mechanism across consecutive detections. A custom YOLOv8-segmentation model was trained using a dataset of over 3500 segmented images of Knorr noodle packs across four flavors. The system is optimized to run on an Intel i5 7th generation mini PC and uses a dashboard for shelf configuration in offline environments. We report a mean Average Precision (mAP0.5) of 97.4% on validation data and demonstrate stable real-time performance in retail environments.

This paper presents a comprehensive review of the work "A Dynamic Algorithm for Approximate Flow Computations" by Professor Prabhakar and Professor Viswanathan. The original paper introduces an algorithm that improves the efficiency of reachability analysis in linear dynamical systems by dynamically determining time intervals and using polynomial approximations to maintain a specified error bound. This review summarizes the key contributions, including the use of Bernstein polynomials and error control mechanisms, and critically evaluates the algorithm’s performance, scalability, and limitations. Experimental insights, theoretical underpinnings, and potential directions for future research are discussed to highlight the broader implications of the work in formal verification and control systems.