SearcharxivSearch

arXiv subjects

Rachel Chan

Publications and source records attributed to Rachel Chan.

4 recordsLinked to original sources

Monocular passive event-based range-finding of airborne objects using the Scheimpflug principle

Passive 3D sensing is increasingly critical for early detection and tracking of small aerial vehicles (UAVs), where traditional active ranging can be tactically undesirable. We present SCHeimpflug for Optical Ranging TechnologY (SCHORTY), a single-aperture passive and active ranging architecture that exploits the Scheimpflug principle to encode range along a tilted object space plane by tilting the sensor relative to the imaging optics. SCHORTY requires only a one-time geometric calibration to map pixel coordinates to range and is inherently sensor and waveband agnostic. We implement SCHORTY using both a visible frame-based camera and an event-based camera (EBC) with closely matched pixel sizes for comparable horizontal resolutions and range binning. Controlled flights of an octocopter and a fixed-wing UAV equipped with GPS provide ground truth distances out to 1.1 km. Experimental results show that SCHORTY achieves deterministic range assignment limited primarily by the projected pixel size, which grows squared distance, while avoiding computationally intensive inverse reconstructions common in coded aperture and PSF engineered systems. In the EBC configuration, EBC-SCHORTY inherently suppresses static background and emphasizes motion, improving UAV detectability in cluttered natural scenes and under turbulence and motion blur. Additionally, we observe an asymmetric defocus blur about the object plane that depends on UAV trajectory, suggesting an extra cue for localization and trajectory inference. These results demonstrate SCHORTY as a practical and Size, Weight, and Power (SWaP) efficient passive ranging solution for medium-range UAV observation and motivate future integration with 2.5D/3D PSF engineering and event-based deconvolution to enhance 3D sensing performance.

physics.optics

A Hybrid GA LLM Framework for Structured Task Optimization

GA LLM is a hybrid framework that combines Genetic Algorithms with Large Language Models to handle structured generation tasks under strict constraints. Each output, such as a plan or report, is treated as a gene, and evolutionary operations like selection, crossover, and mutation are guided by the language model to iteratively improve solutions. The language model provides domain knowledge and creative variation, while the genetic algorithm ensures structural integrity and global optimization. GA LLM has proven effective in tasks such as itinerary planning, academic outlining, and business reporting, consistently producing well structured and requirement satisfying results. Its modular design also makes it easy to adapt to new tasks. Compared to using a language model alone, GA LLM achieves better constraint satisfaction and higher quality solutions by combining the strengths of both components.

cs.CL

Augmenting a Firefighters PPE -- Gas Mask SCBA

PPE (Personal Protective Equipment) has allowed firefighters to perform their everyday tasks without getting harmed since the mid 1800s. Now, the advancement of technology has given rise to the improvements of PPE. PPE can now include sensors to detect any number of environmental hazards (chemical, biological, temperature etc.). As the GT class of CS3750, we have decided to create a version of an interface design sensor that will help firefighters in two ways: navigation and communication. In order to augment a firefighter display when they are within a building, we chose to augment their SCBA (self-contained breathing apparatus). The gas mask will include a small screen that displays vital information directly towards the firefighter without need of any other support. We used the Google Glass to display vital information directly towards the eye in a minimalistic manner, while also augmenting that by adding LED lights to simulate someone calling their name or other auditory signals.While our prototype focuses on two main components of a firefighters search and rescue in a building, both of them combine to augment a firefighters display when searching throughout a building to help improve accuracy, speed and overall experience.

cs.HC

Raphtory: The temporal graph engine for Rust and Python

Raphtory is a platform for building and analysing temporal networks. The library includes methods for creating networks from a variety of data sources; algorithms to explore their structure and evolution; and an extensible GraphQL server for deployment of applications built on top. Raphtory's core engine is built in Rust, for efficiency, with Python interfaces, for ease of use. Raphtory is developed by network scientists, with a background in Physics, Applied Mathematics, Engineering and Computer Science, for use across academia and industry.

cs.SI