Josiah Hester

The erasure of Indigenous cultures is a painfully enduring legacy of U.S. imperial history, producing disparities in culturally-relevant curricula, thus barring Indigenous students from resonant educational experiences. In Hawaiʻi, where Hawaiian instruction was banned for 91 years, culturally-relevant curricula gaps are most severe in Hawaiian-language education and technical fields like Computer Science (CS). Large Language Models (LLMs) are positioned to address these disparities, yet raise concerns of cultural misrepresentation and data sovereignty. We ask: how can LLMs serve educators of varying familiarity with Hawaiian culture while safeguarding Hawaiian knowledge? Addressing this question, we present Kumu Connect, an LLM-assisted culturally-revitalizing CS lesson plan generator, co-designed with Native Hawaiian educators. This work contributes design requirements, a situated case study, and “future imaginaries” paired with design provocations to highlight how community-engaged development can inform culturally-authentic and pedagogically useful AI tools that address CS and culture knowledge gaps and uphold Indigenous data sovereignty.

Climate change, urbanization, and pollution threaten ecosystems and the treaty-guaranteed rights of Native Nations in the Great Lakes region. Tools that support culturally relevant implementation of policy and meaningful access to environmental data for sentinel species like Manoomin, wild rice, can help uphold treaty rights and ensure environmental stewardship. This paper presents Noondawind, an interactive data platform co-designed with Ojibwe partners to support community members and Tribal staff in interpreting and acting on environmental data and policy resources. We engaged in a participatory design process informed by and deeply integrated with Ojibwe worldviews. Our results highlight how participatory and culturally relevant co-design approaches can enhance environmental governance, support data sovereignty, and foster engagement with environmental data. We offer design implications and lessons learned for projects developing tools in partnership with Indigenous communities. These findings contribute to the growing field of Indigenous HCI and social justice literature in HCI.

Although generative AI is being deployed into classrooms with promises of aiding teachers, educators caution that these tools can have unintended pedagogical repercussions, including cultural misrepresentation and bias. These concerns are heightened in low-resource language and Indigenous education settings, where AI systems frequently underperform. We investigate these challenges in Hawai‘i, where public schools operate under a statewide mandate to integrate Hawaiian language and culture into education. Through four co-design workshops with 22 public school educators, we surfaced concerns about using generative AI in educational settings, particularly around cultural misrepresentation, and corresponding designs for auditing tools that address these issues. We find that educators envision tools grounded in specific Hawaiian cultural values and practices, such as tracing the genealogy of knowledge in source materials. Building on these insights, we conceptualize AI auditing as a community-oriented process rather than the work of isolated individuals, and discuss implications for designing auditing tools.

MicroGreen is a design-time framework that enables carbon-aware design for edge ML system. This artifact contains code and instructions on how to reproduce the results presented in the MicroGreen MobiSys 2026 paper.

MicroGreen is a design-time framework that enables carbon-aware design for edge ML system. This artifact contains code and instructions on how to reproduce the results presented in the MicroGreen MobiSys 2026 paper.

Living organisms achieve adaptive actuation through the seamless integration of neural motor control circuitry and proprioceptive feedback. While biohybrid robotics aims to replicate these capabilities by merging engineered muscle with synthetic scaffolds, the field remains limited by interfaces that lack the efficiency and closed-loop regulation of natural neuromuscular systems. Here, we introduce a biohybrid muscle actuator system featuring a bioelectronic interface based on soft poly(3,4-ethylenedioxythiophene) (PEDOT) fibers for stimulation and sensing. These fibers conformally couple to muscle tissues, eliciting robust contractions at voltages as low as 1 V-requiring ultra-low power (0.376 ± 0.034 mW) and preserving long-term tissue viability. By leveraging the independent addressability of these fibers, we demonstrate selective actuation of individual muscle units to achieve precise spatiotemporal control of a two-muscle-powered walking biohybrid robot, reaching a locomotion speed of 5.43 ± 0.79 mm/min. When configured as strain sensors, the fibers exhibit a high gauge factor of 155.45 ± 6.59 and resolve contractile displacements within tens of micrometers. We demonstrate that this sensing modality can be integrated into a closed-loop controller to autonomously modulate stimulation based on real-time feedback, significantly mitigating muscle fatigue (p = 0.038) during continuous operation. This work establishes a versatile platform for efficient actuation and intrinsic feedback sensing, providing a blueprint for efficient, autonomous, and adaptive biohybrid machines.

We introduce SpiderCam, an FPGA-based snapshot depth-from-defocus camera which produces 480x400 sparse depth maps in real-time at 32.5 FPS over a working range of 52 cm while consuming 624 mW of power in total. SpiderCam comprises a custom camera that simultaneously captures two differently focused images of the same scene, processed with a SystemVerilog implementation of depth from differential defocus (DfDD) on a low-power FPGA. To achieve state-of-the-art power consumption, we present algorithmic improvements to DfDD that overcome challenges caused by low-power sensors, and design a memory-local implementation for streaming depth computation on a device that is too small to store even a single image pair. We report the first sub-Watt total power measurement for passive FPGA-based 3D cameras in the literature.

We introduce SpiderCam, an FPGA-based snapshot depth-from-defocus camera which produces 480x400 sparse depth maps in real-time at 32.5 FPS over a working range of 52 cm while consuming 624 mW of power in total. SpiderCam comprises a custom camera that simultaneously captures two differently focused images of the same scene, processed with a SystemVerilog implementation of depth from differential defocus (DfDD) on a low-power FPGA. To achieve state-of-the-art power consumption, we present algorithmic improvements to DfDD that overcome challenges caused by low-power sensors, and design a memory-local implementation for streaming depth computation on a device that is too small to store even a single image pair. We report the first sub-Watt total power measurement for passive FPGA-based 3D cameras in the literature.

Although generative AI is being deployed into classrooms with promises of aiding teachers, educators caution that these tools can have unintended pedagogical repercussions, including cultural misrepresentation and bias. These concerns are heightened in low-resource language and Indigenous education settings, where AI systems frequently underperform. We investigate these challenges in Hawai`i, where public schools operate under a statewide mandate to integrate Hawaiian language and culture into education. Through four co-design workshops with 22 public school educators, we surfaced concerns about using generative AI in educational settings, particularly around cultural misrepresentation, and corresponding designs for auditing tools that address these issues. We find that educators envision tools grounded in specific Hawaiian cultural values and practices, such as tracing the genealogy of knowledge in source materials. Building on these insights, we conceptualize AI auditing as a community-oriented process rather than the work of isolated individuals, and discuss implications for designing auditing tools.

Although generative AI is being deployed into classrooms with promises of aiding teachers, educators caution that these tools can have unintended pedagogical repercussions, including cultural misrepresentation and bias. These concerns are heightened in low-resource language and Indigenous education settings, where AI systems frequently underperform. We investigate these challenges in Hawai`i, where public schools operate under a statewide mandate to integrate Hawaiian language and culture into education. Through four co-design workshops with 22 public school educators, we surfaced concerns about using generative AI in educational settings, particularly around cultural misrepresentation, and corresponding designs for auditing tools that address these issues. We find that educators envision tools grounded in specific Hawaiian cultural values and practices, such as tracing the genealogy of knowledge in source materials. Building on these insights, we conceptualize AI auditing as a community-oriented process rather than the work of isolated individuals, and discuss implications for designing auditing tools.

Amid escalating ecological challenges, Human-Computer Interaction (HCI) researchers have begun adopting More-than-Human design (MtHD) approaches as a means of reimagining and strengthening the bonds between the human and non-human world. Taking a MtHD approach, this work investigates how loofah, a plant-based, biodegradable material, can be used to foster ecological awareness in everyday life. We present a material exploration of loofah, including initial encounters with loofah as a uniquely structural material, a design space focused on how loofah can be combined with various indicators that respond to different environmental factors like temperature, UV radiation, pH of water and soil, the Iron and moisture in soil and utilizing loofah as a substrate for plant growth. Based on this design space, we create two wearables: a hat and a glove. These artifacts incorporate environmental sensing capabilities and host live microgreens, highlighting loofah‘s potential as both an interface and habitat. Through a 15-day autoethnographic journaling process by the first and second authors, we reflect on the embodied experiences of “wearing” environmental change and cultivating on-body ecological practices. This pictorial contributes to the HCI community by introducing a biomaterial-based embodied interaction to provoke reflections on the relationships between materials, non-human forms, and the environment, while integrating functional considerations into MtHD.

Amid escalating ecological challenges, Human-Computer Interaction (HCI) researchers have begun adopting More-than-Human design (MtHD) approaches as a means of reimagining and strengthening the bonds between the human and non-human world. Taking a MtHD approach, this work investigates how loofah, a plant-based, biodegradable material, can be used to foster ecological awareness in everyday life. We present a material exploration of loofah, including initial encounters with loofah as a uniquely structural material, a design space focused on how loofah can be combined with various indicators that respond to different environmental factors like temperature, UV radiation, pH of water and soil, the Iron and moisture in soil and utilizing loofah as a substrate for plant growth. Based on this design space, we create two wearables: a hat and a glove. These artifacts incorporate environmental sensing capabilities and host live microgreens, highlighting loofah‘s potential as both an interface and habitat. Through a 15-day autoethnographic journaling process by the first and second authors, we reflect on the embodied experiences of “wearing” environmental change and cultivating on-body ecological practices. This pictorial contributes to the HCI community by introducing a biomaterial-based embodied interaction to provoke reflections on the relationships between materials, non-human forms, and the environment, while integrating functional considerations into MtHD.

BACKGROUND: Digital behaviors such as sleep, social interaction, and productivity reflect how individuals' structure daily life. Among university students, online activity patterns mirror academic schedules, social rhythms, and lifestyle habits, with disruptions linked to sleep, stress, and well-being. Existing approaches-including wearables, apps, and surveys-yield useful insights but depend on self-report or active participation, limiting long-term adherence. Passive sensing of network traffic offers a scalable alternative for unobtrusive capture of smartphone usage patterns that preserves privacy. OBJECTIVE: This study evaluated whether encrypted smartphone network traffic, collected via a standard virtual private network (VPN), can capture patterns of digital behavior. We assessed feasibility (sustained data capture) and acceptability (usability, burden, and privacy perceptions) and examined whether traffic-derived features reveal aspects of digital behavior-including timing, intensity, and regularity-relevant to health and daily functioning. METHODS: We conducted a two-week prospective observational study at New York University. Participants installed the WireGuard VPN client on personal smartphones, enabling passive capture of encrypted network traffic. Feasibility was assessed using a mixed-methods approach combining quantitative measures of user retention and data coverage with qualitative analysis of semi-structured exit interviews. Acceptability was evaluated using the System Usability Scale (SUS), NASA Task Load Index (NASA-TLX), and qualitative interview analysis. Exploratory analyses visualized traffic-derived features in relation to digital activity patterns. RESULTS: Thirty-eight students consented to participate, of whom 29 contributed valid network traffic data and formed the analytic cohort. Within this cohort, 93% of participants (27/29; Wilson 95% CI: 78-98%) contributed at least five days of monitoring, corresponding to 71% retention relative to all consented participants (27/38; Wilson 95% CI: 55-83%). Mean data coverage within the analytic cohort (N=27) was 74.1% (median 77.1%; bootstrap 95% CI: 66.3-81.4%). These participants contributed an average of 311.6 hours of monitored traffic (approximately 13 days, SD 3.5), ranging from 121 to 496 hours. Acceptability outcomes were evaluated among participants completing the exit survey and interview. Usability ratings were high (mean SUS score = 78) and perceived workload low (NASA-TLX scores minimal). Participants described the system as easy to install, unobtrusive, and generally trustworthy, though some reported temporarily disabling the VPN during activities they considered private. No inferential statistical tests were conducted; analyses were descriptive. Exploratory analyses indicated that traffic-derived features reflected daily digital activity rhythms and revealed distinctive lifestyle patterns, including gaming and irregular late-night food delivery use. CONCLUSIONS: VPN-based monitoring of encrypted smartphone traffic was feasible and acceptable, enabling sustained passive data collection with minimal burden. This approach shows promise as a scalable, device-agnostic method for digital phenotyping that captures fine-grained behavioral rhythms while preserving privacy. With broader validation, this technique could expand the toolkit for studying health and well-being in everyday life. CLINICALTRIAL: This study was not registered as a clinical trial because it did not involve randomization.

This is an erratum for the article “The Devil You Know”: Barriers and Opportunities for Co-Designing Microclimate Sensors, A Case Study of Manoomi” published in ACM J. Comput. Sustain. Soc. 2, 3, Article 39 (September 2024), 30 pages.

Multi-agent systems (MAS) are emerging as promising socio-collaborative companions for emotional and cognitive support. However, existing systems frequently suffer from persona collapse, where agents revert to generic, homogenized assistant behaviors, and social sycophancy, where agents produce redundant, non-constructive dialogue. We propose MASCOT, a multi-agent framework for multi-perspective socio-collaborative companions. MASCOT introduces a novel bi-level optimization strategy to harmonize individual and collective behaviors: 1) Persona-Aware Behavioral Alignment, an RLAIF-driven pipeline that fine-tunes individual agents for agent-specific identities; and 2) Collaborative Dialogue Optimization, a group-level adaptation process that promotes complementary, diverse, and productive discourse. We evaluate MASCOT using human-grounded contexts drawn across both in-domain and out-of-domain (OOD) settings against state-of-the-art baselines. MASCOT improves persona consistency by up to +14.1 and social contribution by up to +10.6. A broad evaluation suite, including human evaluation, multiple LLM judges, three-way comparisons, and automatic metrics, further shows that MASCOT produces more role-consistent and less redundant multi-agent dialogue.

Multi-agent systems (MAS) are emerging as promising socio-collaborative companions for emotional and cognitive support. However, existing systems frequently suffer from persona collapse, where agents revert to generic, homogenized assistant behaviors, and social sycophancy, where agents produce redundant, non-constructive dialogue. We propose MASCOT, a multi-agent framework for multi-perspective socio-collaborative companions. MASCOT introduces a novel bi-level optimization strategy to harmonize individual and collective behaviors: 1) Persona-Aware Behavioral Alignment, an RLAIF-driven pipeline that fine-tunes individual agents for agent-specific identities; and 2) Collaborative Dialogue Optimization, a group-level adaptation process that promotes complementary, diverse, and productive discourse. We evaluate MASCOT using human-grounded contexts drawn across both in-domain and out-of-domain (OOD) settings against state-of-the-art baselines. MASCOT improves persona consistency by up to +14.1 and social contribution by up to +10.6. A broad evaluation suite, including human evaluation, multiple LLM judges, three-way comparisons, and automatic metrics, further shows that MASCOT produces more role-consistent and less redundant multi-agent dialogue.

Shared experiences are fundamental to social connection, yet media consumption is increasingly solitary. While AI companions offer real-time reactions and emotional regulation, existing systems either rely on single-agent designs or lack the social awareness and multi-party interaction required to replicate authentic group dynamics. We present CompanionCast, a general framework for orchestrating multiple specialized AI agents as social collaborators within a live shared context. CompanionCast integrates multimodal event detection, rolling context caching for improved grounding, and spatial audio to enhance co-presence. We validate CompanionCast through sports viewing, a domain with rich dynamics and strong social traditions. Pilot studies with soccer fans demonstrate that CompanionCast significantly improves perceived social presence and emotional sharing compared to solitary viewing. We conclude by discussing implications and open challenges for multi-agent systems as social collaborators in shared experiences.

Understanding in real time how we move, operate, and interact within living and working spaces is crucial to applications in smart buildings, elder care, and the automation industry. Myriads of systems have used complex electronics to capture this information, yet they often rely on independent sensors or devices that require active power and circuitry, also raising privacy concerns when using cameras or microphones. These solutions are often difficult to setup, have incompatible ecosystems, and require extra cost when upgrading already existing devices. In this paper, we propose SoundOff, a system that deploys ultra-low-cost, easily manufacturable, passive-ultrasound-emitting tags in any indoor environment (i.e., door knobs, toilet lids, cabinets, faucets, and windows), where the movement of this furniture causes a unique ultrasonic emission identifiable by a wearable device worn by users. These tags generate ultrasound signals during everyday interaction, above the range of human hearing, making them non-intrusive and non-invasive. This electronics-free, zero-infrastructure solution provides a scalable way to instrument spaces. Through a series of performance evaluations, we show that the ultrasonic emissions generated by SoundOff with different geometrical designs are not only robust to various environmental noises but also easily distinguishable from each other. Through physics-based modeling, we demonstrate how to systematically generate thousands of designs with unique and easily distinguishable ultrasonic emissions, enabling a wide range of interactions that can be mapped to custom automation systurce the work, including a geometric modeling pipeline and fabrication guide that enables design exploration, as well as an easily modifiable recognition system, allowing SoundOff tags to be replicated and disseminated throughout any indoor environment.

Sheep temperament, safety, and well-being relies on herd structure and dynamics. Prior works have shown that 1) microclimate can affect sheep flocking behavior and group decisions, 2) and bleating frequency and patterning correspond with individual sheep mood and temperament. We aim to better understand individual sheep and resulting flocking and herd behavior by using electronic headwear that measures microclimate, head movement, eye movement, and bleating patterns to monitor the factors that may affect sheep stress levels and to understand individuals’ responses to those factors. In this paper, we present the design and development of a prototype sensing system and our plans to investigate sheep acceptance of the system with habituated petting zoo sheep. The prototype consists of audio, microclimate, and head movement sensors mounted to a similar halter system that is worn by the petting zoo sheep daily. Based on extensive conversations with sheep specialists, we expect that the sheep will grow accustomed to the placement and attachment of the headpiece after acquaintance with the smell of the device. In the future, we will add cameras to the system to analyze field-of-view and ocular movement to determine how the number of conspecifics in view affects individuals’ behavior. This work is an effort to develop wearable devices for sheep in the field, whose group behaviors remain poorly understood despite individuals being regularly used in behavioral neuroscience research.

Local workers and their knowledge are essential for sustainable and effective conservation efforts. However, many technology-assisted conservation programs are guided by global benchmarks (e.g., forest cover) and industry metrics (e.g., cost per acre), which often devalue local knowledge and fail to consider the economic and conservation goals of local workers. Assets-based design is well-suited to center workers and their strengths, yet it may fail to fully address the complexities of long-term conservation programs by not explicitly emphasizing workers' goals or bolstering their assets. We extend recent approaches in assets-based design literature that address these limitations through our case studies of reforestation, biodiversity monitoring, and carbon sequestration programs in three protected areas in Madagascar. We leverage a mixed-methods approach of direct reactive observations, unstructured interviews, and an informal design workshop, revealing emergent themes surrounding economic sustainability and the value of local ecological knowledge in conservation. Finally, we explore examples, tensions, and design considerations for worker-centered conservation technology to: (1) prioritize local knowledge, (2) foster love of nature, (3) center economic goals, and (4) embrace local autonomy. This work advances the dialogue on assets-based design, promoting the co-creation of equitable and sustainable conservation technologies with workers in Global South settings by centering local economic priorities and enhancing workers' strengths.

Human bodies are almost always in contact with soft materials like clothing, for warmth, protection, self-expression, etc. Recent advancements in intelligent soft wearables have augmented these on-body soft objects with computational functions and intelligence with little compromise on the softness and comforts of wearables, allowing prolonged wear. These innovations, which combine advanced soft sensor design, fabrication, and computational power, offer unprecedented opportunities to improve our health, productivity, and overall well-being with monitoring and assistive capabilities. However, the inherent physical properties of soft materials present unique challenges in achieving practical interactions. The complexity of intelligent soft wearables, multiplexing intricate designs, soft materials, flexible electronics, advanced signal processing algorithms, and machine learning models, necessitates collaborative efforts from experts across diverse domains. This workshop aims to bring together interested researchers and practitioners across relevant domains to discuss the challenges and opportunities of intelligent soft wearables.

A robot's morphology is pivotal to its functionality, as biological organisms demonstrate through shape adjustments - octopi squeeze through small apertures, and caterpillars use peristaltic transformations to navigate complex environments. While existing robotic systems struggle to achieve precise volumetric transformations, Variable Geometry Trusses offer rich morphing capabilities by coordinating hundreds of actuating beams. However, control complexity scales exponentially with beam count, limiting implementations to trusses with only a handful of beams or to designs where only a subset of beams are actuable. Previous work introduced the metatruss, a truss robot that simplifies control by grouping actuators into interconnected pneumatic control networks, but relies on manual network design and control sequences. Here, we introduce a multi-objective optimization framework based on a tailored genetic algorithm to automate actuator grouping, contraction ratios, and actuation timing. We develop a highly damped dynamic simulator that balances computational efficiency with physical accuracy and validate our approach with experimental prototypes. Across multiple tasks, we demonstrate that the metatruss achieves complex shape adaptations with minimal control units. Our results reveal an optimal number of control networks, beyond which additional networks yield diminishing performance gains.

<sec> <title>BACKGROUND</title> Digital behaviors such as sleep, social interaction, and productivity reflect how individuals structure daily life. Among university students, online activity patterns mirror academic schedules, social rhythms, and lifestyle habits, with disruptions linked to sleep, stress, and well-being. Existing approaches—including wearables, apps, and surveys—yield useful insights but depend on self-report or active participation, limiting adherence in real-world use. Passive sensing of network traffic provides a scalable and less burdensome alternative, enabling unobtrusive capture of smartphone usage patterns while preserving privacy. </sec> <sec> <title>OBJECTIVE</title> This study evaluated whether encrypted smartphone network traffic, collected via a standard virtual private network (VPN), can be used to capture patterns of digital behavior. We assessed feasibility (sustained data capture) and acceptability (usability, burden, and privacy perceptions), and examined whether traffic-derived features reveal aspects of digital behavior—including timing, intensity, and regularity—relevant to health and daily functioning. </sec> <sec> <title>METHODS</title> We conducted a two-week prospective observational study at New York University. Thirty-eight students enrolled; 29 provided valid network data, 27 remained active for more than five days, and 25 completed the exit interview. Participants installed the WireGuard VPN client on personal smartphones, which enabled passive capture of encrypted network traffic. Feasibility was assessed across two domains: user retention and data coverage. Acceptability was evaluated using the System Usability Scale (SUS), NASA Task Load Index (NASA-TLX), and semi-structured exit interviews. Beyond evaluating feasibility and acceptability, we conducted exploratory analyses that visualized traffic-derived features in relation to digital activity patterns. </sec> <sec> <title>RESULTS</title> Of the 29 participants who contributed valid data, 27 (93%) remained active for more than five days. Mean data coverage was 74.1% (median 77.1%). Participants contributed an average of 311.6 hours of monitored traffic (~13 days, SD 3.5), with totals ranging from 121 to 496 hours. Usability ratings were high (mean SUS score = 78) and perceived workload low (NASA-TLX scores minimal). Participants described the system as easy to install, unobtrusive, and generally trustworthy, though some reported temporarily disabling the VPN during activities they considered private. No inferential statistical tests were conducted; analyses were descriptive. Exploratory analyses indicated that traffic-derived features reflected daily digital activity rhythms and revealed distinctive lifestyle patterns, including gaming and irregular late-night food delivery use. </sec> <sec> <title>CONCLUSIONS</title> VPN-based monitoring of encrypted smartphone traffic was feasible and acceptable, enabling sustained passive data collection with minimal burden. The findings demonstrate the potential of this approach as a scalable and device-agnostic method for digital phenotyping—capable of capturing fine-grained behavioral rhythms while preserving privacy. With broader validation and deployment, the technique could expand the toolkit for studying health, well-being, and cognitive function in everyday life. </sec> <sec> <title>CLINICALTRIAL</title> Not applicable. This study was not registered as a clinical trial because it did not involve randomization. </sec>

In the pursuit of place-based, generative AI educational technologies, the field of Human-Computer Interaction (HCI) offers a powerful framework for identifying and addressing diverse user needs. In partnership an Hawaiian language immersion (Kaiapuni) school and 13 educators, this 1-year case study presents a research approach rooted in assets-based design and Design Thinking that leverages rapid iteration, usability testing, and speculative prototyping to co-design a generative AI tool for Kaiapuni educators. Our synthesis of observations, participant reflections, and usability testing feedback provides evidence for such methods in their ability to envision ideal outcomes for Kaiapuni education supported by generative AI technologies.

As edge devices see increasing adoption across a wide range of applications, understanding their environmental impact has become increasingly urgent. Unlike cloud systems, edge deployments consist of tightly integrated microcontrollers, sensors, and energy sources that collectively shape their carbon footprint. In this paper, we present a carbon-aware design framework tailored to embedded edge systems. We analyze the embodied emissions of several off-the-shelf microcontroller boards and peripheral components and examine how deployment context—such as workload type, power source, and usage duration—alters the carbon-optimal configuration. Through empirical case studies comparing battery- and solar-powered scenarios, we find that the lowest-emission choice is often workload- and context-specific, challenging assumptions that energy-efficient or renewable powered systems are always the most sustainable. Our results highlight the need for fine-grained, system-level reasoning when designing for sustainability at the edge and provide actionable insights for researchers and practitioners seeking to reduce the carbon cost of future deployments.

Exploring the shared, intersectional problem space of empowered users making carbon-aware decisions that guide computer systems operation.

With the increasing prevalence of Electronic Nicotine Delivery Systems (ENDS), understanding vaping behaviors and nicotine intake is essential. Existing methods such as self-reports, gesture-based monitoring, and sensor-based methods lack reliability, adaptability, and real-time vaping event data. We introduce PuffEM, a low-power, versatile system that detects vaping events reliably using a touch sensor and on-the-surface magnetometer to estimate nicotine intake. Integrated with a mobile app, it collects sensor and contextual data, supporting vaping and addiction research and health interventions. Lab tests confirm PuffEM's ability to detect vaping events, function across three ENDS devices, and estimate vaporized nicotine liquid. A five-day in-wild study (2 participants, 753 puffs, 41 sessions, 5.79 hours) demonstrated high usability and low burden, validating real-world feasibility. By combining specialized hardware (touch, magnetometer, IMU sensors) with a mobile platform, PuffEM enables reliable vaping detection, behavioral insights, and scalable smoking cessation interventions.

We introduce Focal Split, a handheld, snapshot depth camera with fully onboard power and computing based on depth-from-differential-defocus (DfDD). Focal Split is passive, avoiding power consumption of light sources. Its achromatic optical system simultaneously forms two differentially defocused images of the scene, which can be independently captured using two photosensors in a snapshot. The data processing is based on the DfDD theory, which efficiently computes a depth and a confidence value for each pixel with only 500 floating point operations (FLOPs) per pixel from the camera measurements. We demonstrate a Focal Split prototype, which comprises a handheld custom camera system connected to a Raspberry Pi 5 for real-time data processing. The system consumes 4.9 W and is powered on a 5 V, 10,000 mAh battery. The prototype can measure objects with distances from 0.4 m to 1.2 m, outputting 480×360 sparse depth maps at 2.1 frames per second (FPS) using unoptimized Python scripts. Focal Split is DIY friendly. A comprehensive guide to building your own Focal Split depth camera, code, and additional data can be found at https://focal-split.qiguo.org.

The development of intelligent camouflage systems demands advanced materials and control strategies for dynamic environmental adaptation. Here, we demonstrate a bioinspired camouflage system using hydroxypropyl cellulose (HPC), a cellulose derivative that forms cholesteric liquid crystals with mechanically tunable structural colors. By integrating HPC fibers with computer vision-assisted control, we achieve autonomous color matching with the environment through precise mechanical manipulation. Our system employs computer vision and a custom-designed wavelength-value (WV) mapping algorithm to analyze surroundings and control fiber tension, enabling direct modulation of the reflected wavelength. The closed-loop control system achieves color matching with less than 5% error at room temperature and maintains over 95% accuracy across temperatures from 15 to 35 °C. The HPC fibers exhibit reversible color transitions spanning the visible spectrum (400-700 nm). This integration of sustainable biomaterials with computer vision-guided mechanical control demonstrates an alternative approach for advanced camouflage applications, including military concealment and anticounterfeiting technologies.

The growth of urban centers, the impacts of climate change, and the need for information to inform city planning and community organizing have intensified the need for monitoring solutions in cities. The advancement of low-cost sensor technologies and digital twin frameworks presents an opportunity for cities to deploy extensive, real-time monitoring systems. However, few examples of long-term, large-scale, publicly available urban sensor datasets exist, limiting the ability of researchers and planners to explore important topics such as hyperlocal environmental variations. In this paper, we introduce a comprehensive dataset from a 118-node LTE-M connected, solar-powered air quality sensor network deployed across Chicago, Illinois, from April 2021 to April 2023. This dataset comprises 94,915,745 readings of United States EPA criteria air pollutants, in two parts: 1) 75,932,596 gas sensor readings for carbon monoxide (CO), ozone (O3), nitrogen dioxide (NO2), and sulfur dioxide (SO2), and 2) 18,983,149 particulate matter (PM) readings. This open-access dataset uniquely enables researchers to examine critical aspects of smart city sensing, including environmental equity, sensor network deployment dynamics, and interactions between urban infrastructure, natural environments, and residents. Via integration with open datasets from the City of Chicago and other sources, this dataset serves as a foundational tool for advancing research in environmental justice, public health, and urban planning, empowering researchers to build and test ideas before going to deployment to move closer to achieving smart, sustainable urban environments.

We introduce Focal Split, a handheld, snapshot depth camera with fully onboard power and computing based on depth-from-differential-defocus (DfDD). Focal Split is passive, avoiding power consumption of light sources. Its achromatic optical system simultaneously forms two differentially defocused images of the scene, which can be independently captured using two photosensors in a snapshot. The data processing is based on the DfDD theory, which efficiently computes a depth and a confidence value for each pixel with only 500 floating point operations (FLOPs) per pixel from the camera measurements. We demonstrate a Focal Split prototype, which comprises a handheld custom camera system connected to a Raspberry Pi 5 for real-time data processing. The system consumes 4.9 W and is powered on a 5 V, 10,000 mAh battery. The prototype can measure objects with distances from 0.4 m to 1.2 m, outputting 480$\times$360 sparse depth maps at 2.1 frames per second (FPS) using unoptimized Python scripts. Focal Split is DIY friendly. A comprehensive guide to building your own Focal Split depth camera, code, and additional data can be found at https://focal-split.qiguo.org.

The carbon emissions of modern information and communication technologies (ICT) present a significant environmental challenge, accounting for approximately 4% of global greenhouse gases, and are on par with the aviation industry. Modern internet services levy high carbon emissions due to the significant infrastructure resources required to operate them, owing to strict service requirements expected by users. One opportunity to reduce emissions is relaxing strict service requirements by leveraging eco-feedback. In this study, we explore the effect of the carbon reduction impact of allowing longer internet service response time based on user preferences and feedback. Across four services (i.e., Amazon, Google, ChatGPT, Social Media) our study reveals opportunities to relax latency requirements of services based on user feedback; this feedback is application-specific, with ChatGPT having the most favorable eco-feedback tradeoff. Further system studies suggest leveraging the reduced latency can bring down the carbon footprint of an average service request by 93.1%.

Unmanned Aerial Vehicles (UAVs) hold immense potential across various fields, including precision agriculture, rescue missions, delivery services, weather monitoring, and many more. Despite this promise, the limited flight duration of the current UAVs stands as a significant obstacle to their broadscale deployment. Attempting to extend flight time by solar panel charging during midflight is not viable due to battery limitations and the eventual need for replacement. This paper details our investigation of a battery-free fixed-wing UAV, built from cost-effective off-the-shelf components, that takes off, remains airborne, and lands safely using only solar energy. In particular, we perform a comprehensive analysis and design space exploration in the contemporary solar harvesting context and provide a detailed accounting of the prototype's mechanical and electrical capabilities. We also derive the Greedy Energy-Aware Control (GEAC) and Predictive Energy-Aware Control (PEAC) solar control algorithm that overcomes power system brownouts and total-loss-of-thrust events, enabling the prototype to perform maneuvers without a battery. Next, we evaluate the developed prototype in a bench-top setting using artificial light to demonstrate the feasibility of batteryless flight, followed by testing in an outdoor setting using natural light. Finally, we analyze the potential for scaling up the evaluation of batteryless UAVs across multiple locations and report our findings.

As large language models (LLMs) become increasingly integrated into educational technology, their potential to assist in developing curricula has gained interest among educators. Despite this growing attention, their applicability in culturally responsive Indigenous educational settings like Hawai`i's public schools and Kaiapuni (immersion language) programs, remains understudied. Additionally, `Olelo Hawai`i, the Hawaiian language, as a low-resource language, poses unique challenges and concerns about cultural sensitivity and the reliability of generated content. Through surveys and interviews with kumu (educators), this study explores the perceived benefits and limitations of using LLMs for culturally revitalizing computer science (CS) education in Hawaiian public schools with Kaiapuni programs. Our findings highlight AI's time-saving advantages while exposing challenges such as cultural misalignment and reliability concerns. We conclude with design recommendations for future AI tools to better align with Hawaiian cultural values and pedagogical practices, towards the broader goal of trustworthy, effective, and culturally grounded AI technologies.

Mark Weiser predicted in 1991 that computing would lead to individuals interacting with countless computing devices, seamlessly integrating them into their daily lives until they disappear into the background [42]. However, achieving this seamless integration while addressing the associated environmental concerns is challenging. Trillions of smart devices with varied capabilities and form-factor are needed to build a networked environment of this magnitude. Yet, conventional computing paradigms require plastic housings, PCB boards, and rare-earth minerals, coupled with hazardous waste, and challenging reclamation and recycling, leading to significant e-waste. The current linear lifecycle design of electronic devices does not allow circulation among different life stages, neglecting features like recyclability and repairability during the design process. In this position paper, we present the concept of computational materials designed for transiency as a substitute for current devices. We envision that not all devices must be designed with performance, robustness, or even longevity as the sole goal. We detail computer systems challenges to the circular economy of computational materials and provide strategies and sketches of tools to assess a device's entire lifetime environmental impact.