Running multiple deep neural networks (DNNs) simultaneously on mobile devices introduces challenges due to constrained computing resources. Previous research has explored the use of heterogeneous processors for accelerating DNN inference but often overlooks thermal issues, which can degrade computing power. In this article, we propose Phoenix, a system specifically designed to enhance the performance of multi-instance DNNs in video applications by maximizing accuracy and ensuring the achievement of a required frame rate. Phoenix allocates DNN tasks to the most suitable hardware processors, understanding complex thermal dynamics through reinforcement learning, and postpones the onset of thermal throttling. Despite optimized task allocation, continuous inference of multiple DNNs can still lead to thermal throttling. To manage performance degradation, Phoenix employs a multi-exit network, adaptively executing inference tasks to ensure consistent frame rates. Phoenix minimizes accuracy loss from early exits by optimally generating and operating multi-exit networks. We evaluated Phoenix using two different benchmarks and Virtual Youtuber streaming application. The results demonstrated that Phoenix effectively enhances device performance by delaying thermal throttling and achieving optimal accuracy while maintaining a consistent frame rate.

Although on-device video super-resolution enables high-quality live 360-degree streaming on mobile devices, existing methods often waste energy by overlooking perceived visual quality. In this paper, we present EOS, an energy-efficient on-device super-resolution system for mobile omnidirectional video (ODV) live streaming. EOS reduces energy waste by dynamically adjusting super-resolution complexity based on the predicted visual quality of super-resolved frames. This approach raises two challenges: (1) designing an adaptive inference policy that maximizes energy savings while minimizing degradation in Quality-of-Experience (QoE), and (2) developing a method to predict visual quality under the constraints of mobile ODV live streaming. To tackle these challenges, EOS introduces EOS SR and a No-Reference Up-scaling Quality Prediction scheme. EOS SR employs a device-agnostic, scalable deep neural network optimized for mobile devices, with an energy-aware scheduler that jointly selects the optimal super-resolution model and GPU frequency. The No-Reference Upscaling Quality Prediction scheme estimates visual quality across arbitrary viewpoints in real time without requiring high-resolution reference videos. Experiments on commodity smartphones show that EOS reduces average power consumption by 34.6%–49.9% compared to baseline methods, while preserving high visual quality and frame rates.

For highly immersive mobile volumetric video streaming, it is essential to deliver photo-realistic full-scene content with smooth playback. Unlike traditional representations such as point clouds, 3D Gaussian Splatting (3DGS) has gained attention for its ability to represent high-quality full-scene 3D content. However, our preliminary experiments show that existing methods for 3DGS-based videos fail to achieve smooth playback on mobile devices. In this paper, we propose Vega, a 3DGS-based photo-realistic full-scene volumetric video streaming system that ensures real-time playback on mobile devices. The core idea behind Vega's real-time rendering is object-level selective computation, which allocates computational resources to visually important objects to meet strict rendering deadlines. To enable mobile streaming based on the selective computation, Vega addresses two challenges: (1) designing an encoding scheme that optimizes the data size of videos while being compatible with object-level prioritization, and (2) developing a rendering pipeline that efficiently operates on resource-constrained mobile devices. We implemented an end-to-end Vega system, consisting of a streaming server and an Android application. Experimental results on commodity smartphones show that Vega achieves 30 frames per second (FPS) for full-scene volumetric video streaming while maintaining competitive data size and visual quality compared to existing baselines.

Deep neural networks (DNNs) have enabled a wide range of artificial intelligence of things (AIoT) applications, but their increasing complexity poses challenges for deployment on resource-constrained devices. Model compression techniques such as pruning and quantization have been widely adopted to address these challenges; however, they inevitably incur accuracy loss due to information loss. Recently, class-aware pruning has emerged as a promising approach, but existing methods often lack flexibility, as they are typically tailored to fixed target class sets and fail to generalize well to dynamic or broad class distributions. To address this limitation, we propose Mixture of Class-aware Experts (MoCE), a novel framework that combines class-aware pruning with a Mixture of Experts (MoE) architecture. MoCE constructs multiple lightweight experts using class-aware pruning, each specialized for a subset of classes, and employs a shared encoder and a lightweight router to dynamically select the appropriate expert at runtime. Our preliminary results demonstrate the potential of combining class-aware pruning and expert selection to enable accurate and efficient inference on resource-limited AIoT devices.

Mobile Augmented Reality (AR) applications demand high-quality, real-time visual prediction, including pixel-level depth and semantics, to enable immersive and context-aware user experiences. Recently, Vision Foundation Models (VFMs) offer strong generalization capabilities on diverse and unseen data, supporting scalable mobile AR experiences. However, deploying VFMs on mobile devices is challenging due to computational limitations, particularly in maintaining both prediction accuracy and real-time performance. In this paper, we present ARIA, the first system that enables on-device inference acceleration of a VFM. ARIA employs the heterogeneity of mobile processors through a parallel and selective inference scheme: full-frame prediction is periodically offloaded to a processor with high parallelism capability like GPU, while low-latency updates on dynamic regions are conducted via a specialized accelerator like NPU. Implemented and evaluated using mobile devices, ARIA achieved significant improvements in accuracy and deadline success rate on diverse real-world mobile AR scenarios.

Accurate, real-time machine perception is a key enabler of emerging mobile applications such as augmented reality and autonomous driving. However, running complex vision models within the tight latency budgets of resource-limited platforms remains challenging. We address two root causes: (i) the growing computational demands of state-of-the-art vision models and (ii) the variability of compute resource availability in on-device AI deployments. In this extended abstract, we introduce two adaptive perception systems that leverage AI-system co-design. Deployed on commercial devices and evaluated on representative perception workloads, our systems demonstrate high-performance perception under practical latency and resource constraints.

Existing Android systems exhibit energy inefficiency during mobile web browsing due to the lack of awareness of application-level context. Inferring such context from system-level data alone is challenging, but one promising opportunity is using the sequence of task wakeup events, where one task activates another. These sequences show correlation with the type of webpage being used. In this article, we present Ember, a lightweight and responsive power management system for mobile web browsing using only task wakeup sequences. Ember introduces a neural network–based approach to predict optimal CPU clamping values by addressing three key challenges: (1) embedding task names, given as natural-language strings, into meaningful vectors using a Word2Vec-based embedding scheme tailored for task wakeup sequences; (2) minimizing inference overhead with a touch-driven hierarchical inference method that combines lightweight logistic regression with high-accuracy neural networks to balance responsiveness and efficiency; and (3) adapting to within-page interaction dynamics through an interaction-adaptive clamping mechanism that adjusts constraints across different user interaction phases. Implemented on commercial Android smartphones, Ember reduced power consumption by 6.2%–31.2% across a wide range of webpages while maintaining user-perceived quality of experience (QoE).

Reckless usage of electric (e-) scooters causes many injury accidents, raising critical safety concerns. Despite newly introduced regulations, specifically, speed limits and sidewalk driving prohibitions, the number of accidents increases due to the challenges in enforcement. Therefore, a reliable method to detect safety-threatening illegal behaviors of e-scooters is essential to mitigate this growing problem. In this article, we propose SecureRide, a system that accurately detects illegal e-scooter behaviors, i.e., speeding violation and sidewalk riding, at runtime using only battery information, without the need for additional sensors. To this end, we first design a neural network-based illegal behavior predictor that takes sequences of three battery factors, i.e., voltage, current, and capacity, as inputs. The model architecture is optimized based on time constraints, target accuracy, and resource constraints of the target devices. Next, we devise a runtime detection strategy to achieve both high accuracy and low detection time. SecureRide operates in two modes with different predictors– lightweight-quick and complex-accurate models–depending on the driving situation, ensuring both high accuracy and low detection time. We extensively validate SecureRide based on actual driving experiments. Our results show that SecureRide detects illegal behaviors with an accuracy of up to 99.77% within 1.01 seconds.

We propose a cross-device web solution, called Vulture, which distributes graphical user interface (GUI) elements of apps across multiple devices without requiring modifications of web apps or browsers. Several challenges should be resolved to achieve the goals. First, the peer–server configuration should be efficiently established to distribute web resources in cross-device web environments. Vulture exploits an in-browser virtual proxy that runs the web server’s functionality in web browsers using a virtual HTTP scheme and a relevant API. Second, the functional consistency of web apps must be ensured in GUI-distributed environments. Vulture solves this challenge by providing a single-browser illusion with a two-tier document object models (DOM) architecture, which handles view state changes and user input seamlessly in cross-device environments. We implemented Vulture and extensively evaluated the system under various combinations of operating platforms, devices, and network capabilities while running 50 real web apps. The experiment results show that the proposed scheme provides functionally consistent cross-device web experiences by allowing fine-grained GUI distribution. We also confirmed that the in-browser virtual proxy reduces the GUI distribution time and the view change reproduction time by averages of 38.47% and 20.46%, respectively.

3D object detection with omnidirectional views enables safety-critical applications such as mobile robot navigation. Such applications increasingly operate on resource-constrained edge devices, facilitating reliable processing without privacy concerns or network delays. To enable cost-effective deployment, cameras have been widely adopted as a low-cost alternative to LiDAR sensors. However, the compute-intensive workload to achieve high performance of camera-based solutions remains challenging due to the computational limitations of edge devices. In this paper, we present Panopticus, a carefully designed system for omnidirectional and camera-based 3D detection on edge devices. Panopticus employs an adaptive multi-branch detection scheme that accounts for spatial complexities. To optimize the accuracy within latency limits, Panopticus dynamically adjusts the model's architecture and operations based on available edge resources and spatial characteristics. We implemented Panopticus on three edge devices and conducted experiments across real-world environments based on the public self-driving dataset and our mobile 360° camera dataset. Experiment results showed that Panopticus improves accuracy by 62% on average given the strict latency objective of 33ms. Also, Panopticus achieves a 2.1× latency reduction on average compared to baselines.

The live streaming of omnidirectional video (ODV) on mobile devices demands considerable network resources; thus, current mobile networks are incapable of providing users with high-quality ODV equivalent to conventional flat videos. We observe that mobile devices, in fact, underutilize graphics processing units (GPUs) while processing ODVs; hence, we envisage an opportunity exists in exploiting video super-resolution (VSR) for improved ODV quality. However, the device-specific discrepancy in GPU capability and dynamic behavior of GPU frequency in mobile devices create a challenge in providing VSR-enhanced ODV streaming. In this paper, we propose OmniLive, an on-device VSR system for mobile ODV live streaming. OmniLive addresses the dynamicity of GPU capability with an anytime inference-based VSR technique called Omni SR. For Omni SR, we design a VSR deep neural network (DNN) model with multiple exits and an inference scheduler that decides on the exit of the model at runtime. OmniLive also solves the performance heterogeneity of mobile GPUs using the Omni neural architecture search (NAS) scheme. Omni NAS finds an appropriate DNN model for each mobile device with Omni SR-specific neural architecture search techniques. We implemented OmniLive as a fully functioning system encompassing a streaming server and Android application. The experiment results show that our anytime VSR model provides four times upscaled videos while saving up to 57.15% of inference time compared with the previous super-resolution model showing the lowest inference time on mobile devices. Moreover, OmniLive can maintain 30 frames per second while fully utilizing GPUs on various mobile devices.

We present the results of propagation measurements at 2.485 GHz for fixed wireless loops. Path loss measurements were performed and characterized at 43 subscriber locations around a base station antenna located on top of Crawford Hill in Holmdel, NJ. This suburban location is characterized by rolling hills, foliage, and terrain blockages. Temporal and horizontal motion path loss fluctuations were found to be uncorrelated, each characterized by a different Ricean distribution. Lower r.m.s. delay spreads were obtained with directive subscriber antennas than with omni-directional antennas. No substantial gain loss (less than 2 dB) of subscribers' directive antennas was observed. The effects of trees, with foliage, surrounding the base station upon the path loss and the ratio of scattered power to specular power are also examined. The distance exponent of path loss versus distance (about 1.5) was observed to be less than free-space. Diffraction loss from hilltop trees, shadowing the base station, are suspected to be the cause. This loss decreases as the remote moves further away and comes out of the shadow. Scattered power from directions other than line-of-sight was observed to be as high as one half of the specular contribution when tree scattering near the base station was significant.