MobiCom 2021 Papers

To take advantage of Visible Light Communication (VLC) for low-power applications, such as IoT tags, researchers have been developing systems to modulate (backscatter) ambient light using LC shutters. Various approaches have been explored for single-pixel transmitters, but without following a principled approach. This has resulted in either relatively low data rates, short ranges, or the need for powerful artificial light sources. This paper takes a step back and proposes a more theoretical framework: ChromaLux. By considering the fundamental characteristics of liquid crystals (birefringence and thickness), we demonstrate that the design space is way larger than previously explored, allowing for much better systems. In particular, we uncover the existence of a transient state where the switching time can be reduced by an order of magnitude without lowering the contrast significantly, improving both range and data rate. Using a prototype, we demonstrate that our framework is applicable to different LCs. Our results show significant improvements over state-of-the-art single-pixel systems, achieving ranges of 50 meters at 1 kbps and with bit-error-rates below 1%.

Object detection is a fundamental building block of video analytics applications. While Neural Networks (NNs)-based object detection models have shown excellent accuracy on benchmark datasets, they are not well positioned for high-resolution images inference on resource-constrained edge devices. Common approaches, including down-sampling inputs and scaling up neural networks, fall short of adapting to video content changes and various latency requirements. This paper presents Remix, a flexible framework for high-resolution object detection on edge devices. Remix takes as input a latency budget, and come up with an image partition and model execution plan which runs off-the-shelf neural networks on non-uniformly partitioned image blocks. As a result, it maximizes the overall detection accuracy by allocating various amount of compute power onto different areas of an image. We evaluate Remix on public dataset as well as real-world videos collected by ourselves. Experimental results show that Remix can either improve the detection accuracy by 18%-120% for a given latency budget, or achieve up to 8.1× inference speedup with accuracy on par with the state-of-the-art NNs.

This paper presents the design and implementation of PCube, a phase-based parallel packet decoder for concurrent transmissions of LoRa nodes. The key enabling technology behind PCube is a novel air-channel phase measurement technique which is able to extract phase differences of air-channels between LoRa nodes and multiple antennas of a gateway. PCube leverages the reception diversities of multiple receiving antennas of a gateway and scales the concurrent transmissions of a large number of LoRa nodes, even exceeding the number of receiving antennas at a gateway. As a phase-based parallel decoder, PCube provides a new dimension to resolve collisions and supports more concurrent transmissions by complementing time and frequency based parallel decoders. PCube is implemented and evaluated with synchronized software defined radios and off-the-shelf LoRa nodes in both indoors and outdoors. Results demonstrate that PCube can substantially outperform state-of-the-art works in terms of aggregated throughput by 4.9× and the number of concurrent nodes by up to 5×. More importantly, PCube scales well with the number of receiving antennas of a gateway, which is promising to break the barrier of concurrent transmissions.

Federated learning (FL) has been a popular method to achieve distributed machine learning among numerous devices without sharing their data to a cloud server. FL aims to learn a shared global model with the participation of massive devices under the orchestration of a central server. However, mobile devices usually have limited communication bandwidth to transfer local updates to the central server. In addition, the data residing across devices is intrinsically statistically heterogeneous (i.e., non-IID data distribution). Learning a single global model may not work well for all devices participating in the FL under data heterogeneity. Such communication cost and data heterogeneity are two critical bottlenecks that hinder from applying FL in practice. Moreover, mobile devices usually have limited computational resources. Improving the inference efficiency of the learned model is critical to deploy deep learning applications on mobile devices. In this paper, we present Hermes - a communication and inference-efficient FL framework under data heterogeneity. To this end, each device finds a small subnetwork by applying the structured pruning; only the updates of these subnetworks will be communicated between the server and the devices. Instead of taking the average over all parameters of all devices as conventional FL frameworks, the server performs the average on only overlapped parameters across each subnetwork. By applying Hermes, each device can learn a personalized and structured sparse deep neural network, which can run efficiently on devices. Experiment results show the remarkable advantages of Hermes over the status quo approaches. Hermes achieves as high as 32.17% increase in inference accuracy, 3.48× reduction on the communication cost, 1.83× speedup in inference efficiency, and 1.8× savings on energy consumption.

Maximizing the quality of experience (QoE) for real-time video is a long-standing challenge. Traditional video transport protocols, represented by a few deterministic rules, can hardly adapt to the heterogeneous and highly dynamic modern Internet. Emerging learning-based algorithms have demonstrated potential to meet the challenge. However, our measurement study reveals an alarming long tail performance issue: these algorithms tend to be bottle-necked by occasional catastrophic events due to the built-in exploration mechanisms. In this work, we propose Loki, which improves the robustness of learning-based model by coherently integrating it with a rule-based algorithm. To enable integration at feature level, we first reverse-engineer the rule-based algorithm into an equivalent "black-box" neural network. Then, we design a dual-attention feature fusion mechanism to fuse it with a reinforcement learning model. We train Loki in a commercial real-time video system through online learning, and evaluate it over 101 million video sessions, in comparison to state-of-the-art rule-based and learning-based solutions. The results show that Loki improves not only the average but also the tail performance substantially (26.30% to 44.24% reduction of stall rate and 1.76% to 2.17% increase in video throughput at 95-percentile).

LoRa, as a representative Low-Power Wide-Area Network (LPWAN) technology, can provide long-range communication for battery-powered IoT devices with a 10-year lifetime. LoRa links in practice, however, experience high dynamics in various environments. When the SNR falls below the threshold (e.g., in the building), a LoRa device disconnects from the network. We propose Falcon, which addresses the link dynamics by enabling data transmission for very low SNR or even disconnected LoRa links. At the heart of Falcon, we reveal that low SNR LoRa links that cannot deliver packets can still introduce interference to other LoRa transmissions. Therefore, Falcon transmits data bits on the low SNR link by selectively interfering with other LoRa transmissions. We address practical challenges in Falcon design. We propose a low-power channel activity detection method to detect other LoRa transmissions for selective interference. To interfere with the so-called interference-resilient LoRa, we accurately estimate the time and frequency offsets on LoRa packets and propose an adaptive frequency adjusting strategy to maximize the interference. We implement Falcon, all using commercial off-the-shelf LoRa devices, and extensively evaluate its performance. The results show that Falcon can provide reliable communication links for disconnected LoRa devices and achieves the SNR boundary upto 7.5 dB lower than that of standard LoRa.

Beamforming can improve the communication and sensing capabilities for a wide range of IoT applications. However, most existing IoT devices cannot perform beamforming due to form factor, energy, and cost constraints. This paper presents RFlens, a reconfigurable metasurface that empowers low-profile IoT devices with beamforming capabilities. The metasurface consists of many unit-cells, each acting as a phase shifter for signals going through it. By encoding the phase shifting values, RFlens can manipulate electromagnetic waves to "reshape" and resteer the beam pattern. We prototype RFlens for 5 GHz Wi-Fi signals. Extensive experiments demonstrate that RFlens can achieve a 4.6 dB median signal strength improvement (up to 9.3 dB) even with a relatively small 16 × 16 array of unit-cells. In addition, RFlens can effectively improve the secrecy capacity of IoT links and enable passive NLoS wireless sensing applications.

Human gait is a widely used biometric trait for user identification and recognition. Given the wide-spreading, steady diffusion of ear-worn wearables (Earables) as the new frontier of wearable devices, we investigate the feasibility of earable-based gait identification. Specifically, we look at gait-based identification from the sounds induced by walking and propagated through the musculoskeletal system in the body. Our system, EarGate, leverages an in-ear facing microphone which exploits the earable's occlusion effect to reliably detect the user's gait from inside the ear canal, without impairing the general usage of earphones. With data collected from 31 subjects, we show that EarGate achieves up to 97.26% Balanced Accuracy (BAC) with very low False Acceptance Rate (FAR) and False Rejection Rate (FRR) of 3.23% and 2.25%, respectively. Further, our measurement of power consumption and latency investigates how this gait identification model could live both as a stand-alone or cloud-coupled earable system.

Each fall, millions of monarch butterflies across the northern US and Canada migrate up to 4,000 km to overwinter in the exact same cluster of mountain peaks in central Mexico. To track monarchs precisely and study their navigation, a monarch tracker must obtain daily localization of the butterfly as it progresses on its 3-month journey. And, the tracker must perform this task while having a weight in the tens of milligram (mg) and measuring a few millimeters (mm) in size to avoid interfering with monarch's flight. This paper proposes mSAIL, 8 × 8 × 2.6 mm and 62 mg embedded system for monarch migration tracking, constructed using 8 prior custom-designed ICs providing solar energy harvesting, an ultra-low power processor, light/temperature sensors, power management, and a wireless transceiver, all integrated and 3D stacked on a micro PCB with an 8 × 8 mm printed antenna. The proposed system is designed to record and compress light and temperature data during the migration path while harvesting solar energy for energy autonomy, and wirelessly transmit the data at the overwintering site in Mexico, from which the daily location of the butterfly can be estimated using a deep learning-based localization algorithm. A 2-day trial experiment of mSAIL attached on a live butterfly in an outdoor botanical garden demonstrates the feasibility of individual butterfly localization and tracking.

5G networks are gaining pace in development and deployment in recent years. One of 5G's key objective is to support a variety of use cases with different Service Level Objectives (SLOs). Slicing is a key part of 5G that allows operators to provide a tailored set of resources to different use cases in order to meet their SLOs. Existing works focus on slicing in the frontend or the C-RAN. However, slicing is missing in the fronthaul network that connects the frontend to the C-RAN. This leads to over-provisioning in the fronthaul and the C-RAN, and also limits the scalability of the network.

Traditional liquid identification instruments are often unavailable to the general public. This paper shows the feasibility of identifying unknown liquids with commercial lightweight devices, such as a smartphone. The wisdom arises from the fact that different liquid molecules have various viscosity coefficients, so they need to overcome dissimilitude energy barriers during relative motion. With this intuition in mind, we introduce a novel model that measures liquids' viscosity based on active vibration. Yet, it is challenging to build up a robust system utilizing the built-in accelerometer in smartphones. Practical issues include under-sampling, self-interference, and volume change impact. Instead of machine learning, we tackle these issues through multiple signal processing stages to reconstruct the original signals and cancel out the interference. Our approach could achieve the liquid viscosity estimates with a mean relative error of 2.9% and distinguish 30 kinds of liquid with an average accuracy of 95.47%.

Modern acoustic wearables with microphone arrays are promising to offer rich experience (e.g., 360° sound and acoustic imaging) to consumers. Realtime multi-track audio streaming with precise synchronization however poses significant challenges to the existing wireless microphone array designs that depend on complex digital synchronization as well as bulky and power-hungry hardware.

Massive MIMO forms a crucial component for 5G because of its ability to improve quality of service and support multiple streams simultaneously. However, for real-world MIMO deployments, estimating the downlink wireless channel from each antenna on the base station to every client device is a critical bottleneck, especially for the widely used frequency duplexed designs that cannot utilize reciprocity. Typically, this channel estimation requires explicit feedback from client devices and is prohibitive for large antenna deployments. In this paper, we present FIRE, a system that uses an end-to-end machine learning approach to enable accurate channel estimation without requiring any feedback from client devices. FIRE is interpretable, accurate, and has low compute overhead. We show that FIRE can successfully support MIMO transmissions in a real-world testbed and achieves SNR improvement over 10 dB in MIMO transmissions compared to the current state-of-the-art.

Accurate positioning in global navigation satellite system (GNSS)-denied environments, such as tunnels and underpasses, remains a challenge. Navigation systems for such environments need to strike a balance between price and precision. In this paper, we propose magnetic vehicular positioning (MVP), a navigation system that guides drivers in GNSS-denied environments. The key idea of MVP is to extract magnetic fingerprints from geomagnetic field anomalies. By comparing the measured magnetic field against a magnetic map, positioning can be achieved without GNSS signals. Our proposed matching algorithm allows MVP to provide 5.14 m positioning accuracy. We conducted large-scale real-road experiments for 36 months in two countries and 56 tunnels to demonstrate the effectiveness of the proposed system. Because MVP can be deployed on off-the-shelf smartphones, our approach makes accurate navigation in GNSS-denied environments affordable.

Should you decide to launch a nano-satellite today in Low-Earth Orbit (LEO), the cost of renting ground station communication infrastructure is likely to significantly exceed your launch costs. While space launch costs have lowered significantly with innovative launch vehicles, private players, and smaller payloads, access to ground infrastructure remains a luxury. This is especially true for smaller LEO satellites that are only visible at any location for a few tens of minutes a day and whose signals are extremely weak, necessitating bulky and expensive ground station infrastructure.

In this paper, we devise new attacks exploiting the unprotected data-plane signaling in cellular IoT networks (aka both NB-IoT and Cat-M). We show that, despite the deployed security mechanisms on both control-plane signaling and data-plane packet forwarding, novel data-plane signaling attacks are still feasible. Such attacks exhibit a variety of attack forms beyond simplistic packet-blasting, denial-of-service (DoS) threats, including location privacy breach, packet delivery loop, prolonged data delivery, throughput limiting, radio resource draining, and connection reset. Our testbed evaluation and operational network validation have confirmed the viability. We further propose a new defense solution within the 3GPP C-IoT standard framework.

Carrying over 75% of the last-mile mobile Internet traffic, WiFi has inevitably become an enticing target for various security threats. In this work, we characterize a wide variety of real-world WiFi threats at an unprecedented scale, involving 19 million WiFi access points (APs) mostly located in China, by deploying a crowdsourced security checking system on 14 million mobile devices in the wild. Leveraging the collected data, we reveal the landscape of nationwide WiFi threats for the first time. We find that the prevalence, riskiness, and breakdown of WiFi threats deviate significantly from common understandings and prior studies. In particular, we detect attacks at around 4% of all WiFi APs, uncover that most WiFi attacks are driven by an underground economy, and provide strong evidence of web analytics platforms being the bottleneck of its monetization chain. Further, we provide insightful guidance for defending against WiFi attacks at scale, and some of our efforts have already yielded real-world impact---effectively disrupted the WiFi attack ecosystem.

Being able to use a single app across multiple devices can bring novel experiences to the users in various domains including entertainment and productivity. For instance, a user of a video editing app would be able to use a smart pad as a canvas and a smartphone as a remote toolbox so that the toolbox does not occlude the canvas during editing. However, existing approaches do not properly support the single-app multi-device execution due to several limitations, including high development cost, device heterogeneity, and high performance requirement. In this paper, we introduce FLUID-XP, a novel cross-platform multi-device system that enables UIs of a single app to be executed across heterogeneous platforms, while overcoming the limitations of previous approaches. FLUID-XP provides flexible, efficient, and seamless interactions by addressing three main challenges: i) how to transparently enable a single-display app to use multiple displays, ii) how to distribute UIs across heterogeneous devices with minimal network traffic, and iii) how to optimize the UI distribution process when multiple UIs have different distribution requirements. Our experiments with a working prototype of FLUID-XP on Android confirm that FLUID-XP successfully supports a variety of unmodified real-world apps across heterogeneous platforms (Android, iOS, and Linux). We also conduct a lab study with 25 participants to demonstrate the effectiveness of FLUID-XP with real users.

Deep neural networks (DNNs) have become ubiquitous techniques in mobile and embedded systems for applications such as image/object recognition and classification. The trend of executing multiple DNNs simultaneously exacerbate the existing limitations of meeting stringent latency/accuracy requirements on resource constrained mobile devices. The prior art sheds light on exploring the accuracy-resource tradeoff by scaling the model sizes in accordance to resource dynamics. However, such model scaling approaches face to imminent challenges: (i) large space exploration of model sizes, and (ii) prohibitively long training time for different model combinations. In this paper, we present LegoDNN, a lightweight, block-grained scaling solution for running multi-DNN workloads in mobile vision systems. LegoDNN guarantees short model training times by only extracting and training a small number of common blocks (e.g. 5 in VGG and 8 in ResNet) in a DNN. At run-time, LegoDNN optimally combines the descendant models of these blocks to maximize accuracy under specific resources and latency constraints, while reducing switching overhead via smart block-level scaling of the DNN. We implement LegoDNN in TensorFlow Lite and extensively evaluate it against state-of-the-art techniques (FLOP scaling, knowledge distillation and model compression) using a set of 12 popular DNN models. Evaluation results show that LegoDNN provides 1,296x to 279,936x more options in model sizes without increasing training time, thus achieving as much as 31.74% improvement in inference accuracy and 71.07% reduction in scaling energy consumptions.