Polygonal planes are used to simplify the modeling of the building point cloud, which is widely applied for city planning, 3-D cadastral management, and model rendering. Much of the current research mainly focuses on watertight polygon model reconstruction from the 3-D model data with few missing values, which is highly hypothetical and restricted in application. Therefore, this letter proposes a non-watertight PolyFit (NW-PolyFit) algorithm based on the fit degree of connection and data fitting to reconstruct a non-watertight polygon model from the data with missing values. First, to refine the planar structure in buildings, the refined supporting planes are generated from the detected point cloud primitives. Second, to retain the sharp features, the$\delta $expansion planes are generated from their corresponding supporting planes. We get the relationship between primitives by intersecting those$\delta $expansion planes. Finally, to eliminate the influence of missing data, a fit-and-remove strategy is proposed to filter the generated candidate’s faces, which achieves non-watertight modeling. Experiments show that the NW-PolyFit achieved similar modeling effects for completed data compared with the state-of-the-art methods. The NW-PolyFit achieves non-watertight modeling from the point cloud data with massive missing values while other methods do not.

In applications such as natural language processing or computer vision, one is given a large $n \times n$ matrix $A = (a_{i,j})$ and would like to compute a matrix decomposition, e.g., a low rank approximation, of a function $f(A) = (f(a_{i,j}))$ applied entrywise to $A$. A very important special case is the likelihood function $f\left( A \right ) = \log{\left( \left| a_{ij}\right| +1\right)}$. A natural way to do this would be to simply apply $f$ to each entry of $A$, and then compute the matrix decomposition, but this requires storing all of $A$ as well as multiple passes over its entries. Recent work of Liang et al. shows how to find a rank-$k$ factorization to $f(A)$ using only $n \cdot \poly(\eps^{-1}k\log n)$ words of memory, with overall error $10\|f(A)-[f(A)]_k\|_F^2 + \poly(\epsilon/k) \|f(A)\|_{1,2}^2$, where $[f(A)]_k$ is the best rank-$k$ approximation to $f(A)$ and $\|f(A)\|_{1,2}^2$ is the square of the sum of Euclidean lengths of rows of $f(A)$. Their algorithm uses $3$ passes over the entries of $A$. The authors pose the open question of obtaining an algorithm with $n \cdot \poly(\eps^{-1}k\log n)$ words of memory using only a single pass over the entries of $A$. In this paper we resolve this open question, obtaining the first single-pass algorithm for this problem and for the same class of functions $f$ studied by Liang et al. Moreover, our error is $\|f(A)-[f(A)]_k\|_F^2 + \poly(\epsilon/k) \|f(A)\|_F^2$, where $\|f(A)\|_F^2$ is the sum of squares of Euclidean lengths of rows of $f(A)$. Thus our error is significantly smaller, as it removes the factor of $10$ and also $\|f(A)\|_F^2 \leq \|f(A)\|_{1,2}^2$.

We present a learning-based planner that aims to robustly drive a vehicle by mimicking human drivers’ driving behavior. We leverage a mid-to-mid approach that allows us to manipulate the input to our imitation learning network freely. With that in mind, we propose a novel feedback synthesizer for data augmentation. It allows our agent to gain more driving experience in various previously unseen environments that are likely to encounter, thus improving overall performance. This is in contrast to prior works that rely purely on random synthesizers. Furthermore, rather than completely commit to imitating, we introduce task losses that penalize undesirable behaviors, such as collision, off-road, and so on. Unlike prior works, this is done by introducing a differentiable vehicle rasterizer that directly converts the waypoints output by the network into images. This effectively avoids the usage of heavyweight ConvLSTM networks, therefore, yields a faster model inference time. About the network architecture, we exploit an attention mechanism that allows the network to reason critical objects in the scene and produce better interpretable attention heatmaps. To further enhance the safety and robustness of the network, we add an optional optimization-based post-processing planner improving the driving comfort. We comprehensively validate our method’s effectiveness in different scenarios that are specifically created for evaluating self-driving vehicles. Results demonstrate that our learning-based planner achieves high intelligence and can handle complex situations. Detailed ablation and visualization analysis are included to further demonstrate each of our proposed modules’ effectiveness in our method.

Accurately forecasting the future movements of surrounding vehicles is essential for safe and efficient operations of autonomous driving cars. This task is difficult because a vehicle's moving trajectory is greatly determined by its driver's intention, which is often hard to estimate. By leveraging attention mechanisms along with long short-term memory (LSTM) networks, this work learns the relation between a driver's intention and the vehicle's changing positions relative to road infrastructures, and uses it to guide the prediction. Different from other state-of-the-art solutions, our work treats the on-road lanes as non-Euclidean structures, unfolds the vehicle's moving history to form a spatio-temporal graph, and uses methods from Graph Neural Networks to solve the problem. Not only is our approach a pioneering attempt in using non-Euclidean methods to process static environmental features around a predicted object, our model also outperforms other state-of-the-art models in several metrics. The practicability and interpretability analysis of the model shows great potential for large-scale deployment in various autonomous driving systems in addition to our own.

People around the world are more connected today than ever before. By making phone calls, sending text messages and participating in online chats, mobile users are frequently interacting with their social connections through multiple communication channels. This trend is expected to continue with the emergence of immensely popular communication apps on mobile devices. Intuitively, these interactions on users' mobile phones can reveal valuable information regarding their social relationship with their phone contacts. Understanding such relationship can help provide new services and improve users' mobile experience. In this paper, we explore the opportunity to deeply understand these social relationship through mining mobile communication data. By building an on-device mining framework called Commsense, we show that automatically learning and understanding such relationship can efficiently support useful applications such as categorizing mobile contacts, identifying their relative importance, and automatically managing mobile contacts with very little human interference.

People spend approximately 70% of their time indoors. Understanding the indoor environments is therefore important for a wide range of emerging mobile personal and social applications. Knowledge of indoor floorplans is often required by these applications. However, indoor floorplans are either unavailable or obtaining them requires slow, tedious, and error-prone manual labor.

Electrophoretic displays (EPDs) and organic light emitting diode (OLEDs) are two key technologies used in mobile de-vices. In this paper, we propose the design of an integrated hybrid display combining a transparent OLED (TOLED) and a low power EPD, which is adaptive to show contents of a frame partially on either the TOLED or the EPD. A windows-based predictive model and a calibration algorithm on TOLED are introduced to decide how frame contents can be split between the two displays for achieving the best tradeoff between power reduction and user experiences. A simulation environment that can estimate both the energy consumption and optical properties of the proposed hybrid display is set up based on actual physical measurements. Simulation results show that the predictive model can make right decisions on choosing proper displays in over 90% of the test cases, and this new display design can save over 70% power under many mobile application contexts and still sup-port contents that require fast update rates.

People spend the majority of their time indoors, and human indoor activities are strongly correlated with the rooms they are in. Room localization, which identifies the room a person or mobile phone is in, provides a powerful tool for characterizing human indoor activities and helping address challenges in public health, productivity, building management, etc. Existing room localization methods, however, require labor-intensive manual annotation of individual rooms.

Most people spend more than 90% of their time indoors; indoor air quality (IAQ) influences human health, safety, productivity, and comfort. This paper describes MAQS, a personalized mobile sensing system for IAQ monitoring. In contrast with existing stationary or outdoor air quality sensing systems, MAQS users carry portable, indoor location tracking sensors that provide personalized IAQ information. To improve accuracy and energy efficiency, MAQS incorporates three novel techniques: (1) an accurate temporal n-gram augmented Bayesian room localization method that requires few Wi-Fi fingerprints; (2) an air exchange rate based IAQ sensing method, which measures general IAQ using only CO2 sensors; and (3) a zone-based proximity detection method for collaborative sensing, which saves energy and enables data sharing among users. MAQS has been deployed and evaluated via user study. Detailed evaluation results demonstrate that MAQS supports accurate personalized IAQ monitoring and quantitative analysis with high energy efficiency.

Most people spend more than 90% of their time indoors. Indoor air quality (IAQ) influences human health, safety, productivity, and comfort. This demo introduces MAQS, a personalized mobile sensing system for IAQ monitoring. In contrast with existing stationary or outdoor air quality sensing systems, MAQS users carry portable, indoor location tracking sensors that provide personalized IAQ information. To improve accuracy and energy efficiency, MAQS incorporates three novel techniques: (1) an accurate temporal n-gram augmented Bayesian room localization method; (2) an air exchange rate based IAQ sensing method; and (3) a zone-based proximity detection method for collaborative sensing.

Trip detection is a fundamental issue in many context-sensitive information services on mobile devices. It aims to automatically recognize significant places and trips between them. The key challenge is how to minimize energy consumption while maintaining high accuracy. Previous works that use GPS/WiFi sampling are accurate but energy efficiency is low and does not improve over time. Learning from the human decision making process, we propose an energy-efficient trip detection framework that consists of two modes: The deliberation mode learns cell-id patterns using GPS/WiFi based localization methods; the intuition mode only uses cell-ids and learned patterns for trip detection; transition between the two modes is controlled by parameters that are also learned. We evaluated our framework using real-life traces of six people over five months. Our experiments demonstrate that its energy consumption decreases rapidly as users' activities manifest regularity over time.

In order to improve system performance, many modern processors support super page technology, which is also called variable page. It improves TLB coverage greatly without increasing TLB size. But supporting super page brings great challenge to operating systems. In this paper, we implement two variable page technologies, including the static variable page and dynamic variable page, in Linux kernel on Godson-3 four-core processors. The static variable page technology is implemented on the basis of hugetlbfs pseudo file system. The dynamic one chooses proper page sizes according to the address spaces of applications adaptively. Experiment results show that static variable page technology improves the performance of scientific applications such as large matrix multiplication significantly. The number of TLB misses is reduced by 99% and the speed of large-scale matrix multiplication is improved by over 50%. The dynamic self-adapting variable page technology can bring an average performance improvement of 15% to SPECCPU 2000 benchmarks. Compared with system of single large page (64KB), the file system performance of system supporting dynamic variable page is significantly improved by about 25%.

User identification and access control have become a high demand feature on mobile devices because those devices are wildly used by employees in corporations and government agencies for business and store increasing amount of sensitive data. This paper describes SenGuard, a user identification framework that enables continuous and implicit user identification service for smartphone. Different from traditional active user authentication and access control, SenGuard leverages availability of multiple sensors on today's smartphones and passively use sensor inputs as sources of user authentication. It extracts sensor modality dependent user identification features from captured sensor data and performs user identification at background. SenGuard invokes active user authentication when there is a mounting evidence that the phone user has changed. In addition, SenGuard uses a novel virtualization based system architecture as a safeguard to prevent subversion of the background user identification mechanism by moving it into a privileged virtual domain. An initial prototype of SenGuard was created using four sensor modalities including, voice, location, multitouch, and locomotion. Preliminary empirical studies with a set of users indicate that those four modalities are suited as data sources for implicit mobile user identification.

Emerging electric-drive vehicles demonstrate the potential for significant reduction of petroleum consumption and greenhouse gas emissions. Existing electric-drive vehicles typi- cally include a battery system consisting of thousands of Lithium-ion battery cells. Therefore, large-scale battery-system modeling and analysis is essential for battery system performance analysis, next-generation battery system design, and transportation electrification.

Texture compression is a specialized form of still image compression employed in computer graphics systems to reduce memory bandwidth consumption. Modern texture compression schemes cannot generate satisfactory qualities for both alpha channel and color channel of texture images. We propose a novel texture compression scheme, named ImTC, based on the insight into the essential difference between transparency and color. ImTC defines new data formats and compresses the two channels flexibly. While keeping the same compression ratio as the de facto standard texture compression scheme, ImTC improves compression qualities of both channels. The average PSNR score of alpha channel is improved by about 0.2 dB, and that of color channel can be increased by 6.50 dB over a set of test images, which makes ImTC a better substitute for the standard scheme.

Tuple spaces have been used to build a wide variety of parallel and distributed applications. Two popular systems for building tuple space applications at present are JavaSpaces and Rinda. JavaSpaces uses Java RMI for distributed communication, while Rinda uses Ruby's DRb. Despite an abundance of applications built using JavaSpaces and Rinda, there hasn't been any comparative evaluation of these two systems. This paper provides a detailed ex perimental evaluation of these two systems under several different networking configurations and operating scenarios.