GNSS localization using everyday mobile devices is challenging in urban environments, as ranging errors caused by the complex propagation of satellite signals and low-quality onboard GNSS hardware are blamed for undermining positioning accuracy. Researchers have pinned their hopes on data-driven methods to regress such ranging errors from raw measurements. However, the grueling annotation of ranging errors impedes their pace. This paper presents a robust end-to-end Neural Ranging Correction (NeRC) framework, where localization-related metrics serve as the task objective for training the neural modules. Instead of seeking impractical ranging error labels, we train the neural network using ground-truth locations that are relatively easy to obtain. This functionality is supported by differentiable moving horizon location estimation (MHE) that handles a horizon of measurements for positioning and backpropagates the gradients for training. Even better, as a blessing of end-to-end learning, we propose a new training paradigm using Euclidean Distance Field (EDF) cost maps, which alleviates the demands on labeled locations. We evaluate the proposed NeRC on public benchmarks and our collected datasets, demonstrating its distinguished improvement in positioning accuracy. We also deploy NeRC on the edge to verify its real-time performance for mobile devices.

In low-altitude urban intelligent transportation systems, efficient cooperative task allocation and path planning for multiple unmanned aerial vehicles (UAV) are critical for ensuring the effective execution of complex tasks. This paper proposes a distributed decision-making and autonomous planning framework to achieve cooperative task allocation and path planning for multi-UAVs in low-altitude urban traffic environment. The mission requirements of task allocation and path planning are modeled using evolutionary potential games and show that there exists a Nash equilibrium for the proposed potential function. An Improved Log-linear Learning Algorithm (ILLA) is proposed, and suitable Boltzmann parameters are derived which will enable the proposed ILLA to converge to the optimal Nash equilibrium with a probability one. Furthermore, a Constraint-Based Multi-layer Bidirectional Adaptive A-Star (CBMBA A-Star) algorithm is designed to find optimal and collision free paths for each UAV. Compared with the baseline method, simulation results demonstrate that the proposed approach improves the task reward by 11.67%, reduces the task execution time by 37.41%, and decreases run time by 61.02%, confirming its effectiveness and efficiency in the complex low-altitude urban traffic scenario.

The Position information is essential for large-scale Internet of Things (IoT) devices and services. Multipath and non-line of sight (NLOS) effects introduce additional delays in pseudorange measurements in urban areas. It is one of the main unmodeled errors in Global Navigation Satellite Systems (GNSS). To mitigate interference, various techniques have been developed, including antenna design and sensor fusion. However, traditional estimation approaches often produce biased estimates under the additional path delays. To improve estimation accuracy and robustness, we present an Observation Space Representation Refinement (OSRR) algorithm. The initial position is estimated by least squares without the additional path error. Then, the multipath projection method is used to get possible compensation in pseudorange measurements. Subsequently, the Moving Horizontal Estimation (MHE) is leveraged to get the position with corrected observation space. Field experiments demonstrate that the proposed OSRR algorithm significantly reduces the impact of interference on positioning accuracy. There is no empirical constraint to easily adapt to real-time static and kinematic GNSS pseudorange positioning with unilateral obstruction scenarios.

The coordinated operations of Stealth Unmanned Aerial Vehicle (SUAV) and Swarming Drones (SD) have demonstrated formidable power in the military domain. Efficient path planning is a critical technology that enhances combat effectiveness. This paper proposes a game-theoretic optimization approach to achieve cooperative penetration and target search path planning for swarm UAVs. Multi-task and multi-objective optimization models are developed for complex scenarios whose optimal solutions are NP-hard. Consequently, SUAV and SD are defined as leader and followers, respectively. The formulated Stackelberg game model enables distributed intelligent decision-making for SUAV and SD. We theoretically prove that by selecting an appropriate potential function, subgames within the leader-level and followers-level become an Ordinal Potential Game (OPG) with a Nash equilibrium, thereby ensuring the existence of a Stackelberg Equilibrium (SE) through leader-follower interactions. We propose a Gradient-based Hierarchical Learning and Optimization Algorithm (GHLOA) to achieve SE. At the leader-level, a Stochastic Gradient Ascent (SGA) algorithm optimizes the penetration path for SUAV, while in the followers-level, we demonstrate that the designed Hybrid Learning-based Multimodal Adaptive Pigeon-Inspired Optimization (HLMAPIO) algorithm converges with probability one to the suboptimal solution for each SD. Numerical results indicate that our approach is suboptimal, scalable, and fast adaptable to dynamical scenarios, and it outperforms the state-of-the-art techniques.

The pseudorange error is one of the root causes of localization inaccuracy in GPS. Previous data-driven methods regress and eliminate pseudorange errors using handcrafted intermediate labels. Unlike them, we propose an end-to-end GPS localization framework, E2E-PrNet, to train a neural network for pseudorange correction (PrNet) directly using the final task loss calculated with the ground truth of GPS receiver states. The gradients of the loss with respect to learnable parameters are backpropagated through a Differentiable Nonlinear Least Squares (DNLS) optimizer to PrNet. The feasibility of fusing the data-driven neural network and the model-based DNLS module is verified with GPS data collected by Android phones, showing that E2E-PrNet outperforms the baseline weighted least squares method and the state-of-the-art end-to-end data-driven approach. Finally, we discuss the explainability of E2E-PrNet.

Robotics and automation is one of crucial trend in smart manufacturing to improve production efficiency. Au-tomated guided vehicles (AGVs) are a type of mobile robot used for material handling and have become widely utilized to achieve transportation automation. The usage of multiple AGVs introduces potential risks, such as traffic conflicts and safety risk. To meet high production demands, numerous shop floors are set up for large-scale manufacturing. Thus, reasonable and efficient AGV scheduling is vital for real-world operations. This paper proposes an efficient and scalable prioritized planning algorithm for large-scale multiple-AGV scheduling problem in manufacturing. The algorithm sequentially addresses two primary sub-problems: job assignment and conflict-free routing. The results of job assignment dictate the routes taken by the AGVs. In job assignment, jobs are allocated sequentially based on their pickup times. In conflict-free routing, AGV priorities are predefined, ensuring that higher priority AGVs maintain their movement while adjustments are made only to lower priority AGV plans when conflicts arise. Simulation is conducted on two real shop floor layouts and demonstrates the effectiveness and high efficiency of proposed algorithm. Even in a large-scale layout with 500 jobs and 20 AGVs, the computation time is only around 21 seconds.

Human-Robot Collaboration (HRC) is essential for enhancing productivity and flexibility in smart manufacturing, which poses requirements on accurately predicting the future movements of human operators, especially the trajectories of their upper limbs. However, existing model-based studies on human manipulation prediction lacks consideration of stochasticity and variability while the emerging deep learning-based methods are demanding on data size, which yet makes real-time deployment challenging. Therefore, combining the advantages of both model-based and deep learning-based methods, a method for predicting human arm motion, specifically, the position of a worker's wrist in less than 0.5 second, is proposed by hybridizing a Long Short-Term Memory (LSTM) network with an Inverse Kinematics (IK) model. Using historical coordinate sequences of the wrist joint in three-dimensional space in the past multiple frames as input, a neural network is trained to output the predicted coordinates of the wrist joint for the next frame. Then IK (Inverse Kinematics) is used to calculate the arm's motion trajectory based on the predicted wrist coordinates. As the predicted wrist coordinates are sequentially used as the input for the next prediction cycle, the prediction is realized over a sliding time window. Evaluation was conducted using both proprietary and open datasets, results demonstrated that our LSTM-IK method achieved high prediction accuracy, with an average distance error of approximately 5 cm, and can adapt to various task scenarios and individual differences. Additionally, comparison with ground truth illustrated the model's ability to handle complex motion patterns, even with partial occlusions or rapid movements.

The multi-user Augmented Reality (AR) is powered by shared mapping and localization obtained from Visual Inertial Odometry (VIO) using AR sensors, including cameras and Inertial Measurement Units (IMU). However, VIO is vulnerable to sparse environment features, low lighting conditions, and dynamic motions. The Ultra-Wideband (UWB) transceiver, as a radio-sensing modality robust to visual and dynamic defects, has been considered as a formfitting patch on VIO to flatter multi-user AR. Nevertheless, like other wireless sensors, UWB suffers from noise and interference. Therefore, how to fuse UWB and VIO for multi-user AR is a promising but challenging research direction. To facilitate this process, we designed and released a tool, UarLogger, to log the relative location measurements from UWB and AR sensors mounted on iOS devices, as well as context-related data. We provide two examples–environmental condition evaluation and sensor fusion–to demonstrate its usefulness and showcase how it can boost the development of new algorithms with daily devices in hand.

This poster introduces an Augmented Reality (AR)-assisted framework to help exclude Non-Line-of-Sight (NLOS) signals from the Global Navigation Satellite Systems (GNSS). We developed an AR mobile app named GnssQuest, augmenting the user's real-time camera view with a visualization of GNSS satellites. Our real-world experiment demonstrates that GnssQuest helps users to exclude NLOS satellites blocked by surrounding buildings, leading to significant improvements in GNSS positioning performance.

We present a neural network for mitigating pseudoranges errors to improve localization performance with data collected from mobile phones. A satellite-wise Multilayer Perceptron (MLP) is designed to regress the pseudorange error correction from six satellite, receiver, context-related features derived from Android raw Global Navigation Satellite System (GNSS) measurements. To train the MLP, we carefully calculate the target values of pseudorange errors using location ground truth and smoothing techniques and optimize a loss function involving the estimation residuals of smartphone clock offsets. The corrected pseudoranges are then used by a model-based localization engine to compute locations. The Google Smartphone Decimeter Challenge (GSDC) dataset, which contains Android smartphone data collected from both rural and urban areas, is utilized for evaluation. Both fingerprinting and cross-trace localization results demonstrate that our proposed method outperforms model-based and state-of-the-art data-driven approaches.

To further the penetration level of renewable energy sources (RESs) in power systems, the paper integrates wind turbines into conventional load frequency control (LFC). A resilient model predictive control (MPC) framework is constructed in the context of potential false data injection (FDI) attacks on vulnerable communication networks of multi-area power systems. To reduce the power generation cost, an economic cost function for MPC is firstly formulated. Then, a decentralized-model-based$\chi^{2}$detection unit is presented to distinguish the attacked measurements sent from neighbors. Moreover, to reduce the computation burden of executing the distributed MPC strategy, an intensified event-triggered scheme that can handle incomplete and inaccurate modeling issues is proposed. Validation results illustrate the efficacy of the detection unit and the intensified event-triggered scheme, and conclude the relationships between alarming thresholds and key performance indicators.Note to Practitioners—This paper explores the applicability of LFC with the integration of wind turbines under economic MPC framework. Motivated by the underlying FDI attacks on vulnerable communication networks among different control areas, an intrusion detection unit is proposed to install at each controller side to realize resiliency enhancement. Different from the existing works, this paper meticulously investigates the relationships between alarming thresholds and key performance indicators (KPIs), aiming at providing some valuable references for power operators and managers. Besides, this paper proposes an intensified event-triggered scheme to relieve the computation burden of MPC algorithm. This intensified event-triggered scheme has two advantages. One is that it considers the historic released signals in event-triggered conditions, which makes sure the critical signals at crests or troughs of frequency dynamic curves can be triggered. The other advantage is that the supplementary event-trigged condition can well tolerant the incomplete and inaccurate modeling problems existing in conventional event-triggered conditions. Simulations verify the efficacy and feasibility of the resilient event-triggered MPC strategy for frequency regulation under FDI attacks.

Automated Guided Vehicles AGVs play a vital role in enhancing productivity and efficiency within factory environments. However, their safe and effective operation heavily relies on the ability to navigate through complex spaces while avoiding obstacles. The significance of obstacle avoidance in AGV systems is emphasized, considering its impact on ensuring smooth material flow, minimizing collision risks, and optimizing production processes. The existing state of obstacle avoidance applications in factory settings reveals certain limitations and challenges. Current research and industrial implementations often rely on rule-based approaches or predefined paths, which may not adequately adapt to dynamic environments or unexpected obstacles. Additionally, some methods lack the ability to handle diverse obstacle types or efficiently plan optimal paths, leading to sub-optimal navigation or reduced throughput. In response to these challenges, this study proposes a novel approach for dynamic obstacle avoidance of AGVs based on deep reinforcement learning. By leveraging the Deep Deterministic Policy Gradient (DDPG) model, the AGV learns to make real-time decisions and navigate through dynamic obstacles effectively. The integration of deep neural networks with the actor-critic framework enables the AGV to learn and adapt optimal policies for obstacle avoidance in real-time, overcoming the limitations of rule-based methods. Simulation experiments are conducted to validate the performance and feasibility of the proposed approach. The results demonstrate that the DDPG-based method allows the AGV to successfully navigate through both dynamic and static obstacles in a dynamic environment, improving safety and efficiency in intelligent manufacturing applications.

To facilitate 5G-based positioning applications, Release 16 of the 3GPP 5G standard has defined the Positioning Reference Signal (PRS), which can be used to measure Time of Arrival (TOA) for downlink positioning. However, Orthogonal Frequency Division Multiplexing (OFDM) signals are sensitive and vulnerable to synchronization errors. Moreover, the highly configurable 5G PRS in Release 16 calls for a unique allocation pattern on the subcarriers. Existing timing recovery methods that have been employed for reference signals, which are evenly inserted in the subcarrier symbols, may not perform well. To solve the timing recovery issue of the OFDM signal through 5G standard-compliant PRS, we propose a three-stage timing recovery scheme. We use the 5G PRS as pilot symbols to estimate the path time delay and complete receiver sampling clock synchronization. We propose a generalized path time delay estimation method that can correct timing errors larger than one sample. In addition, we incorporate a delay-locked loop (DLL) that can track the PRS code-phase when the phase errors are within one sample, which showcases the precise positioning possible with a standard-compliant 5G New Radio (NR) signal.

Deep neural networks have achieved state-of-the-art performance in a wide range of recognition/classification tasks. However, when applying deep learning to real-world applications, there are still multiple challenges. A typical challenge is that unknown samples may be fed into the system during the testing phase and traditional deep neural networks will wrongly recognize the unknown sample as one of the known classes. Open set recognition is a potential solution to overcome this problem, where the open set classifier should have the ability to reject unknown samples as well as maintain high classification accuracy on known classes. The variational auto-encoder (VAE) is a popular model to detect unknowns, but it cannot provide discriminative representations for known classification. In this paper, we propose a novel method, Conditional Gaussian Distribution Learning (CGDL), for open set recognition. In addition to detecting unknown samples, this method can also classify known samples by forcing different latent features to approximate different Gaussian models. Meanwhile, to avoid information hidden in the input vanishing in the middle layers, we also adopt the probabilistic ladder architecture to extract high-level abstract features. Experiments on several standard image datasets reveal that the proposed method significantly outperforms the baseline method and achieves new state-of-the-art results.

We present a scalable design for accelerating the problem of solving a dense linear system of equations using LU Decomposition. A novel systolic array architecture that can be used as a building block in scientific applications is described and prototyped on a Xilinx Virtex 6 FPGA. This solver has a throughput of around 3.2 million linear systems per second for matrices of size N=4 and around 80 thousand linear systems per second for matrices of size N=16. In comparison with similar work, our design offers up to a 12-fold improvement in speed whilst requiring up to 50% less hardware resources. As a result, a linear system of size N=64 can be implemented on a single FPGA, whereas previous work was limited to a size of N=12 and resorted to complex multi-FPGA architectures to scale. Finally, the scalable design can be adapted to different sized problems with minimum effort.

Second order Quadratic Programming (QP) solvers such as interior-point method (IPM) require the solution of a system of linear equations at every iteration and could be a factor limiting the implementation of IPM to miniaturized devices or embedded systems. In contrast, first order QP solvers such as alternating direction method of multipliers (ADMM) does not require the solution of a system of linear equations. Thus first order QP solver is cheaper and easier to be implemented in embedded systems such as FPGA which has limited hardware resources. In this paper an FPGA implementation of ADMM which solves QP problems arising from Moving Horizon Estimation is proposed to demonstrate the "MHE on a Chip" idea. Our design has been implemented in both fixed-point and floating point arithmetic on the Xilinx Zynq-7000 XC7Z020-1CLG484C AP SoC and clocks at 50 MHz.

The use of wireless sensor networks for indoor localization application has emerged as a significant area of interest over the last decade, primarily motivated by its low cost and convenient deployment. The weighted centroid localization algorithm is a suitable positioning technique in a wireless sensor network due to its easy implementation. However, the performance of this method is easily affected by outliers and interference in the measurement of radio signal strength. In order to overcome this limitation, a more robust ARMA filter using generalized t-distribution noise model based on influence function approach is proposed. A hardware prototype was implemented to demonstrate that the ARMA filter could improve system performance, especially when dealing with the case of measurement outliers.

Simulations for the quadratically-constrained model predictive control (qc-MPC) with power system linear models are studied in this work. In qc-MPC, the optimization is imposed with two additional constraints to achieve the closed-loop system stability and the recursive-feasibility simultaneously. Instead of engaging the traditional terminal constraint for MPC, both constraints in qc-MPC are imposed on the first control vector of the MPC control sequence. As a result, qc-MPC has the potential for further extension to the control of network centric power systems. The algorithm of qc-MPC has been developed in a previous paper. Here, simulation studies with small-signal linear models of three typical power systems are presented to demonstrate its efficacy. We also develop a computational strategy for the decentralized static state-feedback control using the same quadratic dissipativity constraint as of the qc-MPC. Only state constraints are considered in the state feedback design. A comparison is then provided in the simulation study of qc-MPC relatively to the constrained-state feedback control.

This paper presents a development for the model predictive control (MPC) of nonlinear systems employing the quadratic dissipativity constraint (QDC). In this QDC strategy for nonlinear input-affine systems, a compound output vector is engaged to the supply rate such that the stability condition based on linear matrix inequality (LMI) can be rendered for nonlinear systems. The compound vector shares similar properties of the so-called manifest variable defined in the behaviourial framework for dynamical systems. Unlike linear systems, the LMI-based condition for nonlinear systems has not been found widespread used in the control literature. The present method introduces an application of such condition to nonlinear systems in this paper. In conjunction with QDC, the MPC recursive feasibility is achievable by having a bounded condition on the local divergence of the Lyapunov function. Both the storage function and supply rate of the dissipation inequality are parameterized in this development. The multiplier matrices need to be re-computed at every time step in this approach. Numerical simulation with a network of two chemical reactors has demonstrated the success of QDC approach in MPC, employing the compound output vector.

Quick and accurate error-rate prediction of Probabilistic CMOS (PCMOS) circuits is crucial for their systematic design and performance evaluation. While still in the early stage of research, PCMOS has shown potential to drastically reduce energy consumption at a cost of increased errors. Recently, a methodology has been proposed which could predict the error rates of cascade structures of blocks in PCMOS. This methodology requires error rates of unique blocks to predict the error rates of multiblock cascade structures composed of these unique blocks. In this article we present a new model for characterization of probabilistic circuits/blocks and present a procedure to find and characterize unique circuits/blocks. Unlike prior approaches, our new model distinguishes distinct filtering effects per output, thereby improving prediction accuracy by an average of 95% over the prior art by Palem and coauthors. Furthermore, we show two models where our new model with three stages is 18% more accurate, on average, than our simpler two-stage model. We apply our proposed models to Ripple Carry Adders and Wallace Tree Multipliers and show that using our models, the methodology of cascade structures can predict error rates of PCMOS circuits with reasonable accuracy (within 9%) in PCMOS for uniform voltages as well as multiple voltages. Finally, our approach takes seconds of simulation time whereas using HSPICE would take days of simulation time.

The paper describes a Yoga training system that is built based on motion replication technique (MoRep), including hardware, virtual scenario and feedback design. The motion replication technique proposed here can determine the similarity between Yoga master and student's postures and then provide feedback on the incorrect body posture of the student through multimodal channels. The key innovations of this project are also discussed.

It is well-known that faster sampling increases computational load but gives better performance. Multiplexed Model Predictive Control (MMPC) has been proposed recently. Its motivation was to reduce real-time computational load. The reduction in computational load can be used gainfully to increase sampling rate and improve performance. Hence, in this paper, we derive a formula to compute the Integral-Square-Error (ISE) performance of a MMPC controlled system. Given the plant and disturbance models, the ISE formula derived allows one to investigate how the ISE changes with control design parameters, such as the sampling interval and control weighting. This enables one to select, for example, a suitable sampling interval for the MMPC design to achieve the desired ISE performance. In addition, we validated the ISE formula on a multizone semiconductor manufacturing thermal process.

Replication of human body motion is a very important means to maintain a subject's emotion, knowledge and experience. The replication process requires accurate motion capturing system with sensor technologies to measure postures of human bodies and posture transmission, as well as feedback systems to adjust postures to fit into targeted ones. The sensing and feedback technologies are fundamental building blocks of motion capturing systems, work training system and rehabilitation systems. In particular, the construction of sensing and feedback systems for dynamic postures is much more complicated than that for static postures in terms of the time evolution and non-ridge body. We believe that dynamic postures can be represented by a set of blueprint or code like trajectories of particle of human body movement. Furthermore, we derive that the sensing and feedback systems should be able to establish to directly measure those critical particles without relying on external infrastructures. In the paper, some of those kinds of sensing and feedback devices are presented and some evidences such as feature contours are obtained through analysis of captured data by those devices in order to prove our estimation. We confess that our work is preliminary for this new field, but we hope that the work presented here can lead more efforts to bring out systematic approaches of feature detection and extraction of human postures whose characteristics are different from those of video and audio.

Body Sensor Networks (BSN) is a low-power wireless sensing technology for healthcare applications. Low-power wireless sensing is achieved by the integration of various miniaturized low-power on-chip devices, and is termed Body Sensor Node. The heart of the Body Sensor Node is an ultra low power mixed signal microcontroller. These mixed signal microcontrollers are typically resource-constrained - limited computational speed and memory is available. As a result, complex signal processing algorithms are unlikely to be implemented on these resource-constrained platforms. In this paper, we benchmarked a Fast Fourier Transform (FFT) algorithm, on a Texas Instruments MSP430F1612. The benchmarking results are reported and discussed.

Probabilistic CMOS is considered a promising technology for future generations of computing devices. By embracing possibly incorrect calculations, the technology makes it possible to trade correctness of circuit operations for potentially significant energy saving. For systematic design of probabilistic circuits, accurate mathematical models are indispensable. To this end, we propose a model of probabilistic ripple-carry adders. Compared to existing models, ours is applicable under a wide range of noise assumptions, including the popular additive-noise assumption. Our model provides recursive equations that can accurately capture propagation of carry errors. The proposed model is validated by HSPICE simulation, and we find that the model is able to predict multi-bit error-rates of a simulated probabilistic ripple-carry adder with reasonable accuracy.

Failures of critical factory equipment are one of the main reasons for production stoppages and a regular maintenance schedule is required to ensure continuous operation of production line. However, regular maintenance can be both costly and inefficient. Vibrations are present in all machinery with moving parts and it had always been regarded as an indicator of the health and condition of rotating machinery. With advances in technology, what used to be a mechanic's hunch can now be measured and with reasonable accuracy. This paper present a proof-of-concept implementation of machine diagnostic on chip using an intelligent wireless sensor node, capable of processing and analyzing vibration signals from Micro-electromechanical (MEMS) sensors. Special design considerations were also made to allow the wireless sensor node to be reconfigurable and modular at both the hardware and software level. A custom fixed-point library was implemented to tackle accuracy issues in dealing with small numbers without the use of floating point numbers in resource scarce platforms.

The objective is to develop a control system that will enable the Membrane Distillation Bio-Reactor (MDBR) water recycling plant to become self sufficient and fully automatic. In order to ensure continuous operation, the control system must maintain conditions required for the micro-organisms to survive. These requirements need to be met even when solar radiation and weather conditions are not conducive for water production. Hence, maintenance of a back up reserve is necessary. A balance has to be struck between minimization of power consumed and maximization of output produced to facilitate smooth operation. Control is implemented through a hierarchical framework that keeps track of the different plant objectives. Operational set points are determined through convex steady state optimization. Current set points are based on predictions regarding variation in solar radiation within the planning horizon.

Temperature uniformities within a wafer and from wafer to wafer have significant impact on the smallest feature size or critical dimension of integrated circuits. These are important issues with stringent specifications. To obtain temperature uniformity, a wafer is heated by multiple independently controlled heating elements simultaneously. The accuracy of temperature sensing is hence an important issue. In this paper, sensor bias is estimated for the difficult problem where measurement outliers are close to good data such that they cannot be separated easily. Equations are derived to predict the variance of the estimates from sample size. This information enables the selection of an efficient estimator. Sensor bias estimation efficiency translates into earlier bias removal and less faulty wafers. The theory is verified experimentally in a multi-zone thermal system for semiconductor wafer processing.

Probabilistic CMOS is considered to be a promising technology for substantial energy savings for computing devices, such as DSPs and graphics chips. The basic principle is to relax the energy requirement by allowing possibly incorrect computation results. For devices with probabilistic components, energy should be assigned to each component wisely, in order to achieve a good trade-off between energy consumption and correctness of the outputs. Recently, a few schemes have been proposed for energy assignment of ripple-carry adders, which are often based on intuitive arguments. In the present paper, we extend the idea of energy assignment to probabilistic multipliers. We focus on a fundamental type of multipliers, known as array multipliers. We derive some analytical results. Guided by these results, we devise an energy assignment scheme. We also find that energy assignment for array multipliers and ripple-carry adders can be quite different, due to differences in their structures. To our best knowledge, our work here is the first attempt in the literature to consider energy assignment for multipliers. Some examples, including digital image enhancement, are presented to demonstrate the effectiveness of the proposed scheme.

MPC outperforms other control strategies through its ability to deal with constraints. This requires on-line optimization, hence computational complexity can become an issue when applying MPC to complex systems with fast response times or to embedded applications where computational resources are limited. Multiplexed MPC (MMPC) has been proposed as a strategy to reduce computational complexity and stability results for MMPC have also been established. It has been suggested that the MMPC strategy of distributing the control moves over a complete update cycle, in contrast to conventional MIMO MPC which updates all the control variables simultaneously in one update cycle, may result in improved performance, despite finding sub-optimal solutions to the original problem. In this paper, we show that MMPC can be interpreted as a piecewise linear periodic state feedback law in the augmented state space. This provides a framework for systematic analysis and investigation of the properties of MMPC. In particular, the formula for computing the quadratic cost of MMPC is derived and illustrated through a numerical example.

With the rapid advances in technologies such as MEMS sensors, low-power embedded processing and wireless networking, sensor networks are becoming more powerful in terms of data acquisition and processing capabilities. Sensor networks can now be deployed in the physical world for various important applications such as environmental monitoring, weather monitoring and modeling, military surveillance, healthcare monitoring, tracking of goods and manufacturing processes, smart homes and offices, etc.

In this paper, the methods to rapidly deploy closed-loop control systems are presented. A flexible real-time embedded platform based on reconfigurable computing technologies is established, on which control blocks consisting of optimized control algorithms are set up. By employing control blocks, a set of tools aiming to shorten the development cycle of embedded control systems are developed. Compare to conventional ways, the tools give controller developers much faster ways to construct required controllers with higher flexibility.

In this article, the work of the NTU-JTC joint project entitled "Development of an intelligent lift door monitoring system for preemptive maintenance" is described. The objective of the project is to develop an intelligent system that could predict imminent lift breakdown. The basic idea is to develop a computerized system to monitor the performance of the lift door system and alert the user when the performance has deteriorated to a level when servicing is necessary, but before a lift breakdown actually occur. Such a system can provide not only a more efficient utilization of maintenance resources but also, at the same time, a reduction of the actual lift downtime in the flatted factories managed by JTC. The on site testing result of the prototype demonstrates the effectiveness of the system.

With the advances of high strength/light weight composite material, high performance magnetic bearings, and power electronics technology, flywheel energy storage systems (FESS) are becoming an exciting alternative to traditional battery systems. One of the challenging problems of the FESS is to stabilize the rotor which is very sensitive to outside disturbances and plant uncertainties. In this paper, a stabilizing synchronization design of the FESS is proposed by incorporating cross-coupling technology into the optimal control architecture, which can be decomposed into two problems: a robust optimal control problem to improve the synchronization performance of the rotor in the radial directions and a stability problem. The control scheme is based on minimization of a new quadratic performance index in which the synchronization errors are embedded. Stability of the control scheme is investigated through linear quadratic Gaussian (LQG) optimal control technique. It is shown that with adequate control parameters the resulting control system can provide satisfactory synchronization performance, and the closed-loop stability can be guaranteed theoretically. Simulations on a compact and efficient flywheel energy storage system with integrated magnetic bearings demonstrate that the proposed approach is very effective to recover the unstable system when outside disturbances are present.

Prostate boundary detection from ultrasound images plays a key role in prostate disease diagnoses and treatments. Due to the poor quality of ultrasound images, however, this still remains as a difficult task. Currently, boundary detection are performed manually, which is arduous and heavily user dependent. This paper presents a new approach derived from level set method to semiautomatically detect the prostate surface from 3D transrectal ultrasound images. In this method, a few initial bubbles are simply specified by the user from five particular slices based on the prostate shape. When bubbles evolve, they expand, shrink merge and split, and finally produce the desired prostate surface. To remedy the "boundary leaking" problem caused by gaps or weak boundaries, both region information and statistical intensity distribution are incorporated into the model. We applied the proposed method to eight 3D TRUS images and the results have shown its effectiveness.