Real-time Smoothing Research Articles

Hi-ROS: Open-source multi-camera sensor fusion for real-time people tracking

This paper presents Hi-ROS (Human Interaction in ROS), an open source framework focused on real-time accurate assessment of human motion. The system offers a series of tools to track multiple people in real-time by exploiting a calibrated camera network. No assumptions are made about the typology or number of cameras, nor about the body pose estimation algorithm used to extract the 3D poses of the people in the scene. The tools provided by Hi-ROS include a Skeleton Tracker to ensure temporal consistency of the detected poses, a Skeleton Merger to fuse the tracks from multiple cameras, thus limiting flickering phenomena, a Skeleton Optimizer to ensure limb length consistency, and a Skeleton Filter to perform real-time smoothing of the detected joint trajectories. Accuracy, tracking robustness, and real-time performance of the proposed system were evaluated on a public dataset, containing both single-person and multi-person sequences with up to 4 people interacting. The results obtained using different subsets of the proposed tools show how the complete Hi-ROS pipeline provides accurate and reliable estimates also in challenging scenarios, with a reduction of the RMSE of up to 27% with respect to a pure tracking approach. This work aims to push forward the development of unobtrusive human–robot interaction applications, multi-person automated posture analyses, rehabilitation performance assessments, and any possible application enabled by real-time accurate assessment of human motion via markerless motion capture.

Real-time smoothing of G01 commands for five-axis machining by constructing an entire spline with the bounded smoothing error

G01 commands generated by computer-aided manufacturing (CAM) software need to be smoothed to eliminate the tangential discontinuities at corners. Although the smoothing methods for five-axis tool paths have been widely studied, it is still challenging to realize the high machining efficiency of short, dense G01 commands with real-time processing. This article proposes a real-time five-axis tool path smoothing algorithm for short G01 commands. The uniform smoothing of adjacent two corners with cubic uniform B-spline is developed. Through connecting the pieces of splines, all the G01 commands are smoothed by an entire B-spline. Therefore, the curvature increment induced by the avoidance of the intersection of adjacent micro-splines is prevented, and thus, high allowable feedrate and high machining efficiency can be achieved. Especially, the smoothing errors are theoretically proved to be constrained by using the convex polyhedron method. At the same time, the required number of control points is greatly reduced by sharing the control points with the adjacent G01 commands. Simulation and experimental results demonstrate that the proposed method can significantly reduce the machining time by more than 26% for short linear trajectories.

Kalman Filter-Based Data Fusion of Wi-Fi RTT and PDR for Indoor Localization

The Fine Time Measurement (FTM) protocol introduced by IEEE 802.11 includes a new ranging method, named Wi-Fi Round Trip Time (Wi-Fi RTT), which can be used for indoor localization. Pedestrian Dead Reckoning (PDR) can provide accurate pedestrian tracking through inertial sensors in a short time. Information fusion of PDR and existing wireless technology is widely used in indoor localization to ensure the robustness and stability. In this paper, we propose a fusion indoor localization method of Wi-Fi RTT and PDR. Firstly, an adaptive filtering system consisting of multiple Extended Kalman Filter (EKF) and a new outlier detection method is proposed to reduce the localization error of Wi-Fi RTT. Secondly, the fusion algorithm based on the Federated Filter (FF) and observability is designed to combine Wi-Fi RTT with PDR. Finally, to further improve the localization performance of the fusion algorithm, a real-time smoothing method with fixed interval is used. We evaluate the proposed method in four different scenarios. The results show that the proposed indoor localization method has better stability and robustness, and the average localization error decreased by 37.4-67.6% compared with the classic EKF-based method.

Real-Time Control Strategy for Aggregated Electric Vehicles to Smooth the Fluctuation of Wind-Power Output

Electric vehicles (EVs) are flexible demand-side response resources in a power distribution system. Reasonable and orderly control of charging/discharging processes of aggregated EVs can improve their coordination and interaction with the distribution system and ensure its efficient and stable operation. Aiming at the problem that the fluctuation of wind power output may affect the stable operation of distribution system, a real-time control strategy for aggregated EVs to smooth the fluctuation of wind power is proposed. Firstly, considering the dispatchability of EVs, the charging/discharging energy boundary model is established to determine the charging/discharging margin of an EV at each moment. Then, first-order low-pass filtering is used to determine the total dispatching power of aggregated EVs. Finally, the total charging power of aggregated EVs is determined and power allocation is carried out. Simulation results show that the proposed strategy can achieve real-time smoothing for the fluctuation of wind power output while meeting the charging requirements of EVs, and the proposed strategy can not only reduce the fluctuation rate of total load, but also realize peak shaving and valley filling for the distribution system.

Real-Time Path Smoothing for Mobile Robots in 2D and 3D Environments

Real-Time Path Smoothing for Mobile Robots in 2D and 3D Environments

Online motion smoothing for video stabilization via constrained multiple-model estimation

Video stabilization smooths camera motion estimates in a way that should adapt to different types of intentional motion. Corrective motion (the difference between smoothed and original motions) should be constrained so that black borders do not intrude into the (cropped) stabilized frames. Although offline smoothing can use all of the frames, online (real-time) smoothing can only use a small number of previous frames. In this paper, we propose an online motion smoothing method based on linear estimation applied to a constant-velocity model. We use estimate projection to ensure that the smoothed motion satisfies black-border constraints, which are modeled exactly by linear inequalities for general 2D motion models. We then combine the estimate projection with multiple-model estimation, which can adaptively smooth the camera motion in a probabilistic way. Experimental results show how the proposed algorithm can better smooth the camera motion and stabilize videos in real time.

Dynamic displacement estimation by fusing LDV and LiDAR measurements via smoothing based Kalman filtering

This paper presents a smoothing based Kalman filter to estimate dynamic displacement in real-time by fusing the velocity measured from a laser Doppler vibrometer (LDV) and the displacement from a light detection and ranging (LiDAR). LiDAR can measure displacement based on the time-of-flight information or the phase-shift of the laser beam reflected off form a target surface, but it typically has a high noise level and a low sampling rate. On the other hand, LDV primarily measures out-of-plane velocity of a moving target, and displacement is estimated by numerical integration of the measured velocity. Here, the displacement estimated by LDV suffers from integration error although LDV can achieve a lower noise level and a much higher sampling rate than LiDAR. The proposed data fusion technique estimates high-precision and high-sampling rate displacement by taking advantage of both LiDAR and LDV measurements and overcomes their limitations by adopting a real-time smoothing based Kalman filter. To verify the performance of the proposed dynamic displacement estimation technique, a series of lab-scale tests are conducted under various loading conditions.

2-4 空中操作インターフェースのための実時間平滑化フィルタの比較評価(第2部門 メディア工学(ヒューマンインタフェース))

2-4 空中操作インターフェースのための実時間平滑化フィルタの比較評価(第2部門 メディア工学(ヒューマンインタフェース))

Optimal Linear Fusion of Smoothed State Estimates

The work presented here deals with estimation fusion of smoothed state estimates. Two problems of fusion for smoothing are considered: fixed point and fixed interval. Optimal linear fusion rules in the sense of the optimal weighted least squares (OWLS) and the linear minimum mean-square error (LMMSE) are obtained. These rules are in recursive forms convenient for implementation. We also propose a more practical method for real-time smoothing, which in essence is fusing smoothed and filtered estimates. Illustrative numerical results are provided to verify the performance and credibility of the fusion rules.

Real-time smoothing of car-following data through sensor-fusion techniques

Observation of vehicles kinematics is an important task for many applications in ITS (Intelligent Transportation Systems). It is at the base of both theoretical analyses and application developments, especially in case of positioning and tracing/tracking of vehicles, car-following analyses and models, navigation and other ATIS (Advanced Traveller Information Systems), ACC (Adaptive Cruise Control) systems, CAS and CWS (Collision Avoidance Systems and Collision Warning Systems) and other ADAS (Advanced Driving Assistance Systems). Modern technologies supply low-cost devices able to collect time series of kinematic and positioning data with medium to very high frequency. Even more data can be (almost continually) collected if vehicle-to-vehicle (V2V) communications come true. However, some of the ITS applications (as well as car-following models, on which many ADAS and ACC are based) require highly accurate measures or, at least, smooth profiles of collected data. Unfortunately, even relatively high-cost devices can collect biased data because of many technical reasons and often this bias could lead to unrealistic kinematics, incorrect absolute positioning and/or inconsistencies between vehicles (e.g. negative spacing). As a consequence, data need filtering in most of the ITS applications. To this aim proper algorithms are required and several sensors and sources of data possibly integrated in order to obtain the maximum quality at the minimal cost. This work addresses the previous issues by developing a specific Kalman smoothing approach. The approach is developed in order to deal with car-following conditions but is conceived to take into account also navigation issues. The performances are analysed with respect to real-world car-following data, voluntarily biased for evaluation purposes. Assessment is carried out with reference to different mixtures of sensors and different sensors accuracies.

Real-time smoothing for network adaptive video streaming

Real-time streaming delivery over the Internet with bandwidth variation is a very challenging task. It is important to smooth the quality variability and improve the utilization of the available network bandwidth. In this paper, we propose a real-time optimal smoothing scheduling algorithm for network adaptive video streaming with the variable network bandwidth and packet loss. The algorithm adopts a rate-distortion optimized framework and real-time scheduling scheme to select and schedule the packets according to the network status. It attempts to minimize the quality variability at the client end while at the same time maximizing the utilization of the variable network bandwidth. Experiments show that, compared with frame-based scheduling algorithm, our proposed real-time smoothing algorithm improves and smoothes the quality in decoded video frames.

Online Smoothing of VBR H.263 Video for the CDMA2000 and IS-95B Uplinks

The variable bit rate nature of compressed video remains a major challenge to the transmission of real-time video over mobile CDMA cellular networks. Indeed, H.263 and MPEG-4 video bitstreams can exhibit high peak rates and frequent rate variations, which are difficult to support in 2.5G and 3G mobile networks. In this paper, we consider smoothing schemes to minimize the peak rate, variance, and average bandwidth of the transmitted signal, in order to obtain a lower bit-error-rate and thus, a higher received video quality than if unsmoothed video were transmitted. By introducing a delay in the decoding process, the video output stream can be buffered at the video decoder end: this buffering capability is used in order to perform real-time smoothing of the video streams. The transmitter chooses among the set of rates provided by the IS-95B or cdma2000 standards, in a mobility scenario, in order to support the rate variations. We simulate two end-to-end video communication systems based on IS-95B and cdma2000. They consist of an H.263 codec, an H.223 multiplexer, a rate management system and the physical layer components of the IS-95B and cdma2000 standards. The transmitted signal is subject to multipath fading and multiple-access interference, obtained by simulating all the other users in the cell. Results show a significant gain in the video quality obtained through smoothing of the video streams.

A smoothing method of lofargram using morphological opening filter (part 3)

Lofargram is a processing method for analyzing the temporal structure of passive sonar signals by spectral analysis. The authors have proposed to suppress the impulse noise on lofargram using morphological opening filter instead of averaging procedure which has been used generally. In our experimental results, the effective improvements of the processing gain and the time resolution have been obtained as compared with averaging procedure.However, a large amount of the calculation time of morphological opening filter occurred by the memory access bottleneck with software implementation has been an obstacle to the real-time processing. In this paper, the parallel algorithm of morphological opening filter for lofargram using the distributivity by the structuring element is proposed for the high speed calculation. This algorithm is inspected by the execution results using MIMD parallel machine. Then, the limitation of this method and the performance of the processor node for the real-time smoothing of lofargram are estimated.

モフォロジカルｏｐｅｎｉｎｇフィルタを用いたローファグラムの平滑化手法

Lofargram is a processing method for analyzing the temporal structure of passive sonar signals by spectral analysis. The authors have proposed to suppress the impulse noise on lofargram using morphological opening filter instead of averaging procedure which has been used generally. In our experimental results, the effective improvements of the processing gain and the time resolution have been obtained as compared with averaging procedure.However, a large amount of the calculation time of morphological opening filter occurred by the memory access bottleneck with software implementation has been an obstacle to the real-time processing. In this paper, the parallel algorithm of morphological opening filter for lofargram using the distributivity by the structuring element is proposed for the high speed calculation. This algorithm is inspected by the execution results using MIMD parallel machine. Then, the limitation of this method and the performance of the processor node for the real-time smoothing of lofargram are estimated.

Design of linear tracking filters via robust H/sub 2/ optimization

The design of a velocity-acceleration tracking filter is considered. The filler achieves real-time smoothing of noisy velocity measurements by using acceleration measurements. Robust H/sub 2/-filtering is used which keeps the bias that occurs in the velocity estimation in maneuvers small, for low-cost accelerometers that attain large scale factors. Three methods of robust H/sub 2/ estimation are applied, and their results, for a simple problem with low-cost accelerometer, are compared. The resulting filters are also compared with the traditional Kalman filter and with the robust H/sub /spl infin//-filter. The results of the present note should encourage navigation engineers to apply robust H/sub 2/-filtering techniques in problems where the measurement equations include significant parameter uncertainties.

A matched filter data smoothing algorithm

An efficient spatial smoothing algorithm for filtering data while preserving spatial detail is obtained using the system (impulse response) function of the sensor. In contrast to the normal procedure for determining the filter coefficients for an arbitrary system function, this technique does not involve the use of Fourier transforms. The algorithm results in optimal smoothing within the constraints of retaining good vertical detail after only three or four iterations. A refinement of this procedure involving higher-order filter functions produced the equivalent operation in a single pass. The spatial filtering coefficients are obtained analytically for the frequently applicable Gaussian system function. The efficiency of this procedure is illustrated by filtering simulation, logs, and spectral data. For real-time smoothing of nuclear log data, a filter length of five times the vertical resolution is required. >

Decentralized two-filter smoothing algorithms in discrete-time systems

We develop some decentralized fixed-interval smoothing algorithms which are realized by the two-filter form in linear discrete-time systems, where the decentralized estimation structure consists of a central processor and of two local processors. Two approaches based on local filtering and local smoothing are considered for solving the problems of decentralized smoothing, smoothing update, and real-time smoothing. It is then shown that the notion of decentralized Kalman filtering plays a significant role in the derivation of the algorithms for the former approach, and that the introduction of the Hamiltonian system in discrete-time immediately gives the solutions of the problems in the latter approach. Moreover, some characteristics of the algorithms arising from each approach are also discussed.

Continuous-time decentralized smoothers based on two-filter form: identical local and global models

Abstract This paper proposes some decentralized smoothing algorithms for a continuous-time linear estimation structure consisting of a central processor and of two local processors, in which the local models are assumed to be identical to the global model. The philosophy of the paper is to solve the problem in terms of the local forward and backward information (or Kalman) filters. The resulting algorithms are somewhat different from those based on the local smoothing estimates which have been studied by some other authors. Smoothing update and real-time smoothing algorithms are also presented, ft is shown that the present algorithms have some advantages: the global filtered estimates can be obtained in the course of computing the decentralized smoothing estimates and the central and local processors can be derived in a completely parallel fashion

Computer Assisted Data Acquisition and Analysis for Size Exclusion Chromatography

Abstract Two programs, ACEACQ and ACEANN, provide data acquisition and analysis for Size Exclusion Chromatography (SEC). ACEACQ is used for data logging, real-time smoothing and data reduction by boxcar averaging. ACEACQ requires that traditional SEC hardware be coupled with a Hewlett Packard data logger and microcomputer. A Mettler electronic balance is connected to the computer in a novel fashion to continuously monitor the mass of solvent eluted during an experiment. ACEANN provides analysis for the data collected by ACEACQ. Calculations include molecular weight averages, molecular hydrodynamic diameter averages and column separation efficiency parameters. Significant reductions in analysis time have been realized along with improvements in accuracy of data acquisition, analysis and experimental reproducibility.

The stochastic analysis of dynamic systems moving through random fields

In this paper we consider dynamic systems that move along specified trajectories across random fields, where the field acts as a driving force to the dynamic system. For a specific class of random fields we develop equations for the evolution of the covariance of the state of the dynamic system, and in the special case in which the trajectory is a straight line path followed by a 180° turn (i.e., an "over-and-back" trajectory) we develop a Markovian model that involves a change in the dimension of the state after the turn. For this case we also briefly discuss the estimation problem using recently developed results on "real-time smoothing."