Monte Carlo planning for active object classification

Classifying objects in complex unknown environments is a challenging problem in robotics and is fundamental in many applications. Modern sensors and sophisticated perception algorithms extract rich 3D textured information, but are limited to the data that are collected from a given location or path. We are interested in closing the loop around perception and planning, in particular to plan paths for better perceptual data, and focus on the problem of planning scanning sequences to improve object classification from range data. We formulate a novel time-constrained active classification problem and propose solution algorithms that employ a variation of Monte Carlo tree search to plan non-myopically. Our algorithms use a particle filter combined with Gaussian process regression to estimate joint distributions of object class and pose. This estimator is used in planning to generate a probabilistic belief about the state of objects in a scene, and also to generate beliefs for predicted sensor observations from future viewpoints. These predictions consider occlusions arising from predicted object positions and shapes. We evaluate our algorithms in simulation, in comparison to passive and greedy strategies. We also describe similar experiments where the algorithms are implemented online, using a mobile ground robot in a farm environment. Results indicate that our non-myopic approach outperforms both passive and myopic strategies, and clearly show the benefit of active perception for outdoor object classification.     

Sequential Monte Carlo for Bayesian matching of objects with occlusions

We consider the problem of locating instances of a known object in a novel scene by matching the fiducial features of the object. The appearance of the features and the shape of the object are modeled separately and combined in a Bayesian framework. In this paper, we present a novel matching scheme based on sequential Monte Carlo, in which the features are matched sequentially, utilizing the information about the locations of previously matched features to constrain the task. The particle representation of hypotheses about the object position allow matching in multimodal and cluttered environments, where batch algorithms may have convergence difficulties. The proposed method requires no initialization or predetermined matching order, as the sequence can be started from any feature. We also utilize a Bayesian model to deal with features that are not detected due to occlusions or abnormal appearance. In our experiments, the proposed matching system shows promising results, with performance equal to batch approaches when the target distribution is unimodal, while surpassing traditional methods under multimodal conditions. Using the occlusion model, the object can be localized from only a few visible features, with the nonvisible parts predicted from the conditional prior model.     

State explosion in almost-sure probabilistic reachability

We show that the problem of reaching a state set with probability 1 in probabilistic-nondeterministic systems operating in parallel is EXPTIME-complete. We then show that this probabilistic reachability problem is EXPTIME-complete also for probabilistic timed automata.     

序贯Monte Carlo算法仿真结果分析

序贯Monte Carlo方法能够解决很多实际问题.它的系统模型与Kalman滤波算法相比具有更广泛的适用性,所以研究Monte Carlo方法是很有实际意义的.文中对序贯Monte Carlo算法进行性能分析,对这一方法的跟踪能力进行了仿真实验.采用的仿真系统模型是非线性系统模型.仿真实验比较了EKF、SIS、SIR算法的性能.通过对不同算法的仿真结果之间的分析和比较,得出了有意义的结论.这对一些工程问题的解决是有重要意义的.     

似然值波门SMC-PHD滤波器

概率假设密度(Probability Hypothesis Density, PHD)滤波器的序贯蒙特卡罗(Sequential Monte Carlo, SMC)实现需要大量的粒子。为了解决其计算的有效性,本文提出一种改进的SMC-PHD滤波器,称之为似然值波门SMC-PHD滤波器。首先,以所有预测粒子为依据,利用全部的多目标后验信息,最大限度地确认出所有目标生成的观测。其次,基于校正器中所有预测粒子的似然值,避免为粒子贴标签以及传统的距离计算,使得算法在各种应用中易于实现,只有有效观测才参与粒子权值的更新。最后,与基本SMC-PHD滤波器相比,其优秀的实时性和更好的滤波精度通过仿真得到证实。     

Structure from Motion Using Sequential Monte Carlo Methods

In this paper, the structure from motion (SfM) problem is addressed using sequential Monte Carlo methods. A new SfM algorithm based on random sampling is derived to estimate the posterior distributions of camera motion and scene structure for the perspective projection camera model. Experimental results show that challenging issues in solving the SfM problem, due to erroneous feature tracking, feature occlusion, motion/structure ambiguity, mixed-domain sequences, mismatched features, and independently moving objects, can be well modeled and effectively addressed using the proposed method.     

A Sequential Monte Carlo Method for Bayesian Analysis of Massive Datasets

Markov chain Monte Carlo (MCMC) techniques revolutionized statistical practice in the 1990s by providing an essential toolkit for making the rigor and flexibility of Bayesian analysis computationally practical. At the same time the increasing prevalence of massive datasets and the expansion of the field of data mining has created the need for statistically sound methods that scale to these large problems. Except for the most trivial examples, current MCMC methods require a complete scan of the dataset for each iteration eliminating their candidacy as feasible data mining techniques.In this article we present a method for making Bayesian analysis of massive datasets computationally feasible. The algorithm simulates from a posterior distribution that conditions on a smaller, more manageable portion of the dataset. The remainder of the dataset may be incorporated by reweighting the initial draws using importance sampling. Computation of the importance weights requires a single scan of the remaining observations. While importance sampling increases efficiency in data access, it comes at the expense of estimation efficiency. A simple modification, based on the rejuvenation step used in particle filters for dynamic systems models, sidesteps the loss of efficiency with only a slight increase in the number of data accesses.To show proof-of-concept, we demonstrate the method on two examples. The first is a mixture of transition models that has been used to model web traffic and robotics. For this example we show that estimation efficiency is not affected while offering a 99% reduction in data accesses. The second example applies the method to Bayesian logistic regression and yields a 98% reduction in data accesses.     

基于序列蒙特卡罗滤波的人脸跟踪算法

结合图像的色彩分布和空间布局,提出了一种基于HSV色彩和空间信息的序列蒙特卡罗滤波人脸跟踪算法.通过比较采样值和期望值的特征距离来计算采样状态对应的权值.利用加权采样值来估计未知量后验概率,当采样数趋于无穷时,由大数定理保证采样值分布逼近于真实值分布.仿真实验给出了利用加权采样对人脸跟踪的结果.实验表明,基于序列蒙特卡罗的人脸跟踪算法计算简单有,M能够准确预测人脸的位置并且很好地跟踪其运动轨迹.     

基于序列蒙特卡罗滤波的车辆跟踪算法研究

针对非线性目标跟踪中模型或函数近似等最优估计缺陷问题,提出了基于帧间预测和特征匹配的序列蒙特卡罗滤波跟踪算法.算法中采用在HSV色彩下的空间加权直方图描述跟踪车辆的状态特征,通过简单的随机漂移模型实现估测样本的帧间传递,利用估测样本与期望目标间的相似度量完成样本权重赋值运算,最终利用加权样本值估计实现待测目标的后验状态.实验结果表明,基于序列蒙特卡罗滤波的车辆跟踪算法计算简单有效,能够在复杂环境下实时、准确跟踪道路上无规律、非线性运动的车辆,并能够有效适应车辆部分遮挡和短时丢失等情况.     

一种基于序贯Monte Carlo算法的交通事件重构方法

为了能准确地重构出当前道路场景中的交通流事件,需要合适的模型与方法以及能够代表交通流状态的实时数据。基于交通流非线性非高斯的特点,提出了一种基于序贯Monte Carlo方法的交通流堵塞事件重构模型。提出的模型能够不断的同化真实道路上实时的传感器数据使仿真中的交通流状态与真实路况不断接近。通过分析仿真数据推测出当前真实道路上的堵塞事件及其相关属性,并据此在仿真环境中模拟堵塞,进而实现对真实道路上堵塞事件的重构。理论研究和实验结果表明该模型能够根据重构结果评估当前的道路状况,合理推测引起拥堵的位置与堵塞范围。     

Sequential Monte Carlo Adaptation in Low‐Anisotropy Participating Media

This paper presents a novel method that effectively combines both control variates and importance sampling in a sequential Monte Carlo context. The radiance estimates computed during the rendering process are cached in a 5D adaptive hierarchical structure that defines dynamic predicate functions for both variance reduction techniques and guarantees well‐behaved PDFs, yielding continually increasing efficiencies thanks to a marginal computational overhead. While remaining unbiased, the technique is effective within a single pass as both estimation and caching are done online, exploiting the coherency in illumination while being independent of the actual scene representation. The method is relatively easy to implement and to tune via a single parameter, and we demonstrate its practical benefits with important gains in convergence rate and competitive results with state of the art techniques.     

Evolutionary Optimization of Dynamics Models in Sequential Monte Carlo Target Tracking

This paper describes a new method for the online parameter optimization of various models used to represent the target dynamics in particle filters. The optimization is performed with an evolutionary strategy algorithm, by using the performance of the particle filter as a basis for the objective function. Two different approaches to forming the objective function are presented: the first assumes knowledge of the true source position during the optimization, and the second uses the position estimates from the particle filter to form an estimate of the current ground-truth data. The new algorithm has low computational complexity and is suitable for real-time implementation. A simple and intuitive real-world application of acoustic source localization and tracking is used to highlight the performance of the algorithm. Results show that the algorithm converges to an optimum tracker for any type of dynamics model that is capable of representing the target dynamics.      

Partial Linear Gaussian Models for Tracking in Image Sequences Using Sequential Monte Carlo Methods

The recent development of Sequential Monte Carlo methods (also called particle filters) has enabled the definition of efficient algorithms for tracking applications in image sequences. The efficiency of these approaches depends on the quality of the state-space exploration, which may be inefficient due to a crude choice of the function used to sample in the associated probability space. A careful study of this issue led us to consider the modeling of the tracked dynamic system with partial linear Gaussian models. Such models are characterized by a non linear dynamic equation, a linear measurement equation and additive Gaussian noises. They allow inferring an analytic expression of the optimal importance function used in the diffusion process of the particle filter, and enable building a relevant approximation of a validation gate. Despite of these potential advantages partial linear Gaussian models have not been investigated. The aim of this paper is therefore to demonstrate that such models can be of real interest facing difficult usual issues such as occlusions, ambiguities due to cluttered backgrounds and large state space. Three instances of these models are proposed. After a theoretical analysis, their significance is demonstrated by their performance for tracking points and planar objects in challenging real-world image sequences.     

Computational Issues in the Sequential Probit Model: A Monte Carlo Study

We discuss computational issues in the sequential probit model that have limited its use in applied research. We estimate parameters of the model by the method of simulated maximum likelihood (SML) and by Bayesian MCMC algorithms. We provide Monte Carlo evidence on the relative performance of both estimators and find that the SML procedure computes standard errors of the estimated correlation coefficients that are less reliable. Given the numerical difficulties associated with the estimation procedures, we advise the applied researcher to use both the stochastic optimization algorithm in the Simulated Maximum Likelihood approach and the Bayesian MCMC algorithm to check the compatibility of the results. JEL Classifications: C11, C15, C35, C63     

基于相对熵的概率假设密度滤波器序贯蒙特卡罗实现方式

概率假设密度滤波器的典型序贯蒙特卡罗实现方式与粒子滤波类似,均是利用大量加权粒子估计多目标状态,典型实现方式是为每个期望目标分配固定数目的粒子,这导致较大的算法时间开销.鉴于此,建立了基于相对熵的序贯蒙特卡罗实现方式.首先计算两个不同规模粒子集合的相对熵,与预设阈值进行比较以确定粒子数目,从而动态调整粒子数目.仿真结果表明,所提出的实现方式提高了跟踪效率,在大部分时间步上优于典型实现方式.     

基于序列蒙特卡罗方法的3D人体运动跟踪

针对人体运动跟踪的特点,在退火粒子滤波方法的基础上,提出基于序列蒙特卡罗方法的3D人体跟踪算法．通过状态空间分解提高了退火系数选择的鲁棒性;同时,在每次退火时采用PERM采样方法,而不是标准的重采样,能在一定程度上抑制观测模型与真实分布之间的误差,从而提高算法的稳定性．通过模拟实验表明,该算法适合3D多关节人体跟踪．     

New heuristics for planning operating rooms

We tackle the operating room planning problem of the Plastic Surgery and Major Burns Specialty of the University Hospital “Virgen del Rocio” in Seville (Spain). The decision problem is to assign an intervention date and an operating room to a set of surgeries on the waiting list, minimizing access time for patients with diverse clinical priority values. This problem has been previously addressed in the literature considering different objective functions. The clinical priority depends on the surgery priority and the number of days spent on the waiting list. We propose a set of 83 heuristics (81 constructive heuristics, a composite heuristic, and a meta-heuristic) based on a new solution encoding, and we compare these methods against existing heuristics from the literature for solving operating room planning problems. The heuristics are adapted to the problem under consideration (i.e. considering all constraints and the new objective function), being re-implemented using the information provided by the authors. In total, after a calibration procedure, we compare 17 heuristics. The computational experiments show that our proposed meta-heuristic is the best for the problem under consideration. Finally, the proposed heuristics are tested using data from the Plastic Surgery and Major Burns Specialty. The results show significant improvements on several key performance indicators (number of scheduled surgeries, quality of surgical plan, resources utilization, etc.) when comparing with the actual results obtained by the specialty in the current practice. The aforementioned hospital is currently implementing the heuristic methods.     

Heuristics for the probabilistic traveling salesman problem with deadlines based on quasi-parallel Monte Carlo sampling

The Probabilistic Traveling Salesman Problem with Deadlines (PTSPD) is a Stochastic Vehicle Routing Problem with a computationally demanding objective function. In this work we propose an approximation for that objective function based on Monte Carlo Sampling and using the novel approach of quasi-parallel evaluation of samples. We perform comprehensive computational studies that reveal the efficiency of this approximation. Additionally, we examine different Local Search Algorithms and present a Random Restart Local Search Algorithm for solving the PTSPD together with an extensive computational study on a large set of benchmark instances.