Lingling Zhao

Interacting multiple model particle filter for prognostics of lithium-ion batteries

Abstract We propose a new data-driven prognostic method based on the interacting multiple model particle filter (IMMPF) for determining the remaining useful life (RUL) of lithium-ion (Li-ion) batteries and the probability distribution function (PDF) of the associated uncertainty. The method applies the IMMPF to different state equations. Modeling the battery capacity degradation is very important for predicting the RUL of Li-ion batteries. In this study, improvements are made on various Li-ion battery capacity models (i.e., polynomial, exponential, and Verhulst models). Further, three different one-step state transition equations are developed, and the IMMPF method is applied to estimate the RUL of Li-ion batteries with the use of the three improved models. The PDF of the predicted RUL is obtained by combining the PDFs obtained with each individual model. We conduct four case studies to validate the proposed method. The results are as follows: (1) the three improved models require fewer parameters than the original models, (2) the proposed prognostic method shows stable and high prediction accuracy, and (3) the proposed method narrows the uncertainty PDF of the predicted RUL of Li-ion batteries.

Online UAV path planning in uncertain and hostile environments

Taking uncertainties of threats and vehicles’ motions and observations into account, the challenge we have to face is how to plan a safe path online in uncertain and dynamic environments. We construct the static threat (ST) model based on an intuitionistic fuzzy set (A-IFS) to deal with the uncertainty of a environmental threat. The problem of avoiding a dynamic threat (DT) is formulated as a pursuit-evasion game. A reachability set (RS) estimator of an uncertain DT is constructed by combining the motion prediction with a RRT-based method. An online path planning framework is proposed by integrating a sub goal selector, a sub tasks allocator and a local path planner. The selector and allocator are presented to accelerate the path searching process. Dynamic domain rapidly-exploring random tree (DDRRT) is combined with the linear quadratic Gaussian motion planning (LQG-MP) method when searching local paths under threats and uncertainties. The path that has been searched is further improved by using a safety adjustment method and the RRT* method in the planning system. The results of Mont Carlo simulations indicate that the proposed algorithm behaves well in planning safe paths online in uncertain and hostile environments.

UAV online path planning algorithm in a low altitude dangerous environment

UAV online path-planning in a low altitude dangerous environment with dense obstacles, static threats (STs) and dynamic threats (DTs), is a complicated, dynamic, uncertain and real-time problem. We propose a novel method to solve the problem to get a feasible and safe path. Firstly STs are modeled based on intuitionistic fuzzy set (IFS) to express the uncertainties in STs. The methods for ST assessment and synthesizing are presented. A reachability set (RS) estimator of DT is developed based on rapidly-exploring random tree (RRT) to predict the threat of DT. Secondly a subgoal selector is proposed and integrated into the planning system to decrease the cost of planning, accelerate the path searching and reduce threats on a path. Receding horizon (RH) is introduced to solve the online path planning problem in a dynamic and partially unknown environment. A local path planner is constructed by improving dynamic domain rapidly-exploring random tree (DDRRT) to deal with complex obstacles. RRT* is embedded into the planner to optimize paths. The results of Monte Carlo simulation comparing the traditional methods prove that our algorithm behaves well on online path planning with high successful penetration probability.

A dual-level approach for lithium-ion battery RUL prognosis

A Dual-level approach is proposed for lithium-ion battery remaining useful life prognosis. The Dual-level approach is based on the Verhulst model, the logarithmic model and the PF. The first level prediction produces a rough capacity prediction and then the second level prediction improves the accuracy of the first prediction result. The logarithmic model has the ability to describe the accelerated degradation process more accurate. Due to it is sensitive to its parameters, in this article the result of the first prediction result based on the Verhulst model is used to initialize the parameters of the logarithmic model. The PF is used in both levels to update the state based on each model. Four case studies are conducted to validate the Dual-level approach. The experimental result shows that the Dual-level approach has efficiently improve the performance of the estimating RUL and the capacity fade trend of the lithium-ion battery compared with the prognosis method based on PF and the Verhulst model.

Online Creating an Improved UAV Path in Complex and Hostile Environments

We propose a cost based rapidly-exploring random tree method (CBRRT) to plan an improved path under complex obstacles and threats, limited deliberation time and constraints. The sampling space reduction of the dynamic domain rapidly-exploring random tree method (DDRRT) is improved and heuristics are applied, to guide the path tree to avoid threats and obstacles quickly, to rapidly find a low threat initial path. Meanwhile, the sampling space reduction is utilized to accelerate the path improving procedure of RRT* by limiting the path improving scope to appropriate areas. The reduction is constructed according to the real-time tree growth which provides the environmental information for DDRRT and RRT* as heuristic clues. The constraints and UAV motion are taken into account during the creation of the path. To make the path easier for UAV to follow, a cost based waypoints pruning method (CBP) is proposed, and the curvature-continuous Dubins curve is applied to smooth the path. The simulation results verify that CBRRT and CBP behave well in our environments.

An improved exponential model for predicting the remaining useful life of lithium-ion batteries

Prognostics and health management has become a subject of great interest to many electrical systems. However, the lithium-ion batteries are a core component of many machines and critical to systems functional capabilities. Remaining useful life prediction is central to the PHM of the lithium-ion batteries. The remaining useful life of lithium-ion batteries is defined as length of time from current time to the end of available life. An efficient method for the lithium-ion batteries monitoring would greatly improve the reliability of these machines and systems. For the lithium-ion batteries, the capacity induced by the charge-discharge operational cycle is suitable feature to represent battery degradation trend. The main challenges in battery remaining useful life prediction are to improve predicting accuracy and narrow the probability distribution function of the uncertainty. A novel data-driven approach for lithium-ion batteries remaining useful life using an improved exponential model by particle filter is proposed. To validate our proposed prognostic approach high prediction accuracy and small uncertainty, four case studies were conducted. We compared the remaining useful life prediction results associated with the original exponential model using the particle filter method. The experimental results show the following: 1) the improved exponential model needs fewer parameters than the original model; 2) the proposed prognostic method has stable and high prediction accuracy; 3) the proposed method has small uncertainty.

Prognostics of lithium-ion batteries based on flexible support vector regression

Accurate estimation of the remaining useful life of lithium-ion batteries plays an important role in the prognostic and health management (PHM). The traditional empirical data-driven approaches for RUL prediction usually need multidimensional input physical characteristics including the current, voltage, usage duration, battery temperature, and ambient temperature. From the capacity fading analysis of lithium-ion batteries, this paper found the energy efficiency and battery working temperature closely related to capacity degradation, which not only consider all performance metrics of lithium-ion batteries with regard to the RUL but also take the relationships between some performance metrics into account. Thus, we devise a non-iterative prediction model based on flexible support vector regression (F-SVR) taking the energy efficiency and battery working temperature as input physical characteristics. The F-SVR method divides the training sample dataset into several regions according to the distribution complexity and then generates different parameters set for each region, so it can accurately fit the RUL trend. The proposed prognostic method has high prediction accuracy and the proposed model needs fewer dimensions input data than the traditional empirical models from the experimental results.

PSO-based feature extraction for high dimension small sample

With the development of application areas of machine learning, we are confronted with more and more small sample datasets. The key to these applications is to solve the problem of mining useful information from these data. There are supervised and non-supervised feature extraction methods, linear or non-linear feature extraction methods. Some methods are not suitable for specific fields, so combing different extraction methods becomes a reasonable solution. We propose an algorithm to combine different extraction methods based on decision level fusion. With the difficulty of selecting parameters in feature extraction algorithms, we use PSO algorithm to find the best parameters value. The experiments on UCI datasets show the validity of our algorithms.

Distributed Particle Flow Filter for Target Tracking in Wireless Sensor Networks

We propose, in this paper, a fully distributed tracking algorithm based on particle flow filter over sensor networks based on the max-consensus. The presented distributed particle flow filter is particularly suitable for the sensor network with limited sensing range and consists of two phases: the estimation phase and consensus phase. The local estimation results are obtained via particle flow filter in the estimation phase; then the sensor nodes agree on the best estimation based on max-consensus protocol in the consensus phase. Numerical simulations and comparisons with other distributed target tracking algorithms are carried out to show the effectiveness and feasibility of our approach.

Prognostics of lithium-ion batteries based on different dimensional state equations in the particle filtering method

Accurate prediction of the remaining useful life of lithium-ion batteries plays a significant role in various devices and many researchers have focused on lithium-ion battery reliability and prognosis. A particle filter (PF) is an effective filter for estimation and prediction of time series data where model structure is available. The prediction accuracy of a PF depends on two key factors: parameter initialization and the state equation. In this paper, parameters are estimated using a PF and two empirical exponential models, i.e. the exponential model and improved exponential model, are used to track the battery capacity degradation; each model uses a different state equation. Experiments were performed to compare prediction accuracy using the related parameters estimation model with that using the capacity decline model; this paper compares the effects of the different state equations on the lithium-ion battery remaining useful life prediction. The experimental results show the merits of the capacity decline model based on particle filtering. The capacity decline model PF is more suitable for estimating the battery capacity trend in the long term.