Interacting Multiple Model Research Articles

An Improved IMM Algorithm Based on STSRCKF for Maneuvering Target Tracking

To overcome the IMM algorithm is easy divergence and low tracking accuracy when dealing with complex maneuvering situations, this paper proposes an improved interactive multiple model strong tracking square room cubature Kalman filter (IIMM-STSRCKF) algorithm under the idea of real-time dynamic adjustment of gain matrix and transition probability matrix. The algorithm has been improved in two aspects: on the one hand, the algorithm uses the idea of a strong tracking filter to deduce a new method for time-varying fading factor and introduce it into the square root of the state error covariance matrix of the SRCKF, which improves the tracking accuracy for strong maneuver; on the other hand, the probability difference between two consecutive time points in the IMM submodel is used to adjust the Markov probability transfer matrix to adaptively improve the switching speed of the submodel and the rationality of the allocation. By comparing with IMM-CKF algorithm by maneuvering target tracking case and results show that the IIMM-STSRCKF algorithm has better tracking performance in nonmaneuvering, weak maneuvering, and strong maneuvering cases.

Deep feature tracking based on interactive multiple model

Existing convolutional neural network (CNN) based trackers have limited tracking performance in some challenging scenarios such as deformation, background clutter and illumination variation, because the features extracted from a single layer or from a linear combination of multiple layers are insufficient to describe the target appearance. To overcome this problem, we propose a novel tracking algorithm based on interactive multiple model (IMM) framework for better exploring deep features from different layers (IMM_DFT). In this method, we first build measurement models from convolutional layers by applying correlation filters on hierarchical features. Then, to effectively estimate the target state for each layer, we design a hybrid system which consists of the foregoing measurement model and an online learning motion model. Finally, in order to achieve the optimal fusion of the systems for adapting diverse appearance variation of the target and background, an IMM estimator is developed to dynamically adjust the weight of each system using likelihood function and transition probabilities. Extensive experiments on OTB-2013, OTB-2015 and VOT-2016 benchmark databases demonstrate that the proposed algorithm achieves more favorable performance than several state-of-the-art methods.

Fully adaptive update rate for non‐linear trackers

Modern phased-array multifunction radars have the ability to change or schedule the tasks of the beam in order to accomplish all their missions in an optimised fashion. The resource manager of the radar must then control the update rate of the measurement task of target active tracking, so as to minimise the radar computer load without losing targets. Scarce measurements lead to low radar load, but they also lead to an increased number of illuminations at each measurement epoch to find the target. On the basis of this rationale, a sound procedure was proposed by Blackman and van Keuk to derive an optimal measurement rate. Their optimisation criterion is established using a linear Singer target model and a linear Kalman filter. In this study, their method is extended, and the authors propose a versatile optimal update rate algorithm that is applicable to virtually any non-linear target model combined with any non-linear filter able to output an error covariance matrix. This includes extended Kalman filter (EKF), Unscented Kalman Filter (UKF), Interacting Multiple Model (IMM) algorithm, and particle filters. For numerical experiments and validation, they consider a non-linear target model based on Frenet–Serret three-dimensional equations, and the tracking is performed by a non-linear invariant EKF.

LPI time‐based TMS against high‐sensitivity ESM

Low probability of interception (LPI) time refers to a time during which a certain airborne radar system is tracking the target without being intercepted by a high-sensitivity electronic support measurement (HSESM) on the target. LPI time is determined by the tracking process of the certain airborne radar system and the interception performance of an HSESM. This study proposes a track mission shifting (TMS) tactic of the distributed airborne radar system (DARS) or distributed airborne formation radar system (DAFRS) based on LPI time. First, this study defines a short-time average interception probability of HSESM on frequency search mode. Then, with adaptive dwell time design and the maximum illumination interval algorithm based on predicted tracking error covariance matrix, this study proposes to take interactive multiple models Kalman filter to describe target tracking process and introduces binary hypothesis test of chi-square and non-central chi-square distributions as target detection criterion during target tracking. Finally, with a partial prior knowledge of the opposed HSEMS on the target, this study proposes a TMS tactic model based on LPI time. Simulations show that LPI time is a significant reference value for DARS or DAFRS to turn over track mission against an opposed HSEMS.

Research on safety modeling and analysis in information fusion system

Avionics system integration is a prominent trend in research and development of civil airplane. It can improve task effectiveness, function efficiency, and resource utilization of system. Information fusion, which includes function information fusion, processing information fusion, and sensor input fusion, is a core process of avionics system integration. Some researches about the impact of information fusion on system safety are done in avionics system which is shown as follows: (1) the concept of Mishap Dilution, Mishap Implication and Mishap Confusion (MD–MI–MC) is first defined in function information fusion of avionics system. (2) The model of multi-source MD–MI–MC is established based on hazard theory. (3) The function fusion of Automatic-Dependent Surveillance–Broadcast (ADS–B) and Traffic Collision Avoidance System (TCAS) is used as a typical example to analyze fusion system during the aircraft climbing or landing state. In this paper, the concept and model of multi-source MD–MI–MC are proposed for safety analysis of integrated avionics system. A fusion model with a variable sampling Variational Bayesian–Interacting Multiple Model (VSVB–IMM) algorithm is used to analyze. At last, a set of theory system and evaluation standards including the positive and negative earning analyses are built based on the presented MD–MI–MC theory and mechanism of integrated avionics.

Vehicle Suspension System State Estimation Combining with Interacting Multiple Model Kalman Filter

Vehicle Suspension System State Estimation Combining with Interacting Multiple Model Kalman Filter

Fault detection and diagnosis based on particle filters combined with interactive multiple-model estimation in dynamic process systems

A new approach to fault detection and diagnosis (FDD) is developed for nonlinear stochastic dynamic process systems in this paper. It is called PFs-IMM, which combines particle filters (PFs) and the interactive multiple model (IMM) estimation. In this method, a multiple-model estimation scheme is first formulated to describe the complex process system poorly represented by a single model. The IMM algorithm can deal with abrupt changes in the behavior of operating processes. The residuals of the multiple models are examined for the likelihood of each model. A decision rule is employed to adaptively determine which model is the most appropriate one at each time step. Then based on IMM, a set of PFs run in parallel is used to estimate the states and the reconciled measurements even when the operating mode changes. Each of the PFs utilizes a particular mode to derive the estimation of the state variables as well as the reconciliation of the measured variables based on the probabilistic weighting scheme. From the multiple filters, the interaction among PFs allows the fusing of dynamic estimates. To achieve higher sensitivity to faults and more robustness to disturbances and noises, a new fault index function is developed for FDD. The proposed PFs-IMM approach provides an integrated framework. It can estimate the current operational or faulty mode of the system and derive the overall state estimation and the measurement reconciliation as well. The simulation solutions to the problems are obtained to demonstrate the effectiveness of the proposed method in highly nonlinear dynamic processes.

Robust CPHD algorithm for maneuvering targets tracking via airborne pulsed Doppler radar

A robust cardinalized probability hypothesis density based on interacting multiple model (RCPHD-IMM) is proposed for multiple maneuvering targets tracking under the Doppler blind zone of airborne pulse Doppler radar. This tracking algorithm integrates the advantages of unbiased converted measurements (UCM), interacting multiple model (IMM), and cardinalized probability hypothesis density (CPHD). Under the architecture of the proposed algorithm, an IMM deals with the model switching, and the UCM accounts for nonlinearity in the dynamic system models, whereas the CPHD estimates the number and state of the targets. The gain adjustment matrix is designed to improve the robustness of the proposed algorithm in the blind zone. Furthermore, an adaptive gating strategy is proposed to remove the irrelevant measurements and guarantee the detection of new targets, thereby effectively reduce the computational complexity of the proposed algorithm. Simulation results demonstrate the validity and feasibility of the proposed algorithm. The number and state of the targets in the loss of measurements can be estimated more accurately, and the amount of calculation needed with the RCPHD-IMM algorithm is reduced by 16.03% compared with the traditional CPHD algorithm.

An IMM-GLR Approach for Marine Gas Turbine Gas Path Fault Diagnosis

An IMM-GLR approach based on interacting multiple model (IMM) and generalized likelihood ratio (GLR) estimation was developed to detect, isolate, and estimate gas turbine gas path fault (including abrupt fault and multiple faults) in the underdetermine estimation conditions. In this approach, a model set representing gas turbine health condition and different fault condition was established, and a corresponding bank of filters was designed. An IMM-based FDI algorithm based on these filters is applied to detect and isolate fault, and a GLR estimation algorithm is used to estimate the fault severity. Then a model set update strategy based on the diagnosed fault was proposed to enable the diagnosis of multiple faults. Several simulation case studies on a marine gas turbine were conducted, and the results show that the IMM-GLR approach not only accurately diagnoses the abrupt gas path fault and multiple gas path faults but also accurately estimates the severity of the detected fault in the underdetermine estimation conditions.

Data Fusion Based on Adaptive Interacting Multiple Model for GPS/INS Integrated Navigation System

The extended Kalman filter (EKF) as a primary integration scheme has been applied in the Global Positioning System (GPS) and inertial navigation system (INS) integrated system. Nevertheless, the inherent drawbacks of EKF contain not only instability caused by linearization, but also massive calculation of Jacobian matrix. To cope with this problem, the adaptive interacting multiple model (AIMM) filter method is proposed to enhance navigation performance. The soft-switching characteristic, which is provided by interacting multiple model algorithm, permits process noise to be converted between upper and lower limits, and the measurement covariance is regulated by Sage adaptive filtering on-line Moreover, since the pseudo-range and Doppler observations need to be updated, an updating policy for classified measurement is considered. Finally, the performance of the GPS/INS integration method on the basis of AIMM is evaluated by a real ship, and comparison results demonstrate that AIMM could achieve a more position accuracy.

Coverage-based cooperative target acquisition for hypersonic interceptions

This paper presents a novel coverage-based cooperative target acquisition algorithm for hypersonic interceptions. Firstly, the difficulties in the hypersonic trajectory prediction are introduced which invalidate the conventionally used predicted impact point based mid-course guidance and seeker acquisition. Secondly, in order to optimally estimate and predict the target trajectory information, the interacting multiple model (IMM) algorithm is used with the constant velocity (CV) model, the constant acceleration (CA) model and the Singer model serving as the model set. The target states are described with the probability density function (PDF) based on the IMM prediction. Thirdly, the interceptor seeker target acquisition model is established which considers the blur edge region of the field of view. The cooperative target acquisition algorithm is designed by maximizing the interceptor seekers cooperative coverage of the target high probability region (HPR). Finally, digital simulations prove the effectiveness of the proposed method and reveal that the real challenge in the hypersonic target acquisition is the poor trajectory prediction accuracy which may further result to the unsteadiness of the interceptor trajectories.

Real-Time Path Planning Based on the Situation Space of UCAVs in a Dynamic Environment

This paper aims to find a reliable, collision-free path in a dynamic environment for highly maneuverable unmanned combat air vehicles (UCAVs). Given the real-time nature of the operational scenario, quick and adaptable reactions of UCAVs are necessary for updates in situational awareness. Therefore, we propose a three dimensional (3D) path planning approach based on the situational space to provide the tactical requirements of UCAVs for tracking targets and avoiding collisions. First, to ensure reliable nonlinear measurements, the interacting multiple model (IMM) algorithm based on a cubature Kalman filter (CKF) is chosen for the tracking and prediction algorithm. A constraint reference frame combining the kinematic model of constant acceleration (CA) is developed to solve the problem of arrival point generation. Second, by analyzing the relative motion between the UCAV and the moving objects, we define the situation space and give the corresponding calculation method. In tracking the moving target, the guidance vector contains the fusion information of displacement and velocity. At the same time, taking advantage of the one-step situation space as the judgment of the threat, we further plan the collision avoidance strategy. Third, as the safety in a practically reachable trajectory of the UCAV possesses the absolute priority, the collision avoidance acceleration accounts for this dominant factor in path planning. Simulations and experimental results prove that the proposed approach can plan a smooth and flyable path in 0.008 s under the premise of soft-landing target tracking.

Adaptive state estimation for tracking of civilian aircraft

Tracking of manoeuvring subsonic aerospace vehicles has traditionally been handled by state estimators. Ordinary state estimators perform poorly as the concerned process model can only be defined imprecisely. This contribution evaluates and compares the performance of adaptive single mode non-linear estimators against several versions of other estimators. The primary comparison is with the recently introduced smooth variable structure filter (SVSF) which is claimed to inherit the robustness of variable structure approach. Both the above types of estimators are then benchmarked with a well-known version of interacting multiple model (IMM) estimator which treats the manoeuvring aircraft as a hybrid system consisting of multiple modes. Monte Carlo simulation has been used and several descriptors have been used for comparison. The comparison demonstrates that a version of non-linear adaptive estimators incorporating sigma points and a single constant turn model performs substantially better than the SVSF and its tracking performance approaches that obtainable by the IMM estimator. Novelty of this contribution lies in providing a detailed comparison of the above three families of estimators, which provides adequate insight for selecting tracking estimators by trading off estimation accuracy, algorithm complexity, tuning requirement, and computational load.

Enhancing Localization Accuracy of MEMS-INS/GPS/In-Vehicle Sensors Integration During GPS Outages

In this paper, we propose a novel localization methodology to enhance the accuracy from two aspects, i.e., adapting to the uncertain noise of microelectromechanical system-based inertial navigation system (MEMS-INS) and accurately predicting INS errors. First, an interacting multiple model (IMM)-based sequential two-stage Kalman filter is proposed to fuse the information of MEMS-INS, global positioning system (GPS), and in-vehicle sensors. Three bias filters are built with different covariance matrices to cover a wide range of noise characteristics. Then, IMM algorithm provides a soft switching among the three bias filters to adapt to the uncertain noise of MEMS-INS. Further, an elaborate predictor is developed to accurately predict INS errors during GPS outages. The elaborate predictor comprises an online trained autoregressive integrated moving average (ARIMA) model and an offline trained extreme learning machine (ELM) model. The ARIMA model is designed to predict the basic accumulation process of INS errors, while the ELM model is designed to correct the errors caused by the changes of noise characteristics. Thus, the INS errors can be properly compensated when GPS observations are not available. In all, the proposed localization methodology can achieve accurate performance when facing uncertain noises of MEMS-INS and GPS outages simultaneously. To verify the effectiveness of the proposed methodology, road test experiments with various driving scenarios were performed. The experimental results illustrate the feasibility and effectiveness of the proposed methodology.

Fault Detection and Isolation via the Interacting Multiple Model Approach Applied to Drive-By-Wire Vehicles.

The place of driving assistance systems is currently increasing drastically for road vehicles. Paving the road to the fully autonomous vehicle, the drive-by-wire technology could improve the potential of the vehicle control. The implementation of these new embedded systems is still limited, mainly for reliability reasons, thus requiring the development of diagnostic mechanisms. In this paper, we investigate the detection and the identification of sensor and actuator faults for a drive-by-wire road vehicle. An Interacting Multiple Model approach is proposed, based on a non-linear vehicle dynamics observer. The adequacy of different probabilistic observers is discussed. The results, based on experimental vehicle signals, show a fast and robust identification of sensor faults while the actuator faults are more challenging.

Vehicle Trajectory Prediction by Integrating Physics- and Maneuver-Based Approaches Using Interactive Multiple Models

Vehicle trajectory prediction helps automated vehicles and advanced driver-assistance systems have a better understanding of traffic environment and perform tasks such as criticality assessment in advance. In this study, an integrated vehicle trajectory prediction method is proposed by combining physics- and maneuver-based approaches. These two methods were combined for the reason that the physics-based trajectory prediction method could ensure accuracy in the short term with the consideration of vehicle running dynamic parameters, and the maneuver-based prediction approach has a long-term insight into future trajectories with maneuver estimation. In this study, the interactive multiple model trajectory prediction (IMMTP) method is proposed by combining the two predicting models. The probability of each model in the interactive multiple models could recursively adjust according to the predicting variance of each model. In addition, prediction uncertainty is considered by employing unscented Kalman filters in the physics-based prediction model. To the maneuver-based method, random elements for uncertainty are introduced to the trajectory of each maneuver inferred by using the dynamic Bayesian network. The approach is applied and analyzed in the lane-changing scenario by using naturalistic driving data. Comparison results indicate that IMMTP could achieve a more accurate prediction trajectory with a long prediction horizon.

Augmented Human State Estimation Using Interacting Multiple Model Particle Filters With Probabilistic Data Association

The accurate human gait tracking is an important factor for various robotic applications, such as robotic walkers aiming to provide assistance to patients with different mobility impairment, social robot companions, etc. A context-aware robot control architecture needs constant knowledge of the user's kinematic state to assess the patient's gait status and adjust its movement properly to provide optimal assistance. In this letter, we present a novel human gait tracking approach that uses two particle filters (PFs) and probabilistic data association (PDA) with an interacting multiple model (IMM) scheme for a real-time selection of the appropriate motion model according to the human gait analysis and the use of the Viterbi algorithm for an augmented human gait state estimation. The gait state estimates also interact with the IMM as a prior information that drives the Markov sampling process, while the PDA ensures that the legs of the same person are coupled. The observation data in this work are provided by a laser range finder mounted on a robotic assistant walker. A detailed experimental validation is presented using ground truth data from a motion capture system, which was used in real experiments with elder subjects who presented various mobility impairments. The validation analysis regards the algorithm's accuracy, robustness to occlusions and clutter, and the gait state classification success, subject to the effect of different number of samples used in the PFs. The results for the elder subjects show the dynamics of the proposed algorithm to be used in a real-time application due to its efficacy to provide accurate and robust augmented human gait estimates with a small number of particles.

Strong tracking SCKF based on adaptive CS model for manoeuvring aircraft tracking

A novel tracking algorithm is proposed by the integration of the adaptive current statistical (CS) model and the modified strong tracking (ST) square-root cubature Kalman filter (SCKF) for the manoeuvring aircraft tracking problem. Firstly, the acceleration recursion equation and the acceleration mean input estimation are combined in order to realise the adaptive adjustment of the CS model. Then, the introduced position of the fading factor is relocated from the orthogonality principle and a new formula is put forward. Additionally, the strong manoeuver detection function is established to adjust the manoeuvring frequency of the CS model. The simulation results show that the proposed algorithm possesses better tracking accuracy than the multiple-fading-factor SCKF based on the CS model, the SCKF-ST filter based on the modified CS model and the interacting-multiple-model (IMM)-SCKF. Moreover, the proposed algorithm decreases the runtime by 40% compared with the IMM-SCKF.

Adaptive Interacting Multiple Model Algorithm Based on Information-Weighted Consensus for Maneuvering Target Tracking.

Networked multiple sensors are used to solve the problem of maneuvering target tracking. To avoid the linearization of nonlinear dynamic functions, and to obtain more accurate estimates for maneuvering targets, a novel adaptive information-weighted consensus filter for maneuvering target tracking is proposed. The pseudo measurement matrix is computed with unscented transform to utilize the information form of measurements, which is necessary for consensus iterations. To improve the maneuvering target tracking accuracy and get a unified estimation in each sensor node across the entire network, the adaptive current statistical model is exploited to update the estimate, and the information-weighted consensus protocol is applied among neighboring nodes for each dynamic model. Based on posterior probabilities of multiple models, the final estimate of each sensor is acquired with weighted combination of model-conditioned estimates. Experimental results illustrate the superior performance of the proposed algorithm with respect tracking accuracy and agreement of estimates in the whole network.

Maneuvering Acceleration Estimation Algorithm Using Doppler Radar Measurement

An algorithm to estimate the tangential and normal accelerations directly using the Doppler radar measurement in an online closed loop form is proposed. Specific works are as follows: first, the tangential acceleration and normal acceleration are taken as the state variables to establish a linear state transition equation; secondly, the decorrelation unbiased conversion measurement Kalman filter (DUCMKF) algorithm is proposed to deal with the strongly nonlinear measurement equation; thirdly, the geometric relationship between the range rate and the velocity direction angle is used to obtain two estimators of the velocity direction angle; finally, the interactive multiple model (IMM) algorithm is used to fuse the estimators of the velocity direction angle and then the adaptive IMM of current statistical model based DUCMKF (AIMM-CS-DUCMKF) is proposed. The simulation experiment results show that the accuracy and stability of DUCMKF are better than the sequential extended Kalman filter algorithm, the sequential unscented Kalman filter algorithm, and converted measurement Kalman filter algorithms; on the other hand they show that the AIMM-CS-DUCMKF can obtain the high accuracy of the tangential and normal accelerations estimation algorithm.