Interacting Multiple Model Method Research Articles

A Gated-Recurrent-Unit-Based Interacting Multiple Model Method for Small Bird Tracking on Lidar System.

Lidar presents a promising solution for bird surveillance in airport environments. However, the low observation refresh rate of Lidar poses challenges for tracking bird targets. To address this problem, we propose a gated recurrent unit (GRU)-based interacting multiple model (IMM) approach for tracking bird targets at low sampling frequencies. The proposed method constructs various GRU-based motion models to extract different motion patterns and to give different predictions of target trajectory in place of traditional target moving models and uses an interacting multiple model mechanism to dynamically select the most suitable GRU-based motion model for trajectory prediction and tracking. In order to fuse the GRU-based motion model and IMM, the approximation state transfer matrix method is proposed to transform the prediction of GRU-based network into an explicit state transfer model, which enables the calculation of the models' probability. The simulation carried out on an open bird trajectory dataset proves that our method outperforms classical tracking methods at low refresh rates with at least 26% improvement in tracking error. The results show that the proposed method is effective for tracking small bird targets based on Lidar systems, as well as for other low-refresh-rate tracking systems.

Improving Treatment of Noise Specification of Kalman Filtering for State Updating of Hydrological Models: Combining the Strengths of the Interacting Multiple Model Method and Cubature Kalman Filter

AbstractFiltering methods have been widely used to improve the forecast performances of hydrological models. Although extensive research has been carried out on filtering methods in the hydrology literature, researchers have not treated the noise statistics of filtering methods in much detail. Wrongly specified noise statistics can lead to degraded state estimates and even filter divergence. A powerful approach for state estimation in the presence of unknown noise statistics is the interacting multiple model method (IMM), which solves the noise specification problem by combining the strengths of multiple filters with different noise statistics. The IMM method has received little to no attention in the hydrologic context. This paper proposes a novel approach (IMM‐CKF) for updating the states of hydrological models in the presence of unknown noise statistics by combining the cubature Kalman filter (CKF) with the IMM. The CKF is a popular nonlinear filter that has received little attention in the hydrology literature. The method is tested using a lumped hydrological model. The results of the synthetic case suggest the IMM‐CKF can yield an accurate estimate for the switch as well as the changed states when there are abrupt changes in the noise statistics of true noises, even if sub‐filter noise statistics are not equal to the real ones. The real case results (a forecast experiment) show that the IMM‐CKF has successfully combined the strengths of different sub‐filters and associated noise statistics. The IMM‐CKF can be useful for hydrological forecasting, especially when useful information about the true noise statistics is unavailable.

Estimation of reactor core water level using gamma distribution measured at multiple axial points outside reactor in severe accident scenarios employing interacting multiple model

The water level inside the core provides information about the reactor’s safe operation. In severe accidents (SA), the water level information inside the reactor core may not be available as the conventional reactor devices for water level measurement are out of range. The gamma radiation collected from the detectors placed outside the reactor vessel is axially affected by the water level and fuel distribution present inside the core. This study predicts the water level in the core from the radiation distribution during events arising after the SA. An interacting multiple model (IMM) algorithm is used to determine the core water level by estimating the time-varying radiation distribution at multiple axial points. Numerical experiments using MELCOR code simulate the gamma radiation distribution in different SA scenarios. The results show that water level is estimated with reasonable accuracy using the proposed IMM method and helps diagnose and develop SA conditions measures.

An interactive multiple model method to identify the in-vessel phenomenon of a nuclear plant during a severe accident from the outer wall temperature of the reactor vessel

Nuclear power plants contain several monitoring systems that can identify the in-vessel phenomena of a severe accident (SA). Though a lot of analysis and research is carried out on SA, right from the development of the nuclear industry, not all the possible circumstances are taken into consideration. Therefore, to improve the efficacy of the safety of nuclear power plants, additional analytical studies are needed that can directly monitor severe accident phenomena. This paper presents an interacting multiple model (IMM) based fault detection and diagnosis (FDD) approach for the identification of in-vessel phenomena to provide the accident propagation information using reactor vessel (RV) out-wall temperature distribution during severe accidents in a nuclear power plant. The estimation of wall temperature is treated as a state estimation problem where the time-varying wall temperature is estimated using IMM employing three multiple models for temperature evolution. From the estimated RV out-wall temperature and rate of temperature, the in-vessel phenomena are identified such as core meltdown, corium relocation, reactor vessel damage, reflooding, etc. We tested the proposed method with five different types of SA scenarios and the results show that the proposed method has estimated the outer wall temperature with good accuracy.

A Hybrid IMM Based INS/DVL Integration Solution for Underwater Vehicles

To enhance the performance of the navigation system under the complex underwater environment, a hybrid interacting multiple model (HIMM) aided inertial navigation system (INS)/Doppler velocity logs (DVL) integration solution is proposed. First, to employ the acoustic Doppler current profiler mode of DVL, a novel INS/DVL mechanism is constructed where the water current velocity is estimated in real time and both bottom-track and water-track velocity measurements of DVL are involved in the observation vector. Meanwhile, to deal with the outliers and observation noises in the DVL's measurements, a HIMM algorithm is proposed to adaptively select the proper models to describe the changing environment. Simulations and field test are conducted to evaluate the effectiveness of the proposed algorithm, where the interacting multiple model (IMM) algorithm is employed for comparison. The results indicate that the proposed HIMM-aided INS/DVL integration solution shows superiority than the traditional IMM method when the observation noises and outliers exist. Meanwhile, the proposed integration scheme can successfully bridge the DVL's bottom-track outages.

Device-Free Human Sensing and Localization in Collaborative Human–Robot Workspaces: A Case Study

Modern robot manufacturing is fostering the implementation of hybrid production systems characterized by human-robot cooperative tasks. Safety technologies for workers protection require advanced sensing capabilities and flexible solutions that are able to monitor the movements of the operator in proximity of moving robots. This paper proposes the use of wireless device-free localization (DFL) methods and architectures to detect and track a human worker in a cooperative human-robot industrial workspace. The DFL system is composed of groups of massively interacting small, low-cost, embedded radio-frequency (RF) transceivers that perform received power measurements. These devices are anchored in fixed peripheral locations of the plant and provide the localization of the worker, who peculiarly carries neither wireless active devices (device-free) nor specific tracking sensors (sensor-free sensing). Operator motion is, in fact, estimated by tracking the perturbations of the radio field induced by the human body, considering the effect of concurrently moving robot as non-stationary interference. The proposed localization and detection algorithm is based on the jump linear Markovian system and the interactive multiple model method, and its positioning accuracy has been validated by experiments performed inside a robotic cell of an industrial test plant. The proposed DFL system has been implemented by employing IEEE 802.15.4 RF devices operating at 2.4 GHz and integrated into a software safety architecture. Finally, a software toolset has been designed to predict DFL accuracy, to verify experimental measurements, and also to support the integration with preinstalled industrial sensors to increase the accuracy of the augmented system.

Recursive estimation for Markov jump linear systems with unknown transition probabilities: A compensation approach

This paper considers the minimum mean square error estimation problem for a class of jump Markov linear systems (JMLSs) with unknown transition probabilities (TPs). Compared with other existing estimators, the underlying real but unknown TPs addressed in our method are allowed to be time-variant. The expectation and covariances of residual error in the traditional interacting multiple-model (IMM) algorithm are computed analytically to show that Kalman filter running under true mode can still perform satisfactorily in the presence of wrong TPs. A compensation operator which heuristically modifies the posterior probabilities by adjusting a compensation parameter automatically is then developed. The proposed method is recursive and reduces to IMM method when the compensation parameter goes to value one. Application results for a simulated system are presented to demonstrate the effectiveness.

A collaborative positioning algorithm for mobile target using multisensor data integration in enclosed environments

The accurate and robust positioning for mobile target is an important technology in enclosed environments. The noisy distance measurement makes the positioning performance of mobile target decline. We propose a collaborative positioning algorithm based on time differences of arrival (TDOA) and angle of arrival (AOA) measurements, combined with three-axis accelerometers (TAA). A coarse position result is estimated using an interacting multiple model (IMM) method. We further improve the positioning performance by utilising the TDOA and AOA values integrated with initial results of the IMM method. To evaluate and verify the proposed method, the Cramer–Rao lower bound (CRLB) is derived by multisensor data measurements. We implement the collaborative positioning algorithm and the CRLB estimation as well as the experiments. The results show that the proposed method can eliminate the cumulative positioning error caused by the TAA approach and reduce large positioning errors caused by the TDOA/AOA approaches. Hence, the proposed method can make up for their shortcomings of noisy measurements and perform highly accurate and robust positioning. Moreover, the proposed method is able to achieve CRLB accuracy in positioning performance analysis.

Missile Guidance Law Estimation Using Modified Interactive Multiple Model Filter

This paper presents an interactive multiple-model-based guidance law estimation filter. An antiship missile’s threat index, such as the remaining time of flight and impact angle, can be estimated by using the guidance law estimation filter. The filter bank of the filter was composed of models based on proportional navigation and impact angle control guidance laws. In the guidance law estimation filter, state variables, rather than fixed values of navigation constant or impact angle, are used for estimation. To that end, the modified interactive multiple model, which is a modified version of the conventional interactive multiple model method is presented in this paper. The modified interactive multiple model method minimizes the degradation of estimation performance by systematic approach for different state vectors. The effectiveness of the modified interactive multiple model method and the performance of the guidance law estimation filter are examined through the numerical simulation of various engagement scenarios.

지능형 운전보조시스템을 위한 IMM 기법을 이용한 전방차량 거동추정기법

In today's automotive industry, there exist several systems that help drivers reduce the possibility of accidents, such as the ADAS (Advanced Driver Assistance System). The ADAS helps drivers make correct and quick decisions during dangerous situations. This study analyzed the performance of the IMM (Interacting Multiple Model) method based on multiple Kalman filters using the data acquired from a driving simulator. An IMM algorithm is developed to identify the current discrete state of neighboring vehicles using the sensor data and the vehicle dynamics. In particular, the driving modes of the neighboring vehicles are classified by the cruising and maneuvering modes, and the transition between the states is modeled using a Markovian switching coefficient. The performance of the IMM algorithm is analyzed through realistic simulations where a target vehicle executes sudden lane change or acceleration maneuver.

Combining the interacting multiple model method with particle filters for manoeuvring target tracking with a multistatic radar system

Practical problems on target localisation and tracking arise in many military and civilian applications. This study investigates a problem on locating and tracking a manoeuvring target for a multistatic radar system. An approach that combines the interacting multiple model (IMM) method with variants of particle filters is implemented. The IMM method accounts for mode switching, whereas the particle filters account for non-linearity and/or non-Gaussianity in the dynamic system models for the posed problem. The approach consists of a constant velocity model, a constant acceleration model and a coordinated turn model. This work considers six combinations of extended Kalman filters (EKFs), unscented Kalman filters (UKFs) and particle filters for the models. A simulation study is carried out to evaluate the implemented IMM algorithm variants. Based on the test results obtained, IMM variants that use computationally efficient particle filters in the coordinated turn models, with EKFs and/or UKFs in the remaining models possess potential to achieve superior accuracy in state estimation for problems with significant non-linearity and/or non-Gaussianity.

Mobile Location Estimation Using Fuzzy-Based IMM and Data Fusion

The location of mobile station is an important issue for wireless communication systems. A location estimation scheme using fuzzy-based Interacting Multiple Model (IMM) smoother is proposed in this paper. It combines the time-of-arrival (TOA) and the received signal strength (RSS) measurements to achieve high location accuracy. The fuzzy technique is used to interpolate several linear equations to approximate the nonlinear RSS measurement. The IMM is employed as a switch between the line-of-sight (LOS) and non-line-of-sight (NLOS) states which are considered to be a Markov process with two interactive modes. By integrating the fuzzy filtering and the IMM method for range estimation between the corresponding base station (BS) and mobile station (MS), the proposed robust scheme, in association with data fusion, can efficiently mitigate the NLOS effects on the measurement range error. Simulation results are given to confirm the performance of the proposed method.

Low cost IMU–Odometer–GPS ego localization for unusual maneuvers

This paper presents the problem of outdoor vehicle localization during unusual maneuvers with the Interacting Multiple Model (IMM) and Extended Kalman Filter (EKF) approaches. IMM, contrary to classical methods, is based on the discretization of the vehicle evolution space into simple maneuvers. Each maneuver is represented by a simple dynamic model such as a constant velocity or a constant turning model. This allows the method to be optimized for highly dynamic vehicles. In this work, we focus on unusual vehicle maneuvers during some special driving situations, including very strong accelerations, high speed turnings or backwards driving with stop stages. The presented results are based on real measurements collected from different scenarios. Based on an EKF vs. IMM comparison, these results show a real interest of using the IMM method in order to take into account unusual maneuvers.

多重モデル法を用いた鉄道車両の状態診断(<小特集>TRANSLOG2009/J-RAIL2009)

This paper demonstrates the possibility to detect suspension failures of railway vehicles using a multiple-model approach from on-board measurement data. The railway vehicle model used includes the lateral and yaw motions of the wheelsets and bogie, and the lateral motion of the vehicle body, with sensors measuring the lateral acceleration and yaw rate of the bogie, and lateral acceleration of the body. The detection algorithm is formulated based on the Interacting Multiple-Model (IMM) algorithm adding a method updating estimation model. The IMM method has been applied for detecting faults in vehicle suspension systems in a simulation study. The mode probabilities and states of vehicle suspension systems are estimated based on a Kalman Filter (KF). This algorithm is evaluated in simulation examples. Simulation results indicate that the algorithm effectively detects on-board faults of railway vehicle suspension systems in realistic situation.

Fault Detection of Railway Vehicle Suspension Systems Using Multiple-Model Approach

This paper demonstrates the possibility to detect suspension failures of railway vehicles using a multiple-model approach from on-board measurement data. The railway vehicle model used includes the lateral and yaw motions of the wheelsets and bogie, and the lateral motion of the vehicle body, with sensors measuring the lateral acceleration and yaw rate of the bogie, and lateral acceleration of the body. The detection algorithm is formulated based on the Interacting Multiple-Model (IMM) algorithm. The IMM method has been applied for detecting faults in vehicle suspension systems in a simulation study. The mode probabilities and states of vehicle suspension systems are estimated based on a Kalman Filter (KF). This algorithm is evaluated in simulation examples. Simulation results indicate that the algorithm effectively detects on-board faults of railway vehicle suspension systems.

Condition Monitoring and Fault Detection of Railway Vehicle Suspension using Multiple-Model Approach

This paper demonstrates the possibility to detect suspension failures of railway vehicles using a multiple-model approach from on-board measurement data. The railway vehicle model used includes the lateral and yaw motions of the wheelsets and bogie, and the lateral motion of the vehicle body, with sensors measuring the lateral acceleration and yaw rate of the bogie, and lateral acceleration of the body. The detection algorithm is formulated based on the Interacting Multiple-Model (IMM) algorithm. The IMM method has been applied for detecting faults in vehicle suspension systems in a simulation study. The mode probabilities and states of vehicle suspension systems are estimated based on a Kalman Filter (KF). This algorithm is evaluated in simulation examples. Simulation results indicate that the algorithm effectively detects on-board faults of railway vehicle suspension systems.

IMM Method Using Kalman Filter with Fuzzy Gain

In this paper, we propose an interacting multiple model (IMM) method using intelligent tracking filter with fuzzy gain to reduce tracking errors for maneuvering targets. In the proposed filter, the unknown acceleration input for each sub-model is determined by mismatches between the modelled target dynamics and the actual target dynamics. After a acceleration input is detected, the state estimates for each sub-filter are modified. To modify the accurate estimation, we propose the fuzzy gain based on the relation between the filter residual and its variation. To optimize each fuzzy system, we utilize the genetic algorithm (GA). The tracking performance of the proposed method is compared with those of the adaptive interacting multiple model(AIMM) method and input estimation (IE) method through computer simulations.

1511 多重モデルを用いた鉄道車両の故障検知に関する研究(OS0,6,8,OS0,6,8交通・物流一般,メンテナンス・信頼性・モニタリング・環境,TRANSLOG 2006)

This paper describes the estimation algorithm of the fault detection of railway vehicles. This algorithm is formulated based on the Interacting Multiple-model (IMM) algorithm. IMM algorithm which choose probable model from number of models is applied to the fault detection. In IMM method, changes of the systems structure and the systems parameters are called mode. We provide several suspension failure modes and sensor failure modes for the fault detection. The mode probabilities and states of vehicle suspension are estimated based on Kalman Filter (KF). This algorithm is evaluated in simulation examples. Simulation results show that the algorithm is effective for on-board fault detection of the railway vehicle suspension.

1P1-B10 Interacting Multiple Model法に基づく移動ロボットの故障診断に関する研究 : センサ系の故障診断機能を有するデッドレコニングシステム(37. 車輪移動ロボット)

1P1-B10 Interacting Multiple Model法に基づく移動ロボットの故障診断に関する研究 : センサ系の故障診断機能を有するデッドレコニングシステム(37. 車輪移動ロボット)

Mobile robot localization using color signboard

Accurate, flexible and low-cost localization is an important issue for achieving autonomous and cooperative motions of mobile robots. This paper proposes a localization method for indoor mobile robots by using vision and artificial signboard. The signboards are placed in an environment and attached to the robots to decide the absolute positions and relative positions, respectively, of the robots. The important contribution is using signboard for the relative localization. Use of forward-looking camera and eye-level signboard allows us to achieve both the absolute and relative localization in a multi-robot environment, however, it brings a potential problem that the signboard viewed with the camera tends to be occasionally occluded with mobile robots hovering in the environment. The signboard is therefore designed such that the forward-looking camera can reliably detect the signboard even though it is partially blocked with unforeseen obstructions. A color-coding pattern is printed on the signboard to identify a lot of signboards in the multi-robots environment. The Kalman filter and the Interacting Multiple Model method are applied to find accurate positions of the robots by using the identified signboards. Experimental results show the feasibility of the proposed localization method.