A Kalman Filter Merging CV and Kinetic Acceleration Estimation Model Using Mode Probabilities

Masataka Hashirao, Tetsuya Kawase, Iwao Sasase

Research output: Contribution to journalArticle

Abstract

For radar tracking, the α-β filter and the Kalman filter, both of which do not require large computational requirements, have been widely utilized. However these filters cannot track a maneuvering target accurately. In recent years, the IMM (Interactive Multiple Model) algorithm has been proposed. The IMM is expected to reduce tracking errors for both non-maneuvering and maneuvering target. However, the IMM requires heavy computational burden, because it utilizes multiple Kaiman filters in parallel. On the other hand, the α-β filter with an acceleration term which can estimate maneuver acceleration from the past target estimated positions using the kinetic model, has been proposed. This filter is not available for tracking targets under clutter environment, since it does not calculate the covariance matrix which is needed for gate setting. In this paper, we apply the acceleration estimate to the Kaiman filter, and propose the hybrid Kalman filter with a constant-velocity filter and an acceleration estimation filter, and it integrates the outputs of two filters using the normalized distance of the prediction error of each filter. The computational requirement of the proposed filter is smaller than that of the IMM since the proposed filter consists of only two Kalman based filters. The proposed method can prevent deteriorating tracking accuracy by reducing the risk of maneuver misdetection when a target maneuvers. We evaluate the performance of the proposed filter by computer simulation, and show the effectiveness of the proposed filter, comparing with the conventional Kaiman filter and the two-stage Kaiman filter.

Original languageEnglish
Pages (from-to)3147-3151
Number of pages5
JournalIEICE Transactions on Communications
VolumeE86-B
Issue number10
Publication statusPublished - 2003 Oct

Fingerprint

Merging
Kalman filters
Kinetics
Radar tracking
Covariance matrix
Target tracking
Computer simulation

Keywords

  • Acceleration estimation
  • Kalman filter
  • Radar
  • Target tracking

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Networks and Communications

Cite this

A Kalman Filter Merging CV and Kinetic Acceleration Estimation Model Using Mode Probabilities. / Hashirao, Masataka; Kawase, Tetsuya; Sasase, Iwao.

In: IEICE Transactions on Communications, Vol. E86-B, No. 10, 10.2003, p. 3147-3151.

Research output: Contribution to journalArticle

@article{736eb9ac4b4e451cbc3f6f0ed750abaa,
title = "A Kalman Filter Merging CV and Kinetic Acceleration Estimation Model Using Mode Probabilities",
abstract = "For radar tracking, the α-β filter and the Kalman filter, both of which do not require large computational requirements, have been widely utilized. However these filters cannot track a maneuvering target accurately. In recent years, the IMM (Interactive Multiple Model) algorithm has been proposed. The IMM is expected to reduce tracking errors for both non-maneuvering and maneuvering target. However, the IMM requires heavy computational burden, because it utilizes multiple Kaiman filters in parallel. On the other hand, the α-β filter with an acceleration term which can estimate maneuver acceleration from the past target estimated positions using the kinetic model, has been proposed. This filter is not available for tracking targets under clutter environment, since it does not calculate the covariance matrix which is needed for gate setting. In this paper, we apply the acceleration estimate to the Kaiman filter, and propose the hybrid Kalman filter with a constant-velocity filter and an acceleration estimation filter, and it integrates the outputs of two filters using the normalized distance of the prediction error of each filter. The computational requirement of the proposed filter is smaller than that of the IMM since the proposed filter consists of only two Kalman based filters. The proposed method can prevent deteriorating tracking accuracy by reducing the risk of maneuver misdetection when a target maneuvers. We evaluate the performance of the proposed filter by computer simulation, and show the effectiveness of the proposed filter, comparing with the conventional Kaiman filter and the two-stage Kaiman filter.",
keywords = "Acceleration estimation, Kalman filter, Radar, Target tracking",
author = "Masataka Hashirao and Tetsuya Kawase and Iwao Sasase",
year = "2003",
month = "10",
language = "English",
volume = "E86-B",
pages = "3147--3151",
journal = "IEICE Transactions on Communications",
issn = "0916-8516",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "10",

}

TY - JOUR

T1 - A Kalman Filter Merging CV and Kinetic Acceleration Estimation Model Using Mode Probabilities

AU - Hashirao, Masataka

AU - Kawase, Tetsuya

AU - Sasase, Iwao

PY - 2003/10

Y1 - 2003/10

N2 - For radar tracking, the α-β filter and the Kalman filter, both of which do not require large computational requirements, have been widely utilized. However these filters cannot track a maneuvering target accurately. In recent years, the IMM (Interactive Multiple Model) algorithm has been proposed. The IMM is expected to reduce tracking errors for both non-maneuvering and maneuvering target. However, the IMM requires heavy computational burden, because it utilizes multiple Kaiman filters in parallel. On the other hand, the α-β filter with an acceleration term which can estimate maneuver acceleration from the past target estimated positions using the kinetic model, has been proposed. This filter is not available for tracking targets under clutter environment, since it does not calculate the covariance matrix which is needed for gate setting. In this paper, we apply the acceleration estimate to the Kaiman filter, and propose the hybrid Kalman filter with a constant-velocity filter and an acceleration estimation filter, and it integrates the outputs of two filters using the normalized distance of the prediction error of each filter. The computational requirement of the proposed filter is smaller than that of the IMM since the proposed filter consists of only two Kalman based filters. The proposed method can prevent deteriorating tracking accuracy by reducing the risk of maneuver misdetection when a target maneuvers. We evaluate the performance of the proposed filter by computer simulation, and show the effectiveness of the proposed filter, comparing with the conventional Kaiman filter and the two-stage Kaiman filter.

AB - For radar tracking, the α-β filter and the Kalman filter, both of which do not require large computational requirements, have been widely utilized. However these filters cannot track a maneuvering target accurately. In recent years, the IMM (Interactive Multiple Model) algorithm has been proposed. The IMM is expected to reduce tracking errors for both non-maneuvering and maneuvering target. However, the IMM requires heavy computational burden, because it utilizes multiple Kaiman filters in parallel. On the other hand, the α-β filter with an acceleration term which can estimate maneuver acceleration from the past target estimated positions using the kinetic model, has been proposed. This filter is not available for tracking targets under clutter environment, since it does not calculate the covariance matrix which is needed for gate setting. In this paper, we apply the acceleration estimate to the Kaiman filter, and propose the hybrid Kalman filter with a constant-velocity filter and an acceleration estimation filter, and it integrates the outputs of two filters using the normalized distance of the prediction error of each filter. The computational requirement of the proposed filter is smaller than that of the IMM since the proposed filter consists of only two Kalman based filters. The proposed method can prevent deteriorating tracking accuracy by reducing the risk of maneuver misdetection when a target maneuvers. We evaluate the performance of the proposed filter by computer simulation, and show the effectiveness of the proposed filter, comparing with the conventional Kaiman filter and the two-stage Kaiman filter.

KW - Acceleration estimation

KW - Kalman filter

KW - Radar

KW - Target tracking

UR - http://www.scopus.com/inward/record.url?scp=0242664603&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0242664603&partnerID=8YFLogxK

M3 - Article

VL - E86-B

SP - 3147

EP - 3151

JO - IEICE Transactions on Communications

JF - IEICE Transactions on Communications

SN - 0916-8516

IS - 10

ER -