Skip to Main content Skip to Navigation

Traitement non-linéaire du signal radar par filtrage particulaire

Abstract : This memoir develops a global approach of the radar tracking problem for manoeuvering targets with low signal/noise ratio, through particle non-linear filtering. Particle filtering was proposed as soon as 1989 and yields solutions to problems where non-linearities are an obstacle to the application of traditional filtering techniques. It is a measure-theoretic approximation of the conditional probability law of state variables to be estimated, by random particles whose dynamics follows the stochastic flow of the system and are weighted, via Bayes' rule, by measurements up to current time. This work deals essentially with direct processing of radar data at the output of the digital sampling stage. It is shown in particular that taking into account target's dynamics in coherent integration of radar recurrences allows to approach theoretical limits, which are far beyond classical techniques. This leads to reconsider the detection problem itself, according to the general principle known as ''track before detect''. Besides, one deals with post-processing of target's coordinates as delivered by usual tracking radars, in particular for a generic model of missile with proportional navigation law. We solve in that case, the problem of determining its likely destination in terms of likelyhood hypothesis testing on the incoming missile's trajectory.
Document type :
Complete list of metadata
Contributor : Emilie Marchand <>
Submitted on : Monday, April 30, 2007 - 2:20:20 PM
Last modification on : Thursday, June 10, 2021 - 3:03:34 AM
Long-term archiving on: : Monday, June 27, 2011 - 3:52:13 PM


  • HAL Id : tel-00144042, version 1


Jean-Charles Noyer. Traitement non-linéaire du signal radar par filtrage particulaire. Automatique / Robotique. Université Paul Sabatier - Toulouse III, 1996. Français. ⟨tel-00144042⟩



Record views


Files downloads