Method and apparatus for recognising a radar target

• US 20030164792A1
• Filed: 01/23/2003
• Published: 09/04/2003
• Est. Priority Date: 07/24/2000
• Status: Active Grant

• Current Assignee: Qinetiq Limited (QinetiQ Group PLC)
• Original Assignee: Qinetiq Limited (QinetiQ Group PLC)
• Inventors: Jahangir, Mohammed, Ponting, Keith M
• Granted Patent: US 6,801,155 B2
• Time in Patent Office: Days
• US Class Current: 342/90
• CPC Class Codes: G01S 13/524 based upon the phase or frequency shift resulting from movement of objects, with reference to the transmitted signals, e.g. coherent MTi
• CPC Class Codes: G01S 7/2923 based on data belonging to a number of consecutive radar periods
• CPC Class Codes: G01S 7/415 Identification of targets based on measurements of movement associated with the target
• CPC Class Codes: G06F 18/295 Markov models or related models, e.g. semi-Markov models; Markov random fields; Networks embedding Markov models
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes

Abstract

A method of recognising a radar target comprises producing a sequence of Doppler spectra of radar returns form a scene and producing therefrom a sequence of Doppler feature vectors for a target in the scene. Hidden Markov modelling (HMM) is then used to identify the sequence of Doppler feature vectors as indicating a member of a particular class of targets. HMM is used to identify the sequence of Doppler feature vectors by assigning to each feature vector an occurrence probability by selecting a probability distribution or state from a set thereof associated with a class of targets, multiplying the occurrence probabilities together to obtain an overall probability, repeating for other probability distributions in the set to determine a combination of probability distributions giving highest overall probability for that class of target, then repeating for at least one other class of targets and selecting the target class as being that which yields the highest overall occurrence probability.

Claims

1. A method of recognising a radar target comprising receiving radar returns from a scene and processing the returns to produce a Doppler spectrum characterised in that the method also includes processing the Doppler spectrum to produce a Doppler feature vector and using hidden Markov modelling (HMM) to identify the Doppler feature vector as indicating a member of a particular class of targets.

2. A method of recognising a radar target according to claim 1 characterised in that the method includes processing radar returns to obtain a sequence of Doppler spectra for each target and producing therefrom a sequence of Doppler feature vectors, and using HMM to identify the sequence of Doppler feature vectors as indicating a member of a particular class of targets.

3. A method of recognising a radar target according to claim 2 characterised in that the method includes using HMM to identify the sequence of Doppler feature vectors by assigning to each feature vector an occurrence probability by selecting a probability distribution or state from a set thereof associated with a class of targets, multiplying the occurrence probabilities together to obtain an overall probability, repeating for other probability distributions in the set to determine a combination of probability distributions giving highest overall probability for that class of targets, then repeating for at least one other class of targets and selecting the target class as being that which yields the highest overall occurrence probability.

4. A method of recognising a radar target according to claim 3 characterised in that it includes selecting probability distributions for successive feature vectors on the basis of some transitions between distributions allocated to successive feature vectors being allowed for the class of targets and others being forbidden.

5. A method of recognising a radar target according to claim 4 characterised in that it includes a preliminary HMM training procedure in which parameters for the states or probability distributions and transition probabilities are produced by deriving Doppler feature vectors for training data obtained from known classes of targets and calculating the mean and variance of vectors corresponding to like targets or target classes.

6. A method of recognising a radar target according to claim 5 characterised in that the HMM training procedure includes a plurality of cycles through state sequences.

7. A method of recognising a radar target according to claim 1 characterised in that targets encompassed within a single radar range cell.

8. Apparatus for recognising a radar target comprising a radar receiver for receiving radar returns from a scene and processing means for processing the returns to produce a Doppler spectrum characterised in that the processing means also includes means for processing the Doppler spectrum to produce a Doppler feature vector and hidden Markov modelling (HMM) means for identifying the Doppler feature vector as indicating a member of a particular class of targets.

9. Apparatus for recognising a radar target according to claim 8 characterised in that the processing means is arranged to process radar returns to obtain a sequence of Doppler spectra for each target and to produce therefrom a sequence of Doppler feature vectors, and the HMM means is arranged to identify the sequence of Doppler feature vectors as indicating a member of a particular class of targets.

10. Apparatus for recognising a radar target according to claim 9 characterised in that the HMM means is arranged to identify the sequence of Doppler feature vectors by assigning to each feature vector an occurrence probability by selecting a probability distribution or state from a set thereof associated with a class of targets, multiplying the occurrence probabilities together to obtain an overall probability, repeating for other probability distributions in the set to determine a combination of probability distributions giving highest overall probability for that class of targets, then repeating for at least one other class of targets and selecting the target class as being that which yields the highest overall occurrence probability.

11. Apparatus for recognising a radar target according to claim 10 characterised in that the HMM means is arranged to select probability distributions for successive feature vectors on the basis of some transitions between distributions allocated to successive feature vectors-being allowed for the class of targets and others being forbidden.

12. Apparatus for recognising a radar target according to claim 11 characterised in that the HMM means is arranged to undergo a preliminary training procedure in which parameters for the states or probability distributions and transition probabilities are produced by deriving Doppler feature vectors for training data obtained from known classes of targets and calculating the mean and variance of vectors corresponding to like targets or target classes.

13. Apparatus for recognising a radar target according to claim 12 characterised in that the HMM training procedure includes a plurality of cycles through state sequences.

14. Apparatus for recognising a radar target according to claim 8 characterised in that it is arranged to encompass targets within a single range cell.