Lecture 5

LERTEKNIK
REG
Lecture Outline
AU L
T O MA RO
T IC C O N T
1. Conceptual MHT
• Fundamental Components
• Simplifications
Target Tracking: Lecture 5 • Summary
Multiple Target Tracking: Part II 2. Hypothesis-Based MHT
• Assignment Problem
• Algorithm
3. Track-Based MHT
Gustaf Hendeby • Implementation Details
hendeby@isy.liu.se • Summary
4. User Interaction
Div. Automatic Control
Dept. Electrical Engineering 5. Examples
Linköping University • SUPPORT
• ADABTS
December 10, 2014 6. Summary
• Concluding Remarks
• Learn More. . .
G. Hendeby (hendeby@isy.liu.se) Target Tracking: Lecture 5 (MHT) December 10, 2014 1 / 36 G. Hendeby (hendeby@isy.liu.se) Target Tracking: Lecture 5 (MHT) December 10, 2014 2 / 36
Last Lecture Multiple Hypothesis Tracking (MHT)
Intro multi-target tracking (MTT) MHT: consider multiple associations hypotheses over time
Single hypothesis tracker (SHT) Started with the conceptual MHT
• Global nearest neighbor (GNN)
Integrated track initialization
• Joint Probabilistic Data Association (JPDA)
Two principal implementations
Auction algorithm • hypotheses based
Fundamental theorem of target tracking • track based
Conceptual MHT Conceptual MHT Fundamental Components
Conceptual MHT: basic idea Representing Hypotheses
Each hypothesis, {Θik−1 }N h

i=1 , is characterized by the number of targets
Described in Reid (1979)
(tracks) and their corresponding sufficient statistics
Intuitive hypothesis based brute force implementation
Between consecutive time instants, different association hypotheses, Θ1k−1 , P(Θ1k−1 ) Θ2k−1 , P(Θ2k−1 )
{Θik−1 }N h
i=1 , are kept in memory
Idea: generate all possible hypotheses, and then prune to avoid
combinatorial hypotheses growth
Hypothesis limiting techniques:
• clustering 2
ŷk|k−1
1 1
ŷk|k−1 ŷk|k−1
• pruning low probability hypotheses
• N-scan pruning
• combining similar hypotheses
• ...
Conceptual MHT Fundamental Components Conceptual MHT Fundamental Components
Representing Hypotheses Generating Hypotheses
Form Θ`k , {θk , Θik−1 }

Each hypothesis, {Θik−1 }N h
i=1 , is characterized by the number of targets
(tracks) and their corresponding sufficient statistics Θ1k−1 , P(Θ1k−1 ) Θ2k−1 , P(Θ2k−1 )
yk1 yk2 yk3 yk1 yk2 yk3
Θ1k−1 , P(Θ1k−1 ) Θ2k−1 , P(Θ2k−1 ) FA
measurements
FA
measurements
FA NT FA NT
FA FA
yk3 yk3 FA NT NT FA NT NT
FA FA
T1 NT T1 NT
FA FA
FA NT FA NT
FA FA
Hypotheses NT NT NT NT NT NT
Hypotheses
FA FA
yk1 yk1 T1 NT T1 NT
2
ŷk|k−1 FA FA
1 1
ŷk|k−1 ŷk|k−1 FA NT FA NT
FA FA
yk2 yk2 T2 NT NT T1 NT NT
FA
T1 NT
FA
FA NT
FA
T1 NT NT
Conceptual MHT Fundamental Components Conceptual MHT Fundamental Components
Computing Hypothesis Probabilities System Overview
New Set of
Measurements
Let Θ`k , {θk , Θik−1 }, then {yki }m k
i=1
(using the “Fundamental Theorem of TT”)
Reduce
P(Θ`k |y0:k ) ∝ p(yk |Θ`k , y0:k−1 )P(θk |Θik−1 , y0:k−1 )P(Θik−1 |y0:k−1 ) Set of
Hypotheses
Generate New
Hypotheses
Calculate
Hyp. Probabilities Number of
{Θik−1 }N {Θik }N {P (Θik )}N Hypotheses Θik
Y h h h
mkfa mknt Y j j θk−1 (j) j j i=1 i=1 i=1
∝ βfa βnt PD pk|k−1 (yk ) (1 − PD PG ) P(Θik−1 |y0:k−1 )
j∈JDi i
j∈JND
z −1
Note User
Presentation
The sets JDi and JNDi depend on Θik−1 ! The number of targets and target Logic
estimates usually differ between hypotheses.
Conceptual MHT Simplifications Conceptual MHT Simplifications
Reducing Complexity Clustering
Group targets without common measurements, and handle the groups

separately
Θ1k−1 Θ2k−1
Clustering
yk3 yk3
Pruning of low probability hypotheses
N-scan pruning
Merging similar hypotheses
yk1 yk1
2
ŷk|k−1
1 1
yk2 yk2
Clustering Clustering: cluster management
Group targets without common measurements, and handle the groups

separately When targets get closer
Θ1k−1 Θ2k−1 • If measurement falls inside the gates of tracks in different clusters,
merge the clusters
yk3 yk3
Cluster-2 Cluster-2
• The hypotheses for each cluster are combined into a super-hypotheses
When targets separate
• If a group of tracks cluster do not share measurements with the other
yk1 yk1 tracks in the cluster (for a period of time), split the cluster
•
2
ŷk|k−1 Hypotheses for the cluster are also divided into smaller hypotheses
1 1
yk2 yk2
corresponding to two smaller clusters.
Cluster-1 Cluster-1
Clustering: process clusters separately (1/2) Clustering: process clusters separately (1/2)
Hypotheses generation Hypotheses generation
Form Θ`k , {θk , Θik−1 } for each cluster as if the other clusters do not Form Θ`k , {θk , Θik−1 } for each cluster as if the other clusters do not
exist. exist.
Without clustering Without clustering
Θ1k−1 , P(Θ1k−1 ) Θ2k−1 , P(Θ2k−1 )
measurements measurements
FA FA
FA NT
FA
Cluster-1 Cluster-2 FA NT
FA
Cluster-1 Cluster-2
FA NT NT FA NT NT
T1
FA
NT
Θ1k−1 , P(Θ1k−1 ) Θ1k−1 , P(Θ1k−1 ) T1
FA
NT
Θ2k−1 , P(Θ2k−1 ) Θ2k−1 , P(Θ2k−1 )
FA FA
FA NT measurements measurements FA NT measurements measurements
Hypotheses
Hypotheses
FA FA
NT NT NT FA FA NT NT NT FA FA
Hypotheses
Hypotheses
FA FA NT FA FA NT
Hypotheses
T1 NT T1 T1 NT T1
Hypotheses
FA NT FA NT
FA NT NT FA NT NT
FA NT T1 FA NT T1
FA FA FA FA
T2 NT NT T2 NT T1 NT NT T1 NT
FA T1
T1 NT FA
T1 NT
FA
FA NT
FA
T1 NT NT
Clustering: process clusters separately (2/2) N-scan Pruning
Hypotheses reduction Case N = 2
For each cluster: Θ12
Delete hypotheses with probability below a threshold, γp (e.g., This scheme assumes that Θ22
γp = 0.001) any uncertainty is perfectly
Θ11 Θ32
Deletion Condition: P(Θik ) < γp resolved after N time steps
Hypotheses
It is a general commonsense Θ10 Θ21 Θ42
Keep only the most probable hypotheses with a total probability mass
to choose N ≥ 5 (situation
above a threshold, γc (e.g., γc = 0.99) Θ20 Θ31 Θ52
dependent)
i
X The N last ancestors of each Θ30 Θ41 Θ62
Deletion Condition: P(Θ`kk ) > γc hypothesis must be stored
k=1 Θ51 Θ72
` time
where `k is a sequence such that P(Θ`kk ) ≥ P(Θkk+1 ) 0 1 2 3 4
N-scan Pruning N-scan Pruning
Case N = 2 Case N = 2
Θ12
This scheme assumes that Θ22

This scheme assumes that
any uncertainty is perfectly any uncertainty is perfectly
resolved after N time steps Θ11 Θ32 resolved after N time steps
Hypotheses
Hypotheses
It is a general commonsense Θ10 Θ21 Θ42 It is a general commonsense Θ42
to choose N ≥ 5 (situation to choose N ≥ 5 (situation
Θ20 Θ31 Θ52 Θ31 Θ52
dependent) dependent)
The N last ancestors of each Θ30 Θ41 Θ62 The N last ancestors of each Θ41 Θ62
hypothesis must be stored hypothesis must be stored
Θ51 Θ72
time time
0 1 2 3 4 0 1 2 3 4
N-scan Pruning N-scan Pruning
Case N = 2 Case N = 2

any uncertainty is perfectly any uncertainty is perfectly
resolved after N time steps Θ23 resolved after N time steps Θ23
Hypotheses
Hypotheses
It is a general commonsense Θ42 Θ33 It is a general commonsense Θ42 Θ33
to choose N ≥ 5 (situation to choose N ≥ 5 (situation
Θ31 Θ52 Θ43 Θ31 Θ52 Θ43
dependent) dependent)
The N last ancestors of each Θ41 Θ62 Θ53 The N last ancestors of each Θ41 Θ62 Θ53
hypothesis must be stored hypothesis must be stored
Θ63 Θ63
time time
0 1 2 3 4 0 1 2 3 4
N-scan Pruning Hypothesis Merging
Case N = 2
Reid’s original paper suggests to check for hypothesis pairs with:

This scheme assumes that the same number of targets (tracks)
any uncertainty is perfectly
similar track estimates
resolved after N time steps
Hypotheses
It is a general commonsense If these conditions are satisfied:

to choose N ≥ 5 (situation
dependent) Θ43 merge the hypotheses
The N last ancestors of each assign the new hypothesis the sum of the combined hypotheses’
Θ62 Θ53
hypothesis must be stored probability
Θ63
time
0 1 2 3 4
Conceptual MHT Summary Hypothesis-Based MHT
Summary Hypothesis-Based MHT
Attractive method since each hypothesis is Proposed by Cox and Hingorani (1996)
• an alternative representation of reality Generate only the best hypotheses, skip hypotheses that will be
• easily interpreted deleted
Drawback: generating all possible hypotheses only to discarding Use the N-best solutions to the assignment problem (introduced last
(most of) them is inefficient lecture with GNN)
Some hypotheses contain the same track; hence fewer unique tracks • Murty’s method, 1968
than hypotheses Find the Nh -best hypothesis, generating as few unnecessary
Track based methods were popular until an efficient way to implement hypothesis as possible
a hypothesis based MHT was given by Cox and Hingorani (1996) Hypothesis reduction techniques still apply
Hypothesis-Based MHT Assignment Problem Hypothesis-Based MHT Assignment Problem
Assignment Problem: repetition (1/2) Assignment Problem: repetition (2/2)
Let Θ`k , {θk , Θik−1 }.

Y P j pj θk−1 (j)
mfa mnt D k|k−1 (yk )
P(Θ`k |y0:k ) ∝ p(yk |Θ`k , y0:k−1 )P(θk |Θik−1 , y0:k−1 )P(Θik−1 |y0:k−1 ) P(Θ`k |y0:k ) ∝ βfak βntk Ci P(Θik−1 |y0:k−1 )
Y
j∈JDi
1 − PDj PGj
mkfa mknt Y j j θk−1 (j) j j
∝ βfa βnt PD pk|k−1 (yk ) (1 − PD PG ) P(Θik−1 |y0:k−1 )
j∈JDi i
j∈JND Logarithmize and form the assignment matrices
A1 T1 T2 fa1 fa2 fa3 nt1 nt2 nt3
Divide and multiply the right hand side by yk1 `11 `12 log βfa × × log βnt × ×
× represents −∞. yk2 `21 × × log βfa × × log βnt ×
nTi yk3 × × × × log βfa × × log βnt
P j pj (yki )
(1 − PDj PGj ) = (1 − PDj PGj ) (1 − PDj PGj )
Y Y Y
Ci , `ij , log D k|k−1
j j .
(1−PD PG ) A2 T1 fa1 fa2 fa3 nt1 nt2 nt3
j=1 j∈JDi i
j∈JND yk1 `11 log βfa × × log βnt × ×
yk2 `21 × log βfa × × log βnt ×
yk3 × × × log βfa × × log βnt
Hypothesis-Based MHT Assignment Problem Hypothesis-Based MHT Assignment Problem
Assignment Problem: N-best solutions Assignment Problem: Murty’s Method
Murty’s Method
Given the assignment matrix Ai ,
Given an assignment matrix Ai , the Auction algorithm (or similar)
Find the best solution using Auction algorithm.
finds the best assignment in polynomial time
2nd best solution:
Generalizations of this problem to find the N-best solutions:
• Express the 2nd best solution as the solution of a number of best
• Formulate as several best assignment problems
solution assignment problems.
• Solve independently using the Auction algorithm
• Find the solution to each of these problems by Auction.
• Murty’s method
• The solution giving the maximum reward (minimum cost) is the second
best solution.
Repeat the procedure for more solutions
Hypothesis-Based MHT Algorithm Hypothesis-Based MHT Algorithm
Algorithm Outline Generating the Nh -best Hypotheses
Aim: Given hypotheses {Θik−1 }N i mk Input {Θik−1 }N i Nh i mk

i=1 , {P(Θk−1 |y0:k−1 )}i=1 , and {yk }i=1
h
i=1 and measurements {yk }i=1 , find

h
the Nh best hypotheses {Θik }N h

i=1 (avoid generating all hypotheses)
Output HYP-LIST (N hypotheses, decreasing probability)
PROB-LIST (matching probabilities)
Reminder of Hypothesis Probability
Y P j pj θk−1 (j) 1. Initialize all elements in HYP-LIST and PROB-LIST to ∅ and −1

mfa mnt D k|k−1 (yk )
P(Θ`k |y0:k ) ∝ βfak βntk Ci P(Θik−1 |y0:k−1 ) 2. Find assignment matrices {Ai }N i Nh
i=1 for {Θk−1 }i=1
h
j∈JDi
1 − PDj PGj | {z }
Legacy 3. For j = 1 . . . Nh
| {z }
Assignment dependent 1. For i = 1 . . . Nh
1. For the assignment matrix Ai find the jth best solution Θjik
Find {Θ`k }N h
`=1 that maximizes P(Θ`k |y0:k ). 2. Compute the probability P(Θjik )
Two steps: 3. Update HYP-LIST and PROB-LIST: If the new hypothesis enters the
list, discard the least probable entry
• Obtain the solution from the assignment (Murty’s method)
4. If P(Θjik ) is lower than the lowest probability in PROB-LIST discard Θjik
• Multiply the obtained quantity by previous hypothesis dependent terms
and never use Ai again in subsequent recursions
Track-Based MHT Track-Based MHT
Track-Based MHT: motivation Track-Based MHT: principle
Hypotheses usually contain identical tracks — significantly fewer Tracks at time k, {Tki }N t
i=1
tracks than hypotheses Track scores, Sc(Tki )
Idea: Store tracks, T i , not hypotheses, Θi , over time Form a track tree, not a hypothesis tree
Delete tracks with low scores
Θ1k−1 , P(Θ1k−1 ) Θ2k−1 , P(Θ2k−1 )
old tracks new tracks
yk1 Tk1 yk3
1
Tk−1 NM Tk2
yk1 Tk3
Track List
2
Tk−1 NM Tk4
yk1
2
ŷk|k−1 yk2 Tk5
2
ŷk|k−1
1 1
ŷk|k−1 ŷk|k−1 1
Tk−1 2
1
ŷk|k−1
yk1 Tk6 Tk−1
yk2
yk2 Tk7
yk3 Tk8
Track-Based MHT Implementation Details Track-Based MHT Implementation Details
Hypotheses Generation Track Scores and Hypotheses Probabilities

Hypothesis: a collection of compatible tracks: old tracks new tracks
Θ1k = {Tk1 , Tk5 , Tk8 }, Θ2k = {Tk2 , Tk3 , Tk7 , Tk8 } Track probability:
yk1 Tk1
Generating hypothesis is needed for reducing the number of tracks
P(Θjk )
X
P(Tki ) =
further and for user presentation 1
Tk−1 NM Tk2
Tki ∈Θjk
Use only tracks with high score
yk1 Tk3
Keep track compatibility information (e.g., in a binary matrix) Hypothesis score:
Track List
old tracks new tracks 2
Tk−1 NM Tk4
Tk1 Tk2 Tk3 Tk4 Tk5 Tk6 Tk7 Tk8 yk2
Sc(Tkj )
X
yk1 Tk1
Sc(Θik ) = Tk5
Tk1 0 0 0 1 1 0 1 1 1
Tk−1 NM Tk2
Tk2 0 1 1 1 1 1 1 Tkj ∈Θik
yk1 Tk3 yk1 Tk6
Tk3 0 0 0 0 1 1
Track List
2
Tk−1 NM Tk4
Tk4 0 0 1 1 1 Hypothesis probability: yk2 Tk7
yk2 Tk5
Tk5 0 1 0 1
exp Sc(Θik )

Tk6 0 1 1
yk1 Tk6
i yk3 Tk8
yk2 Tk7 P(Θk ) =
Tk7 0 1 1+ N j
P h
yk3 Tk8 j=1 exp Sc(Θk )
Tk8 0
Track-Based MHT Implementation Details Track-Based MHT Summary
Complexity Reducing Techniques System Components

New Set of
Measurements
{yki }m k
i=1
Set of Generate New Discard Discard

Tracks Tracks Low Score Low Probability
Cluster incompatible tracks for efficient hypothesis generation i
{Tk−1 }N t
i=1 {Tki }N t
i=1 Tracks Tracks
Apply N-scan pruning to the track trees

Merge tracks with common recent measurement history Generate Calculate
Hypotheses Track
{Θik }N h
i=1 Probabilities
z −1
User
Presentation
Logic
User Interaction Examples SUPPORT
User Presentation Logic Example: harbor protection (SUPPORT)
Maximum probability hypothesis: simplest alternative

• Possibly jumpy; the maximum probability hypothesis can change
erratically
Show track clusters: (weighted) mean, covariance and expected
number of targets
Keep a separate track list: update at each step with a selection of
tracks from different hypotheses
Consult (Blackman and Popoli, 1999) for details
Examples ADABTS Summary Concluding Remarks
Example: busy indoor environments Which Multi-TT Method to Use?
SNR Low Medium High

Computation
Low Group TT / PHD GNN GNN
Medium MHT GNN or JPDA GNN
High TrBD / MHT MHT Any
GNN and JPDA are very bad in low SNR.

When using GNN, one generally has to enlarge the overconfident
covariances to account for neglected data association uncertainty.
JPDA has track coalescence and should not be used with closely
spaced targets, see the “coalescence avoiding” versions.
MHT requires significantly higher computational load but it is said to
be able to work reasonably under 10–100 times worse SNR.
Summary Learn More. . . Summary Learn More. . .
Learning More (1/2) Learning More (2/2)
Samuel S. Blackman.
Multiple hypothesis tracking for multiple target tracking. Ingemar J. Cox, Matthew L. Miller, Roy Danchick, and G. E. Newnam.
IEEE Transactions on Aerospace and Electronic Systems, 19(1):5–18, January 2004. A comparison of two algorithms for determining ranked assignments with application to
multitarget tracking and motion correspondence.
Samuel S. Blackman and Robert Popoli. IEEE Transactions on Aerospace and Electronic Systems, 33(1):295–301, January 1997.
Design and analysis of modern tracking systems.
Artech House radar library. Artech House, Inc, 1999. Roy Danchick and G. E. Newnam.
ISBN 1-5853-006-0. Reformulating Reid’s MHT method with generalised Murty K-best ranked linear
assignment algorithm.
Ingemar J. Cox and Sunita L. Hingorani. IEE Proceedings-F Radar and Sonar Navigation, 153(1):13–22, February 2006.
An efficient implementation of Reid’s multiple hypothesis tracking algorithm and its
evaluation for the purpose of visual tracking. Matthew L. Miller, Harold S. Stone, and Ingemar J. Cox.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 18(2):138–150, February Optimizing Murty’s ranked assignment method.
1996. IEEE Transactions on Aerospace and Electronic Systems, 33(3):851–862, July 1997.
Ingemar J. Cox and Matthew L. Miller. Donald B. Reid.
On finding ranked assignments with application to multitarget tracking and motion An algorithm for tracking multiple tragets.
correspondence. IEEE Transactions on Automatic Control, 24(6):843–854, December 1979.
IEEE Transactions on Aerospace and Electronic Systems, 31(1):486–489, January 1995.

Lecture 5

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture 5

Uploaded by

Copyright:

Available Formats

LERTEKNIK

Last Lecture Multiple Hypothesis Tracking (MHT)

Conceptual MHT: basic idea Representing Hypotheses

Each hypothesis, {Θik−1 }N h

Conceptual MHT Fundamental Components Conceptual MHT Fundamental Components

Representing Hypotheses Generating Hypotheses

Form Θ`k , {θk , Θik−1 }

Computing Hypothesis Probabilities System Overview

estimates usually differ between hypotheses.

Conceptual MHT Simplifications Conceptual MHT Simplifications

Reducing Complexity Clustering

Group targets without common measurements, and handle the groups

Clustering Clustering: cluster management

Group targets without common measurements, and handle the groups

Conceptual MHT Simplifications Conceptual MHT Simplifications

Clustering: process clusters separately (2/2) N-scan Pruning

Hypotheses reduction Case N = 2

For each cluster: Θ12

Conceptual MHT Simplifications Conceptual MHT Simplifications

N-scan Pruning N-scan Pruning

This scheme assumes that Θ22

N-scan Pruning N-scan Pruning

This scheme assumes that Θ13

Conceptual MHT Simplifications Conceptual MHT Simplifications

N-scan Pruning Hypothesis Merging

Reid’s original paper suggests to check for hypothesis pairs with:

It is a general commonsense If these conditions are satisfied:

Summary Hypothesis-Based MHT

Hypothesis-Based MHT Assignment Problem Hypothesis-Based MHT Assignment Problem

Assignment Problem: repetition (1/2) Assignment Problem: repetition (2/2)

Let Θ`k , {θk , Θik−1 }.

Assignment Problem: N-best solutions Assignment Problem: Murty’s Method

Hypothesis-Based MHT Algorithm Hypothesis-Based MHT Algorithm

Algorithm Outline Generating the Nh -best Hypotheses

Aim: Given hypotheses {Θik−1 }N i mk Input {Θik−1 }N i Nh i mk

i=1 and measurements {yk }i=1 , find

the Nh best hypotheses {Θik }N h

 Y P j pj θk−1 (j)  1. Initialize all elements in HYP-LIST and PROB-LIST to ∅ and −1

Track-Based MHT: motivation Track-Based MHT: principle

Track-Based MHT Implementation Details Track-Based MHT Implementation Details

Hypotheses Generation Track Scores and Hypotheses Probabilities

Complexity Reducing Techniques System Components

Set of Generate New Discard Discard

Apply N-scan pruning to the track trees

User Interaction Examples SUPPORT

User Presentation Logic Example: harbor protection (SUPPORT)

Maximum probability hypothesis: simplest alternative

Example: busy indoor environments Which Multi-TT Method to Use?

SNR Low Medium High

GNN and JPDA are very bad in low SNR.

Summary Learn More. . . Summary Learn More. . .

Learning More (1/2) Learning More (2/2)

You might also like

Y P j pj θk−1 (j) 1. Initialize all elements in HYP-LIST and PROB-LIST to ∅ and −1