NCSU/URO NSF ITR Progress(March 2007)
David Kaber, Mo-Yuen Chow, Rob St. Amant &Regina Stoll
Departments of Industrial & SystemsEngineering, Electrical & Computer Engineering,
and Computer Science
North Carolina State University
“Intelligent Human-Machine Interfaces & Control for HighlyAutomated Biochemical Screening Processes”
“Top-3” Outcomes in Focus Areas: Industrial & Systems Engineering (ISE):
1. Conducted cognitive task analyses with high-throughputscreening (HTS) system users and automation engineers.
2. Redesigned and prototyped usable HTS process interfaces.3. Developed high-level cognitive models (GOMSL models) for
testing new interfaces and comparing with “old” designs.
Computer Science (CS):1. Developed new technique for translation of AI planning
language to HCI/cognitive modeling language.2. Reviewed high-level cognitive modeling (GOMS) techniques.3. Developed approach to stochastic modeling of HTS user tasks.
Electrical & Computer Engineering (ECE):1. Developed Petri Net (PN) model of HTS process.2. Developed hybrid simulation (discrete event and system
dynamics) of HTS process using MATLAB SIMULINK.3. Developed novel, efficient HTS scheduling algorithm.
Overall Research Process Model: Links among subteam activities:
ECE provides accurate simulation of HTS process at URO togenerate data for interface prototype testing. ECE develops networkresource scheduling algorithm for multiple robot/device control.
ISE (or IE) develops prototype interfaces to be linked to processsimulation. ISE creates high-level cognitive models for representingoperator behavior and prelim. interface testing.
CS translates cognitive task analysis into low-level ACT-R models ofbiologist behavior in HTS operations.
ECE and CS assist ISE with usability evaluations of old and newprocess control interfaces.
Outcomes in URO Focus Areas: Systems Engineering and Life Science Automation (LSA):
System setup and operation for HTS (automated) experiments aswell as for manual screening procedures.
Developed automated enzymatic and cellular screening methods.
Process Information Technologies (PIT): Developed web-based Operator Information Management System
(OIMS) for storing physiological data during HTS experiments. Reviewed and developed mobile wireless physiological and
workload data acquisition systems. Developed data visualization tools for real-time examination of
occupational physiology data on operators.
Occupational Physiology (OP): Investigated physiological stress response to different types of HTS
experiment workload (manual and automated screening). Evaluated methods for assessing stress tolerance under standard
operating conditions (to be used as “baseline” data for evaluatingother conditions).
Details of ISE Subteam Work: Cognitive task analyses (CTA) and automation modeling:
Used GDTA (Goal-Directed Task Analysis) to identify biologistprocess goals, tasks, decisions and information requirements.
Developed AH (abstraction hierarchy) models of processautomation describing purposes, functions and components.
Compared results to formulate interface design recommendations.
Usable interface prototyping: Used existing usability heuristics, results of CTA and new
design metaphor (“cookbook”) to prototype control dialogs forHTS devices and method editing software interfaces.
Developed Java-based prototypes and conducted usabilityevaluations with expert biologists at URO.
Cognitive model development and testing of interfaces: Developed computational GOMSL models to describe
scripted user behavior with HTS process control software. Applied models to actual interface prototypes using new
cognitive modeling tool (EGLEAN compiler and simulator).
Details of ISE Subteam Work: Images of overall GDTA hierarchy and example
AH model for HTS device:1.
Find “hits” among a huge number of compounds as fast as possible and at low cost
1.1.
Adapt “bench-top” method to High-
throughput
Screening (HTS) line
1.2.
Ensure results of assay meet quality
criteria
1.3.Optimize assay
method for efficiency
1.4.
Conduct accurate analysis of data from
assay
1.5.
Generate understandable
reports of results for
clients
1.2.1.
Identify criteria for quality factors
1.2.2.
Calculate quality statistics based on
sample data
1.2.3.Compare test
statistics with
criteria
1.2.4.
Assess quality of experimental runs
1.3.1.
Avoid bottlenecks
1.3.2.
Determine optimal
(plate) batch size
1.3.3.Establish optimal
sequence of steps
in assay
1.3.4.Optimize pipetting
steps
1.3.5.
Identify available
work-in-process storage space
1.4.1.
Establish criterion
activity level for identifying “hit”
1.4.2.Use basic statistical
analysis to identify and address outliers
in data
1.4.3.
Determine enzyme
activity levels (e.g., Trypsin) (conduct
data analysis)
1.4.4.
Determine
variation in enzyme activity levels
1.4.5.Decide whether
test compound is active
1.5.1.Address customer
information needs for decisions on
test compounds
(i.e., to support determination of
whether further investigation of
compounds is
needed)
1.1.1.
Identify steps that need to be
performed as part of (automated version of ) assay
1.1.2.
Identify appropriate micro -plate types for assay
1.1.3.Establish plate configuration to
achieve statistically valid results
1.1.4.Identify automated devices to use
to perform steps of assay
1.1.5.
Identify time critical steps as part
of assay as basis for sequencing steps
1.1.6.Identify feasible sequences of steps
in assay that allow for successful
execution of experiment (i .e ., Is there any flexibility in sequence of
steps in method ?)
1.1.7.
Adapt manual pipetting steps to automated version of assay
1.1.8.
Design measurement approach to facilitate analysis of inhibition of
enzyme activity by sample
compounds
1.1.9.Develop program for assay method
using HTS line control software
(e.g. Beckman Coulter -SAMI )
1.1.10.
Ensure reproducibility of results from HTS line Specific configuration of bar coding system components
Rotating feed motor
Feed reel for labels
Take-up reel for labelsFeed reel for foil
Take-up reel for foil
Reel position sensors
Process of system
initialization
Process of printing
bar code
Process of applying
bar code
Process of reading
bar code
Printer
Foil feeder
Vacuum gripper
Plate holder
Laser scanner
Manual entry keypad
LCD display
Process of verifying
available paper/foil Process of moving plate
holder to home position Process of paper/foil feed
Process of applying thermal ink to label
Process of removing label
from backing paperProcess of locating label position
and blowing label on micro-plateProcess of verifying label
Process of activating
reader (scanner)Process of moving
plate for scanning Process of positioning
plate holder (begin/end)
Thermal print head
Vertical translation motor
Laser diode
Rotating (scan) mirror
Receiver (photo diode)
Vacuum pump
Rotation motor
Translational motor
Rotational control motor
Translational control motor
Label paper feeder
Label and read micro-plates
Assign ID to,
and recognize,
micro-plates
RS-232 C connection
CPU
RS-232C connection to PC
Apply and read
controller
CPU
RS-232C connection to PC
RS232C to printer
GDTA forHTS Process
AH Model forBarcoder Device
Details of ISE Subteam Work: Images of old HTS process control interfaces:
Device-orientedprogramming of line
Original interfacerequires recall ofmachine language
Details of ISE Subteam Work: Images of HTS control interface prototypes:
Process-orientedprogramming of line
Redesigned interfacetaps biologist recognitionof familiar information
Details of ISE Subteam Work: Image of GOMSL model simulation in run-time
and application to actual HTS control interface:
GOMS codedebugging window
GOMS modeloutput window Java-based interface
presented to model(included system state info,features for user action).
Details of ISE Subteam Work: Usability evaluations:
5 expert biologists at CELISCA (30-40yrs.) and GOMSLmodels of human operators.
Tasks: (1) programming HTS method; (2) selecting andconfiguring micro-plate barcode labels.
Evaluated old and new interfaces in repeated tests. Dependent measures - Task completion time, errors and
usability ratings. Findings:
GOMSL model output not sign. different from actual operatorprocess behaviors and times (for 4 subtasks with oldinterface, 3-of-4 subtasks with new interface).
Expert task performance time with new interface comparableto old interface (even with limited training).
Significant reductions in errors with new interface (25% dec.). Biologist usability ratings for new interfaces (mean=4.0/5.0)
greater than for old interfaces (mean=3.25/5.0).
Major Publications in Focus Area: Industrial & Systems Engineering:
Kaber, D. B., Segall, N. & Green, R. (accepted and in revision).Metaphor-based design of high-throughput screening processinterfaces. Submitted to Int. J. of Usability Studies.
Kaber, D. B., Segall, N., Green, R. S., Entzian, K. & Junginger, S.(in press). Using multiple cognitive task analysis methods forsupervisory control interface design in high-throughput biologicalscreening processes. To appear in the Int. J. of CognitiveTechnology & Work.
Segall, N., Green, R. S. & Kaber, D. B. (2006). User, robot andautomation evaluations in high-throughput biological screeningprocesses. In Proc. of the 2006 ACM Conference on Human-robotInteraction (HRI 06) (pp. 274-281). New York: ACM.
Details of CS Subteam Work: Automated language translation tools:
Created compiler for converting action modeling formalismsin PDDL to operators in ACT-R.(PDDL useful for formalizing results of CTA. New compiler
provides pathway to transform CTA to cognitive model.) Currently developing suite of translation tools. Support and interest in cognitive modeling community.
Survey of cognitive modeling literature: First review of modeling methods research in 10 years. Focus on modern applications, off desktop.
Stochastic modeling of user tasks: Developed stochastic and time-dependent models of user
task performance using Markov chains. Developed approach for using cascading Markov models for
inferring HTS operator goal states. Currently in design and implementation stage.
Details of CS Subteam Work: Images of PDDL code and corresponding ACT-R
model translation:
(define (domain hanoi) (:requirements :strips) (:predicates (clear ?x) (on ?x ?y) (smaller ?x ?y)) (:action move :parameters (?disc ?from ?to) :precondition (and (smaller ?to ?disc) (on ?disc ?from) (clear ?disc) (clear ?to)) :effect (and (clear ?from) (on ?disc ?to) (not (on ?disc ?from)) (not (clear ?to)))))(define (problem hanoi4) (:domain hanoi) (:objects peg1 peg2 peg3 d1 d2 d3 d4) (:init (smaller peg1 d1) (smaller peg1 d2) (smaller peg1 d3) (smaller peg1 d4) (smaller peg2 d1) (smaller peg2 d2) (smaller peg2 d3) (smaller peg2 d4) (smaller peg3 d1) (smaller peg3 d2) (smaller peg3 d3) (smaller peg3 d4) (smaller d2 d1) (smaller d3 d1) (smaller d3 d2) (smaller d4 d1) (smaller d4 d2) (smaller d4 d3) (clear peg2) (clear peg3) (clear d1) (on d4 peg1) (on d3 d4) (on d2 d3) (on d1 d2)) (:goal (and (on d4 peg3) (on d3 d4) (on d2 d3) (on d1 d2))))
(p move =goal> isa move-disk test 0 from =from to =to disk =size =disk> isa disk size =size==> =disk> peg =to !pop!)
Major Publications in Focus Area: Computer Science:
Ritter, F.E., Kukreja, U. & St. Amant, R. (in press). Including a model ofvisual processing with a cognitive architecture to model a simpleteleoperation task. Journal of Cognitive Ergonomics and Decision-Making.
St. Amant, R., Horton, T.E. & Ritter, F.E. (in press). Model-based evaluationof expert cell phone menu interaction. ACM Transactions on Computer-Human Interaction.
Ritter, F.E., Van Rooy, D., St. Amant, R. & Simpson, K (2006). Providinguser models with direct access to computer interfaces: An exploratory studyof a simple human-robot interface. IEEE Transactions on Systems, Man, andCybernetics (Part A), 36(3):592–601, 2006.
St. Amant, R., McBride, S.P. & Ritter, F. E. (2006). An AI planningperspective on abstraction in ACT-R modeling: Toward an HLBR languagemanifesto. In 13th Annual ACT-R Workshop (pp. 72–76).
St. Amant, R., McBride, S.P. & Ritter, F. E. (2006). AI support for buildingcognitive models. In Proc. of the 21st National Conference on ArtificialIntelligence (AAAI) (Nectar Track, pp. 1663–1666). Menlo Park, CA: AAAIPress.
Details of ECE Subteam Work: PN modeling:
Studied HTS process as discrete event system. Investigated scheduling algorithm for HTS process. Provided high-level, event-driven structure for simulator.
Hybrid simulation: Generated data on process under faulty/normal conditions. Providing data for presentation through prototype control
interfaces for GOMSL model evaluation (ISE Team). Process scheduling algorithms:
Used to schedule HTS experiments before running process. Can be used to reschedule in real-time when faults occur.
Details of ECE Subteam Work:
CAD model andsimulation of HTSprocess line.
p32
p30
p33
t 1
p2
t 2
t3
p1
p3
p34
p4
t 4
t5
p6
t6
p5
p7
t7
p8
t8
p9
p31
ORCA takes new
test plate from the
incubator
ORCA places new
test plate at
barcoder
ORCA takes
labeled plate to
Biomek .
ORCA leaves plate
at Biomek
workspace
Biomek prepares
test plate
Biomek finishes
preparing test plate
ORCA moves test
plate to incubator
Incubator caches
the plate
p14
t14
p15
t15
p16
Incubation
Biomek prepares
test plate
ORCA moves test
plate to incubator
Incubator caches
the plate
t 11
p12
t 12
ORCA takes
labeled plate to
Biomek
ORCA leaves plate
at Biomek
workspace
p22
t22
p23
t23
p25
Biomek prepares
test plate
ORCA moves test
plate to Fluostar
ORCA releases the
plate
t19
p20
t20
ORCA takes
labeled plate to
Biomek
ORCA leaves plate
at Biomek
workspace
Fluostar reads the
plate
t26
p27
t27
p29
ORCA moves test
plate to incubator
Incubator caches
the plate
t28
t29
t9
p10
2
2
p13
t132
2
t16
p17
Incubation
2
2
p21
t21
2
2
p28
p11
2
t 10 Reserve
Biomek2
t17
p18
2
2
Release
Biomek
p19
2
t18
2
p24
2
t24
t25
p26
Release
Biomek
Reserve
Biomek
Availability of
Biomek
Availability of
ORCA
Availability of
Barcoder
ORCA Initialization
Two initial test
plates
Barcoder
Initialization
Barcoder is working
Image of PN Model of HTS process.
Details of ECE Subteam Work: Image of SIMULINK simulation of HTS process:
ORCA
ORCA
ORCA
Biomek
Biomek
Biomek
ORCA
ORCA
Biomek
ORCA
ORCA
InDelayedOut
Out
t9
InDelayedOut
Out
t8
InDelayedOut
Out
t7
InDelayedOut
Out
t6
InDelayedOut
Out
t5
InDelayedOut
Out
t4
InDelayedOut
Out
t3
InDelayedOut
Out
t29
InDelayedOut
Out
t28
InDelayedOut
Out
t27
InDelayedOut
Out
t26
InDelayedOut
Out
t25
InDelayedOut
Out
t24
InDelayedOut
Out
t23
InDelayedOut
Out
t22
InDelayedOut
Out
t21
InDelayedOut
Out
t20
InDelayedOut
Out
t2
InDelayedOut
Out
t19
InIn1D
elayedOut
Out
Out1
t18
InDelayedOut
Out
t17
InDelayedOut
Out
Out1
t16
InDelayedOut
Out
t15
InDelayedOut
Out
t14
InDelayedOut
Out
t13
InDelayedOut
Out
t12
InDelayedOut
Out
t11
In In1D
elayedOut
Out
Out1
t10
InDelayedOut
Out
t1
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p9
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p8
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p7
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p6
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p5
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p4
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p34
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p33
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p32
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p31
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p30
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p3
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p29
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p28
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p27
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p26
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p25
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p24
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p23
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p22
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p21
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p20
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p2
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p19
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p18
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p17
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p16
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p15
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p14
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p13
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p12
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p11
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p10
In
T
ra
n
O
u
tT
ra
n
T
o
k
e
n
O
u
t
p1
test9
To Workspace9
test8
To Workspace8
test7
To Workspace7
test6
To Workspace6
test5
To Workspace5
test
To Workspace4
test34
To Workspace35
test33
To Workspace34
test32
To Workspace33
test31
To Workspace32
test30
To Workspace31
test29
To Workspace30
test4
To Workspace3
test28
To Workspace29
test27
To Workspace28
test25
To Workspace27
test24
To Workspace26
test11
To Workspace25
test26
To Workspace24
test23
To Workspace23
test22
To Workspace22
test21
To Workspace21
test20
To Workspace20
test3
To Workspace2
test19
To Workspace19
test18
To Workspace18
test17
To Workspace17
test16
To Workspace16
test15
To Workspace15
test14
To Workspace14
test13
To Workspace13
test12
To Workspace12
test10
To Workspace10
test2
To Workspace1
test1
To Workspace
0
Constant3
0Constant2
0Constant1
0Constant
Details of ECE Subteam Work: Flowchart of HCH algorithm for HTS scheduling:
Major Publications in Focus Area: Electrical & Computer Engineering:
T. Hong, M.-Y. Chow, P. Haaland, D. Wilson, & R. Walker, "Scheduling a LifeScience High-Throughput Platform under Starvation Constraints Using TimedTransition Petri Nets and Heuristic Search," Proc. of IEEE ISIE07, Vigo, Spain, June4-7, 2007.
Z. Li & M.-Y. Chow, "Sampling Rate Scheduling and Digital Filter Co-design ofNetworked Supervisory Control System," Proc. of IEEE ISIE07, Vigo, June 4-7, 2007
T. Hong & M.-Y. Chow, “Timed Petri Nets Modeling of High-Throughput ScreeningProcess for Fault Study”, Proc. of IEEE IECON06, Paris, Nov. 6-10, 2006.
Z. Li & M.-Y. Chow, “Adaptive Multiple Sampling Rate Scheduling of Real-timeNetworked Supervisory Control System – Part I,” Proc. of IEEE IECON06, Paris,Nov. 6-10, 2006.
Z. Li & M.-Y. Chow, “Adaptive Multiple Sampling Rate Scheduling of Real-timeNetworked Supervisory Control System – Part II,” Proc. of IEEE IECON06, Paris,Nov. 6-10, 2006.
Z. Li, R. Vanijjirattikhan, M.-Y. Chow & Y. Viniotis, “Real-time IP Network TrafficDelay Estimation Model Design and Comparison with DSP Technology,” Proc. ofIECON05, Raleigh, NC, Nov. 6-10, 2005, pp. 2439 - 2444
W-L. (Danny) Leung, R. Vanijjirattikhan, Z. Li, L. Xu, T. Richards, B. Ayhan & M.-Y.Chow, “Intelligent Space with Time Sensitive Applications”, 2005 IEEE/ASME Int.Conference on Advanced Intelligent Mechatronics, Monterey, CA, 24-28 July, 2005.
Details of URO Work: Systems Engineering and LSA:
Developed ORCA and SAMI XXbased systems.
Screening methods applied torobotic lines at CELISCA.
Process Info Technologies (PIT): 24/7 individual operator data
recording/representation. Database storage of work-
related data. Occupational Physiology (OP):
Related degree of process autoto work efficiency and HRmeasure of workload.
Measured Cortisol in salivaunder various HTS workloadand auto conditions.
Major Publications by URO: Vilbrandt, R., Kreuzfeld, S., Stoll, R.: Flexible erfassung von belastungs-
und beanspruchungsparametern bei arbeitsmedizinischenfelduntersuchungen. Arbeitsmed.Sozialmed.Umweltmed. 41 (2006) 10,457-462.(Industrial field investigation of workload parameters on medical responses.)
Kumar, M., Weippert, M., Arndt, D., Kreuzfeld, S., Vilbrandt, R., Stoll, R.:Fuzzy evaluation of heart rate signals for mental stressassessment. IEEE Trans. on Fuzzy Systems Vol. 14 (2007) (in press).
Stoll, R., Kreuzfeld, S., Weippert, M., Vilbrandt, R., Stoll, N.: System forflexible field measurement of physiological data of operators working inautomated labs. J. Assn. Lab Auto. 11 (2007) 2, (in press).
Weippert, M., Kreuzfeld, S., Kumar, M., Kaber, D., Stoll, R.: Fuzzymodeling of mental effort using heart rate variability-data. Proc. 4th Int.Forum Life Science Automation, Hohe Düne, 14.-15.09.2006, 80.
Kumar, M., Stoll, N., Vilbrandt, R., Kaber, D., Stoll, R.: Fuzzy-baseddata interpretation – A tool for investigation of laboratory staff’s stresslevel. Proc. 4th Int. Forum Life Science Automation, 14.-15.09.2006,Hohe Düne, 69.
NCSU Students Graduated orFunded by ITR Grant:
Computer Science: Marivic Bonto-Kane (expected 2007-8; passed prelim.) Sean McBride (expected 2007-8)
Electrical & Computer Engineering: YiXin Cai (in progress) Tao Hong (in progress) Zheng Li (expected 2007; passed proposal defense) Rangsarit Vanijjirattikhan (expected 2007)
Industrial & Systems Engineering: Rebecca Green (expected 2008; passed prelim.) Sang-Hwan Kim (in progress) Noa Segall (2006; completed)
Faculty Service on InternationalStudent Dissertation Committees: Chow (NCSU ECE):
Thomas Krueger-Sundhaus, College of Computer Science andElectrical Engineering, University of Rostock (Ph.D., 2006).
Kaber (NCSU ISE): Thomas Roddelkopf, College of Computer Science and Electrical
Engineering, University of Rostock (Ph.D., 2006; completed). Ralf Schroder, College of Computer Science and Electrical
Engineering, University of Rostock (Ph.D., 2007; completed).
St. Amant (NCSU CS): Holger Dahl, College of Computer Science and Electrical
Engineering, University of Rostock (Ph.D., 2006).
Stoll (URO School of Medicine): Noa Segall, Department of Indus. & Sys. Engr., NC State
University (Ph.D., 2006; completed)
URO Students Graduated or Fundedby PEO (German ITR) Grant:
Beyond students graduated through jointcommittees, expected URO Ph.D. graduationsinclude: Mohit Kumar (expected 2007-8) for Habilitation Matthias Weippert (expected 2007) for Ph.D. Reinhard Villbrandt (expected 2008) for Ph.D. Sebastian Neubert (expected 2009) for Ph.D.
(Note: These students are supervised by R. Stoll, N. Stoll &K. Thurow through CELISCA.)