Modelling of Embedded System.pptx

8/9/2019 Modelling of Embedded System.pptx

1/34

Embedded SystemsEEE 540

Session 4

Anshul Mittal

Asst. Professor (Sr.) (SENSE)

VI !ni"ersity


2/34

Session #b$e%ti"es

• !nderstand Embedded SystemModellin& ' lo &ra*h+ SM+ Petrinets,


3/34

-uestions

• hat are the units inside a *ro%essorand hat are their fun%tionalities/

• hat is the nal sta&e of softarethat is *ut into mi%ro%ontroll /

• 1an the re2uirements %han&e in thedesi&n life %y%le/


4/34

Ansers

• !nits inside a *ro%essor 3 Pro&ramlo !nit+ 1ontrol !nit

• inal sta&e of softare is %alled #MIma&e.

• It de*ends on the %om*any *oli%y.


5/34

6 M#7E88IN9 /

• Modellin& a system before ma:in& it into a*rodu%t is mandatory.

for e& 3 If a desi&ner is *osted ith a %hallen&e ofma:in& an ele"ator + the desi&ner should rstmodel it.

• Modellin& ill hel* the desi&ner to understand thesystem better and ill in%rease the %onden%ele"el.

• Many modellin& te%hni2ues are a"ailable and it isu*to the desi&ner to ma:e a %hoi%e out of it.

VI !ni"ersity


6/34

y*i%al Embedded System

VI !ni"ersity


7/34

inite State Ma%hine ( SM )

• inite state ma%hine is a sim*le modellin&a**roa%h that is "ery useful to understandin& asystem+ hoe"er %om*le; the system may be.

• SM a model %om*osed of – the number of nite states+

– the *ossible transitions beteen the nite states +

– a%tions %aused due to transitions and it shall also shohat an a%tion ould be ta:en.

•

he SM modellin& ill start ith one of the statesa"ailable. It %an be 2uoted as SA in thedia&ram. It ill &o throu&h series of transitions andmay nally end ith any of the states a"ailable.

VI !ni"ersity


8/34


9/34

inite State Ma%hine

• here are se"eral a%tion ty*es – Entry a%tion ? hi%h is *erformed hen enterin& a state

– E;it a%tion ? hi%h is *erformed hen e;itin& a state

– In*ut a%tion ? hi%h is *erformed de*endin& on *resent stateand in*ut %onditions

– ransition a%tion ? hi%h is *erformed hen *erformin& a%ertain transition

• he 1urrent state is basi%ally determined by the *aststates. In short or to be more *re%ise it re%ords the%han&es from the startin& *oint to *resent state.

• Number of states is basi%ally &o"erned by the numberof @its in the memory+ for e;am*le+ if memory is 4bits lon& there ould be B *ossible states.

VI !ni"ersity


10/34

inite State Ma%hine

• #ther than state there are fe other terms as ellsu%h as transition and a%tion. – A transition is an indi%ation of a state %han&e that has

ha**ened as a result of meetin& a %ondition

– An a%tion is sim*ly a des%ri*tion of an a%ti"ity that hasto be *erformed at a &i"en *oint of time. A%tions %an be*erformed at entry+ e;it+ transition and in*ut.

• here are to di


11/34

inite State Ma%hine

• All states of an SM are said to be eithera%%e*tin& or not a%%e*tin&.

• At the time hen all in*ut is *ro%essed if the%urrent state is a%%e*tin& state+ the in*ut is

a%%e*ted otherise it is re$e%ted.“FSM basically works on the principle of CURRENTSTATE and INUT to deter!ine the ne"t state#“

•

i&ure indi%ates the SM *rin%i*le

VI !ni"ersity


12/34


13/34

inite State Ma%hine

• 1onsider that the ele"ator has been de*loyed in =Foor buildin&.

• Ele"ator initially shall be in idle state in any one ofthe three Foors.

• Ele"ator system does mo"in& u* and don basedon the re2uest &i"en by the user.

• Assume that ele"ator is in &round Foor and user&ot into ele"ator and has &i"en re2uest to mo"e

to Gnd Foor – rom idle state after re%ei"in& the in*ut from user the

%ontroller desi&ned for the ele"ator ill %om*are thein*ut ith the %urrent Foor number.

VI !ni"ersity


14/34

inite State Ma%hine

– If &reater it is o"ert that the ele"ator has to mo"e u*and %are should be ta:en that door should not o*enhen the lift is in motion. So disablin& the door o*enfun%tion durin& motion is mu%h a**re%iable.

– After rea%hin& the Foor the door has to be :e*t o*en forsome time (5 se%onds) so that user %an ste* out andill :ee* it o*en until a ne re2uest %omes in.

VI !ni"ersity


15/34

inite State Ma%hine

• he di


16/34

Softare A**li%ations

•

he folloin& %on%e*ts are %ommonly used tobuild softare a**li%ations ith nite statema%hines – Automata based *ro&rammin&

– E"ent dri"en SM

– Virtual SM (VSM)

• he model used for modellin& is *urely based onones %omfort and e;*ertise ith the method.

VI !ni"ersity


17/34

Perti Net 9ra*h

• @i*artite ei&hted dire%ted &ra*h – Pla%es 3 1ir%les

– ransitions 3 bars and bo;es

– Ar%s 3 arros labelled ith ei&hts

•

o:ens 3 @la%: dots

VI !ni"ersity


18/34

ransition rin& rule

• A mar:in& is %han&ed a%%ordin& to the folloin&rules – A transition is enabled if there are enou&h to:ens in

ea%h in*ut *la%e.

– An enabled transition may or may not re. – he rin& of a transition modies mar:in& by %onsumin&

to:ens from the in*ut *la%es and *rodu%in& to:ens inout*ut *la%es.


19/34


20/34

Perti Net Modellin&

•Petri net is *rimarily used for studyin& thedynami% %on%urrent beha"ior of netor: basedsystems here there is a dis%rete Fo.

• he basi% elements of *etri net modellin& are

*la%es and transitions. An ar% is the third elementbein& used hi%h e;ists only from a transition to*la%e or *la%e to a transition.

VI !ni"ersity


21/34


22/34

Syn%hroniDation

•

his *ro*erty is sim*le here transition t ill beenabled only hen atleast one to:en is there atea%h of its in*ut *la%es.

VI !ni"ersity


23/34


24/34

1on%urran%y

• his is one of the most im*ortant *ro*erties or*etri nets.

• ere in this %ase shon t and tG are %on%urrent. his ma:es *etri nets to be used to modelsystems of distributed %ontrol ith multi*le

*ro%ess e;e%utin& %on%urrently in time.

VI !ni"ersity


25/34

1onFi%t

•1onsider a situation here t and tG are bothready to re but the rin& of any leads todisablin& of the other transition.

• his is li:e a deadlo%: situation.

VI !ni"ersity


26/34

An E;am*le

•1onsider a situation here t and tG are bothready to re but the rin& of any leads todisablin& of the other transition.

• his is li:e a deadlo%: situation.

VI !ni"ersity


27/34

An E;am*le

•1ho%late Vendin& Ma%hine

VI !ni"ersity


28/34

An E;am*le

VI !ni"ersity


29/34

An E;am*le

VI !ni"ersity


30/34

An E;am*le

VI !ni"ersity


31/34

An E;am*le

VI !ni"ersity


32/34

An E;am*le

VI !ni"ersity

• inally the user ill &et the 51%ho%late.

• Ad"anta&e of *etri net Is that it &i"es"isual e


33/34

1on%lusion

•

Embedded System 1an be modelled usin& Fo&ra*hs+ SM or *etri nets. he sele%tion ofmethodolo&y is *urely based on users%on"inien%e.

•

Petri nets ha"e fe im*ortant *ro*erties su%h asse2uential e;e%ution+ syn%hroniDation+ mer&in&+1on%urran%y and %onFi%t.

VI !ni"ersity


34/34

han: 6ou

Modelling of Embedded System.pptx

Documents