- Spot - LRE

Question regarding Spot and MTL
by Edmond Irani Liu 18 Nov '20

18 Nov '20

Dear Spot team, greetings. This is Edmond Irani Liu, a Ph.D student at the Cyber-Physical Systems Group, Technical University of Munich. First of all, I would say thank you for your effort in developing such a great platform and tool! I will explain a bit about how we have used Spot in our research and then pose my questions below. We are recently working on formalization of traffic rules to further obtain traffic rule-compliant motion trajectories for automated vehicles. As a simple prototype, we have used Spot to translated some simple rules in LTL into a Büchi Automaton and later obtained its product automaton with the transition system of the discrete states of the vehicle. We could obtain some nice results with it : ) We are now trying to take one step further and to formalize more traffic rules, but came across the limitations of LTL, and switched to MTL for the reason that it is able to specify timed constraints, e.g. Always keep a safe distance to a leading vehicle if it did not perform a cut-in maneuver for the last 3 seconds. My questions are as follows: By any chance, is there some kind of workaround in Spot with which one can incorporate the mentioned timed constraints? (Last x seconds, future x seconds, ..) Of course our system state is also discrete with a 0.1 seconds step in time. If it is not possible with Spot, do you know some related model checker which can potentially handle this situation? I have been looking around but so far did not come across a tool which handles MTL/timed-automaton as nicely as Spot handles LTL. Thank you very much for your time reading this mail. I appreciate in advance for any feedback regarding the content. Kind Regards, Edmond

2 1

compare automata with exclusive_word
by Philip Schwartz 16 Nov '20

16 Nov '20

Hello Prof. Alexandre Duret-Lutz Thank you again for all your answers. I tried to compare 2 automats created using spot and got a a counterexample although they seems to be equal: aut2 = spot.translate(spot.formula("p1 U p2")) aut = spot.make_twa_graph(aut2.get_dict()) p1 = buddy.bdd_ithvar(aut.register_ap("p1")) p2 = buddy.bdd_ithvar(aut.register_ap("p2")) aut.new_states(2) aut.set_init_state(0) aut.new_edge(0, 0, p1 & buddy.bdd_not(p2)) aut.new_edge(0, 1, p2) aut.new_edge(1, 1, p1 & buddy.bdd_not(p2)) aut.new_edge(1, 1, p2) aut.set_acceptance(1, "Inf(1)") print(aut2.exclusive_word(aut)) # actual result - “p2; cycle{p1 & !p2}” # expected - None Best Regards, Phillipp

2 8

LTL formula accepts/rejects a word
by Philipp schwartz 12 Nov '20

12 Nov '20

Hello, My name is Phillip and I tried to use Spot tool to "play" with LTL formulas. My question is, is it possible to get boolean result if LTL formula accepts or rejects a word? e.g. In case the LTL formula is "aUb" is it possible to check that the formula is accepting "cycle(b)"? Best regards, Phillip

3 10

comapre spot automata
by רועי פוגלר 08 Nov '20

08 Nov '20

Hello, 1. I tried to compare 2 spot automata as described here -> https://spot-sandbox.lrde.epita.fr/notebooks/examples%20(read%20only)/conta… (last section) but, I got a counterexample which I think is not correct since both automata are not accepting this counterexample: I tried to compare automata which presents P1 U p2 and automata which answers "no" to everything and I got cycle{P1 & !p2} as a counterexample. attached picture with the code and examples from Ide. 2. Is there a way to transfer spot automata to LTL? Regards, Roi

3 8

check if word exists in language
by רועי פוגלר 22 Oct '20

22 Oct '20

Hello, When my input is a word (string such as 'p1 & !p2; cycle{p2}') and spot automata, (e.g. f = spot.formula("p1 U p2") aut = spot.translate(f, 'low') ) is it possible to check if the word exists in automata? Regards, Roi

1 1

print spot automata using graphviz
by רועי פוגלר 21 Oct '20

21 Oct '20

Hello, How can I print the automata graph? aut.show() and display(aut) does not work in simple python file (not jupyter notebook) Regards, Roi

2 1

Spot tool
by רועי פוגלר 20 Oct '20

20 Oct '20

My name is Roi and I'm an Msc student at Bar Ilan University (Israel). I'm trying to use the spot tool (python version) and I'm trying to figure out how to build spot-automata (spot.twa_graph object) type when my input is the edges, states and accepting states. can you help me please with that? In addition, is there an example of Angluin implementation using Spot? Regards, Roi

2 4

Spot tool
by רועי פוגלר 19 Oct '20

19 Oct '20

My name is Roi and I'm an Msc student at Bar Ilan University (Israel). I'm trying to use the spot tool (python version) and I'm trying to figure out how to build spot-automata (spot.twa_graph object) type when my input is the edges, states and accepting states. can you help me please with that? In addition, is there an example of Angluin implementation using Spot? Regards, Roi

1 0

Re: [Spot] Explicit order of sub-formulas in formula constructor
by Alexandre Duret-Lutz 06 Oct '20

06 Oct '20

On Tue, Oct 6, 2020 at 5:09 PM Niklas Metzger <niklas.metzger(a)cispa.saarland> wrote: > > Hi Alexandre, > > thank you for your fast response! I’m a bit confused on the meaning of creation time: > > >>> import spot > >>> f = spot.formula("Fa") > >>> ff = spot.formula("Ga") > >>> spot.formula.Or([f, ff]) > spot.formula("Fa | Ga") > >>> f = spot.formula("Ga") > >>> ff = spot.formula("Fa") > >>> spot.formula.Or([f, ff]) > spot.formula("Fa | Ga") > >>> n = spot.formula("Ga") > >>> nn = spot.formula("Fa") > >>> spot.formula.Or([n, nn]) > spot.formula("Fa | Ga") > > Is there some internal representation that does not create new objects if seen before? Yes. You can also use print(f.id()) to display the "serial number" of formula f. That's an identifier incremented for each new formula (new in the sense "not seen before"). This value is used to sort formulas (e.g., order the operands), or as hash value. But it is not necessarily unique in the case that so many formulas are created that the counter loops. -- Alexandre Duret-Lutz

1 0

Explicit order of sub-formulas in formula constructor
by Niklas Metzger 06 Oct '20

06 Oct '20

Hi! I’m currently using Spot to manipulate LTL formulas (using Python) and encountered the following behaviour: The result of "spot.formula.Or([x,y])” possibly changes the order of the subformulas, e.g. globally will appear in front of finally. Is there a way to build formulas that stick to the given order? Best and thanks in advance Niklas

2 1