Announcement

Collapse
No announcement yet.

Event hit dont fire when one Actor moves into another Actor.

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    #16
    Originally posted by mindfane View Post
    Problem is I think 'Hit' is fired when actor is actually blocked. I played around today with several setup, and this is what I am seeing. Couldn't you try to do a trace from one Actor to the Other within 'Begin Overlap' to get the hit normal and direction?
    I can trace, but how much performance cost trace have? I thought about constant trace ahead before, but I'm afraid that it will be big performance hit.

    Comment


      #17
      Originally posted by Battou View Post
      I can trace, but how much performance cost trace have? I thought about constant trace ahead before, but I'm afraid that it will be big performance hit.
      Thing is, you only need to do a trace when an overlap happens. So it might not be a huge hit on performance. Then again it depends on how frequent overlaps happen. Anyway just try this and see if it works. If it hits performance significantly, we can some up with something else.
      I expected challenges - but I never thought they'd be so bad, so long, and so frequent.

      Comment


        #18
        Originally posted by mindfane View Post
        Thing is, you only need to do a trace when an overlap happens. So it might not be a huge hit on performance. Then again it depends on how frequent overlaps happen. Anyway just try this and see if it works. If it hits performance significantly, we can some up with something else.
        It actually works. Thanks. (But still wandering why cant use EventHit) But now I stumbled another problem. I wanted to use hit normal to catculate angle between Actor forward vector and wall normal to calculate sliding collision, so Actor wont stop if wall isn't perpendicular but slow slide along the wall. I can calculate angle, but I don't know how to determine to witch side to slide because angle always positive. Any suggestion? Maybe there are some another way to make sliding collision?

        Comment


          #19
          Originally posted by Battou View Post
          It actually works. Thanks. (But still wandering why cant use EventHit) But now I stumbled another problem. I wanted to use hit normal to catculate angle between Actor forward vector and wall normal to calculate sliding collision, so Actor wont stop if wall isn't perpendicular but slow slide along the wall. I can calculate angle, but I don't know how to determine to witch side to slide because angle always positive. Any suggestion? Maybe there are some another way to make sliding collision?
          Convert the Forward vector and Wall normal into a Rotators and calculate Delta. Check the Yaw part of the resulting rotator. It should have postive and negetive values depending on the angle difference.
          I expected challenges - but I never thought they'd be so bad, so long, and so frequent.

          Comment


            #20
            Originally posted by mindfane View Post
            Convert the Forward vector and Wall normal into a Rotators and calculate Delta. Check the Yaw part of the resulting rotator. It should have postive and negetive values depending on the angle difference.
            It works great in 2D (mowing in XY plane), but I'm aming on 3D space movement, and then delta calculated in world axises wich in same collision but when both actors turned for example 90 deg and looking up, will give delta in wrong axis. Any way to get that delta in some actors local space axises?

            Comment


              #21
              Originally posted by Battou View Post
              It works great in 2D (mowing in XY plane), but I'm aming on 3D space movement, and then delta calculated in world axises wich in same collision but when both actors turned for example 90 deg and looking up, will give delta in wrong axis. Any way to get that delta in some actors local space axises?
              Have you tried Acosd node? It is supposed to work in this case.

              https://answers.unrealengine.com/que...ooking-at.html

              Checkout that post. Looks at the accepted answer. Pay attenstion to the right-most pasrt of the blueprint.
              I expected challenges - but I never thought they'd be so bad, so long, and so frequent.

              Comment


                #22
                Originally posted by mindfane View Post
                Have you tried Acosd node? It is supposed to work in this case.

                https://answers.unrealengine.com/que...ooking-at.html

                Checkout that post. Looks at the accepted answer. Pay attenstion to the right-most pasrt of the blueprint.
                Nope, it won't work, it also always positive, so I cant figure out do I need to turn to the right or to the left.(

                Comment


                  #23
                  From the component screen you should attach the hit event to the collision sphere -> now you have a hit event to nothing. -> Right Click Sphere1 in the above screenshot and add a hit event (On component hit) to it that way the hit event counts for your spehere.
                  Do Not Argue Toxic Game
                  Balloony PoP get it on Google Play
                  Fishy Blub Blub
                  get it on Google Play

                  Comment


                    #24
                    Originally posted by KillerSneak View Post
                    From the component screen you should attach the hit event to the collision sphere -> now you have a hit event to nothing. -> Right Click Sphere1 in the above screenshot and add a hit event (On component hit) to it that way the hit event counts for your spehere.
                    It still need or Physics simulation or Sweep on in set location. And Now it's different problem, how to get positive/negative angle detween hit point normal and actor forward vector

                    Comment

                    Working...
                    X