Downside definition: I’ve an AI (zombie) that’s destroying a particular object to succeed in a particular location. The thing is saved as a blackboard worth and is a baby actor of a NavLink(To set off Zombie attacking object on nav hyperlink attain)
The AI additionally makes use of AI Notion to react to sounds. When the AI hears a noise (corresponding to a sound from a participant), it ought to abandon the present job (destroying the article) and transfer in direction of the noise location.
Here is the place the issue happens: when the AI is in the midst of destroying the article and hears the noise once more, the trail to the noise location may contain the NavLink used to destroy the article. Because the AI is at present destroying the article, it abandons the duty however can’t attain the noise location as a result of it must move via the NavLink comparable to the article being destroyed. This causes the AI to get caught, because it would not have a sound solution to proceed.
What I Need to Obtain:
I want the AI to proceed destroying the article if the trail to the noise location entails the NavLink associated to the article. In different phrases, the AI ought to solely abandon the destruction job if the trail to the noise location would not depend on the NavLink used to destroy the article. If the trail requires the NavLink to succeed in the noise, the AI ought to keep centered on destroying the article.
What I attempted: I attempted constructing a path between the AI’s present location and the noise location, then extracting the factors of the constructed path. I in contrast the gap between these path factors and the NavLink the AI is actively utilizing. Nonetheless, this strategy works unstably and really poorly. The AI nonetheless abandons the duty too early, and typically it would not appropriately detect the NavLink when it ought to.
I’m in search of a extra dependable solution to test if the trail to the noise location requires the AI to traverse the NavLink it’s at present utilizing to destroy the article.
