PandaAI Path Follow: Change rate of traverse

I tried the example in the Path Follow chapter in the manual.

But I am not able to change the parameters in this way, that Ralph is reaching the set arrow position.

...
   self.AIchar = AICharacter("seeker", self.seeker, 100, 0.05, 5)
        self.AIworld.addAiChar(self.AIchar)
        self.AIbehaviors = self.AIchar.getAiBehaviors()

        #Path follow (note the order is reveresed)
        self.AIbehaviors.pathFollow(0.9)
        self.AIbehaviors.addToPath(self.target4.getPos())
        self.AIbehaviors.addToPath(self.target4.getPos())
        self.AIbehaviors.addToPath(self.target3.getPos())
        self.AIbehaviors.addToPath(self.target2.getPos())
        self.AIbehaviors.addToPath(self.target1.getPos())
        

        self.AIbehaviors.startFollow()

...

He makes a turn to the next target before he reaches the actual target.

How I can change this behavior/traverse?
I was not able to find any sufficient information about that.

Thanks.

The AI goes to the center of the model, so your models have to be centered where you want the AI to go.

Sorry for the bad english XD, I hope you still get what I’m saying.

I ran into this problem too, my actor just stopped before reaching its end point, when I really needed it to go all the way. I found this: AI Path following broken in 1.8.0?

Basically, following the path is cut short for every point. This is hard coded, and non-configurable. With the above solution, I did some tinkering to fit my own situation:

        lastPos = self.node.getPos()
        corrPoints = []
        for idx, wp in enumerate(self.path):
            dirVec = wp - lastPos
            dirVec.normalize()
            dirVec *= 4.5 if idx < len(self.path)-1 else 1.0
            corrPoints.append(wp+dirVec)
            lastPos = wp

The resulting corrPoints list what I feed to the aiBehavior:

        self.aiBehaviors.pathFollow(1.0)
        for pos in reversed(corrPoints):
            self.aiBehaviors.addToPath(pos)

Hope this helps, I had to search hard to find this solution. Documentation and examples on aiBehaviors are pretty scarce, unfortunately.