Chapter 2. Patrolling

Patrolling is a simple extension to pathfinding. Instead of just having a single target in mind, we might have two or more points. We might go back and forth between them, or travel in a never-ending loop.

In this chapter, you will learn about:

• How patrolling works

• Patrolling in Quick Path, React, and RAIN AI packages

• Getting to know more about behavior trees

• Creating patrols that go to different points in a level by not always following the same path

Patrolling is a way to get an object from point A to point B and then to point C, and so on. Pathfinding is still required to get from one waypoint to another, but here, we daisy-chain them into a larger, more meaningful path.

Quick Path AI

Quick Path is back again, with built-in capabilities to handle patrol. With its simple approach to AI, only a few straightforward steps are needed to get a scene finished. Here is a breakdown of these steps:

• Making the world ready for patrol

• Setting up the patrol script

We'll start by expanding on our world from Chapter 1, Pathfinding the quick path demo. Stretch out a couple of the blocks to make a larger surface area. Then, click on the terrain object (the parent of all the cubes forming the terrain), and in the Inspector, click on the Bake button. You can see what happens next in the following screenshot:

If the Bake function isn't covering all the areas, you'll need to check its grid dimensions. X remains the same in the world space, but Y is actually the Z axis. You might need to increase or decrease these numbers to cover everything in the scene:

Quick Path converts the values of the Y or Z axis values internally. By default, it is set to Y as the Up/Down axis, but you can...

React AI

React AI doesn't come equipped with a patrol script, so we provided one. We'll start with this behavior tree script and look at how it works and how to use it. Here are the steps to reproduce it:

1. Create a patrol script.

2. Create a patrol AI.

3. Set up the NPC patrol.

To start with, we've provided a script for you to use. In it, I started with the last script for pathfinding, and then I extended it to use a similar configuration to the patrol path in Quick Path's patrol script. Here are a couple of key points about this script:

• It is based on the FollowThePlayer script from the previous chapter.

• You can find the code in the book's contents at \Scripts\React AI\Patrol.cs.

• It stores a public array of Vector3, so the Inspector UI can allow designers to set the waypoints.

• Instead of the target being a player, it is set to the next waypoint in the list. Once we are close enough, it selects the next waypoint. Close Enough is the float field that allows the inspector to find it.

• If there are no waypoints...

RAIN AI

RAIN has this section put together pretty well. In reality, we only have one small section to change from the pathfinding demo, especially because the pathfinding demo had actually turned off the patrol feature.

Start with the project for RAIN AI from Chapter 1, Pathfinding. From the menu, navigate to RAIN | Behavior Tree Editor. From the editor, select FollowGreenRoad. Under Sequence is a patrol route node called waypointpatrol; select it. Finally, we have a property called Loop Type. Presently, it is on One Way, which stops at the last waypoint. You can switch it to Ping Pong or Loop, as shown in the following screenshot:

Ping Pong bounces you back and forth on the path, while Loop connects the last waypoint to the first to start over.

This works when creating a typical patrolling behavior, where a character loops along a path. However, what if we want to have a character patrol an area by walking around back and forth to different points without always following the same route?...

Summary

We were able to get patrolling operational in all three AIs. Each AI had its own approach.

For patrol as well as pathfinding, Quick Path had very few steps. If your need is mostly pathfinding, it does exactly what it claims to do easily. As for React, it does not have patrol out of the box, as we saw in the last chapter. However, it was not difficult to create a script that operates inside its behavior tree AI editor. This is a powerful system that you can use to allow designers to easily access and apply your awesome scripts, but you have to be comfortable programming to build the pieces for it. For RAIN, which made this about as easy as a big red button. With one setting changed, we changed the pathfinding AI into a patrolling AI. RAIN comes equipped with a huge variety of prebuilt character controlling; we looked at how to use a different waypoint system, a waypoint network, to give variety to our character when patrolling.

This concludes our chapters on setting up basic character...