Lecture 7&8

Question 1

What are steering behaviours?

Answer 1

Make AIs move semi naturally around an environment. No pathing or anything, just reacting to local information

Question 2

What is the seek steering behaviour?

Answer 2

Steers character towards a specified position

Aligns the velocity of the character with the direction to the target

Question 3

What is the flee steering behaviour?

Answer 3

Same as seek, but away from target

Question 4

What is the arrival steering behaviour?

Answer 4

Same as seek, but decelerating as it reaches the target

Question 5

What is the wander steering behaviour?

Answer 5

Random steering

Two imaginary circles in front of agent

Big circle is wandering strength

Little circle is wandering rate

Changes destination to random point on big circle within little circle every N frames

Question 6

What is the pursuit steering behaviour?

Answer 6

Same as seek but with a moving target.

Try to predict where target will be

Question 7

What is the evade steering behaviour?

Answer 7

Same as flee, but with moving target.

Try to maximise distance from predicted target position

Question 8

What needs to be constructed to do path finding?

Answer 8

Graph of nodes (points) that represent the environment

Question 9

What are the dis/advantages of hexagonal grids over rectangular grids?

Answer 9

Hexagonal - more natural movement, but harder to implement

Question 10

HOw does dijkstra’s algorithm work?

Answer 10

Keep track of distance travelled so far

Expand the next node that will increase the cost by the least

Question 11

How does A* work?

Answer 11

Dijkstra + heuristic for cost to reach end

Question 12

What is an admissible heuristic?

Answer 12

One that never overestimates the cost

Question 13

What is the difference between Manhattan, Euclidean and Chebyshev distance?

Answer 13

Manhattan - counts diagonals as 2

Euclidean - absolute distance

Chebyshev - counts diagonals as 1

Question 14

In 4 way connectivity, manhattan distance is

A) Admissible

b) not admissible

what about 8 way?

Answer 14

4 way admissible

8 way not admissible

Question 15

The lower h(n), the

more

less

nodes A* will expand

Answer 15

More

Question 16

What is a navmesh?

Answer 16

models traversable areas of a map using triangles usually

Question 17

Pros/cons of navmeshes

Answer 17

pros: Shorter more natural paths, can ignore static and dynamic obstacles, requires less nodes, pathing isnot zigzag
cons: Not as efficient as grid/wp, procedural generation doesnt work too well, time consuming to make manually

Question 18

What are areas in unity navmeshes?

Answer 18

SOme places like snow or mud can be assigned a higher cost of traversal

Question 19

What must be added to an object to do pathfinding in unity?

Answer 19

Navmesh agent

Question 20

how can an obstacle be added in unity

Answer 20

Add a NavMesh Obstacle component to an object

Question 21

What methods does PlayerPrefs have?

Answer 21

float GetFloat(string key, float devVal=0.0f) / SetFloat(string key, float val)

int GetInt(string key, int defVal = 0) / SetInt(string key, int val)

string GetString(string key, string defVal = “”) / SetString(string key, string val)

bool HasKey(string key), void DeleteKey(string key), void DeleteAll()

void Save()

Question 22

How to make a gameobject not be destroyed when loading a new scene?

Answer 22

public static void DontDestroyONLoad(Object target)

Question 23

What is the difference between flexible and hard-coded FSMs?

Answer 23

Flexible - classes and interfaces as components

Hard-coded - all logic part of the code itself

Question 24

What is an avatar?

Answer 24

Asset created by unity when a 3d humanoid is imported. binds model to the animator