Unity 机器学习代理工具包 (ML-Agents) 是一个开源项目,它使游戏和模拟能够作为训练智能代理的环境。
您最多选择25个主题 主题必须以中文或者字母或数字开头,可以包含连字符 (-),并且长度不得超过35个字符
 
 
 
 
 

366 行
12 KiB

//Put this script on your blue cube.
using System.Collections;
using UnityEngine;
using Unity.MLAgents;
using Unity.Barracuda;
using Unity.MLAgents.Actuators;
using Unity.MLAgents.Sensors;
using Unity.MLAgentsExamples;
public class WallJumpAgent : Agent
{
// Depending on this value, the wall will have different height
protected int m_Configuration;
// Brain to use when no wall is present
public NNModel noWallBrain;
// Brain to use when a jumpable wall is present
public NNModel smallWallBrain;
// Brain to use when a wall requiring a block to jump over is present
public NNModel bigWallBrain;
public GameObject ground;
public GameObject spawnArea;
protected Bounds m_SpawnAreaBounds;
public GameObject goal;
public GameObject shortBlock;
public GameObject wall;
protected Rigidbody m_ShortBlockRb;
protected Rigidbody m_AgentRb;
protected Material m_GroundMaterial;
protected Renderer m_GroundRenderer;
protected WallJumpSettings m_WallJumpSettings;
public float jumpingTime;
public float jumpTime;
// This is a downward force applied when falling to make jumps look
// less floaty
public float fallingForce;
// Use to check the coliding objects
public Collider[] hitGroundColliders = new Collider[3];
Vector3 m_JumpTargetPos;
Vector3 m_JumpStartingPos;
string m_NoWallBehaviorName = "SmallWallJump";
string m_SmallWallBehaviorName = "SmallWallJump";
string m_BigWallBehaviorName = "BigWallJump";
protected EnvironmentParameters m_ResetParams;
public override void Initialize()
{
m_WallJumpSettings = FindObjectOfType<WallJumpSettings>();
m_Configuration = Random.Range(0, 5);
m_AgentRb = GetComponent<Rigidbody>();
m_ShortBlockRb = shortBlock.GetComponent<Rigidbody>();
m_SpawnAreaBounds = spawnArea.GetComponent<Collider>().bounds;
m_GroundRenderer = ground.GetComponent<Renderer>();
m_GroundMaterial = m_GroundRenderer.material;
spawnArea.SetActive(false);
m_ResetParams = Academy.Instance.EnvironmentParameters;
// Update model references if we're overriding
var modelOverrider = GetComponent<ModelOverrider>();
if (modelOverrider.HasOverrides)
{
noWallBrain = modelOverrider.GetModelForBehaviorName(m_NoWallBehaviorName);
m_NoWallBehaviorName = ModelOverrider.GetOverrideBehaviorName(m_NoWallBehaviorName);
smallWallBrain = modelOverrider.GetModelForBehaviorName(m_SmallWallBehaviorName);
m_SmallWallBehaviorName = ModelOverrider.GetOverrideBehaviorName(m_SmallWallBehaviorName);
bigWallBrain = modelOverrider.GetModelForBehaviorName(m_BigWallBehaviorName);
m_BigWallBehaviorName = ModelOverrider.GetOverrideBehaviorName(m_BigWallBehaviorName);
}
}
// Begin the jump sequence
public void Jump()
{
jumpingTime = 0.2f;
m_JumpStartingPos = m_AgentRb.position;
}
/// <summary>
/// Does the ground check.
/// </summary>
/// <returns><c>true</c>, if the agent is on the ground,
/// <c>false</c> otherwise.</returns>
/// <param name="smallCheck"></param>
public bool DoGroundCheck(bool smallCheck)
{
if (!smallCheck)
{
hitGroundColliders = new Collider[3];
var o = gameObject;
Physics.OverlapBoxNonAlloc(
o.transform.position + new Vector3(0, -0.05f, 0),
new Vector3(0.95f / 2f, 0.5f, 0.95f / 2f),
hitGroundColliders,
o.transform.rotation);
var grounded = false;
foreach (var col in hitGroundColliders)
{
if (col != null && col.transform != transform &&
(col.CompareTag("walkableSurface") ||
col.CompareTag("block") ||
col.CompareTag("wall")))
{
grounded = true; //then we're grounded
break;
}
}
return grounded;
}
else
{
RaycastHit hit;
Physics.Raycast(transform.position + new Vector3(0, -0.05f, 0), -Vector3.up, out hit,
1f);
if (hit.collider != null &&
(hit.collider.CompareTag("walkableSurface") ||
hit.collider.CompareTag("block") ||
hit.collider.CompareTag("wall"))
&& hit.normal.y > 0.95f)
{
return true;
}
return false;
}
}
/// <summary>
/// Moves a rigidbody towards a position smoothly.
/// </summary>
/// <param name="targetPos">Target position.</param>
/// <param name="rb">The rigidbody to be moved.</param>
/// <param name="targetVel">The velocity to target during the
/// motion.</param>
/// <param name="maxVel">The maximum velocity posible.</param>
void MoveTowards(
Vector3 targetPos, Rigidbody rb, float targetVel, float maxVel)
{
var moveToPos = targetPos - rb.worldCenterOfMass;
var velocityTarget = Time.fixedDeltaTime * targetVel * moveToPos;
if (float.IsNaN(velocityTarget.x) == false)
{
rb.velocity = Vector3.MoveTowards(
rb.velocity, velocityTarget, maxVel);
}
}
public override void CollectObservations(VectorSensor sensor)
{
var agentPos = m_AgentRb.position - ground.transform.position;
sensor.AddObservation(agentPos / 20f);
sensor.AddObservation(DoGroundCheck(true) ? 1 : 0);
}
/// <summary>
/// Gets a random spawn position in the spawningArea.
/// </summary>
/// <returns>The random spawn position.</returns>
public Vector3 GetRandomSpawnPos()
{
var randomPosX = Random.Range(-m_SpawnAreaBounds.extents.x,
m_SpawnAreaBounds.extents.x);
var randomPosZ = Random.Range(-m_SpawnAreaBounds.extents.z,
m_SpawnAreaBounds.extents.z);
var randomSpawnPos = spawnArea.transform.position +
new Vector3(randomPosX, 0.45f, randomPosZ);
return randomSpawnPos;
}
/// <summary>
/// Changes the color of the ground for a moment.
/// </summary>
/// <returns>The Enumerator to be used in a Coroutine.</returns>
/// <param name="mat">The material to be swapped.</param>
/// <param name="time">The time the material will remain.</param>
protected IEnumerator GoalScoredSwapGroundMaterial(Material mat, float time)
{
m_GroundRenderer.material = mat;
yield return new WaitForSeconds(time); //wait for 2 sec
m_GroundRenderer.material = m_GroundMaterial;
}
public void MoveAgent(ActionSegment<int> act)
{
AddReward(-0.0005f);
var smallGrounded = DoGroundCheck(true);
var largeGrounded = DoGroundCheck(false);
var dirToGo = Vector3.zero;
var rotateDir = Vector3.zero;
var dirToGoForwardAction = act[0];
var rotateDirAction = act[1];
var dirToGoSideAction = act[2];
var jumpAction = act[3];
if (dirToGoForwardAction == 1)
dirToGo = (largeGrounded ? 1f : 0.5f) * 1f * transform.forward;
else if (dirToGoForwardAction == 2)
dirToGo = (largeGrounded ? 1f : 0.5f) * -1f * transform.forward;
if (rotateDirAction == 1)
rotateDir = transform.up * -1f;
else if (rotateDirAction == 2)
rotateDir = transform.up * 1f;
if (dirToGoSideAction == 1)
dirToGo = (largeGrounded ? 1f : 0.5f) * -0.6f * transform.right;
else if (dirToGoSideAction == 2)
dirToGo = (largeGrounded ? 1f : 0.5f) * 0.6f * transform.right;
if (jumpAction == 1)
if ((jumpingTime <= 0f) && smallGrounded)
{
Jump();
}
transform.Rotate(rotateDir, Time.fixedDeltaTime * 300f);
m_AgentRb.AddForce(dirToGo * m_WallJumpSettings.agentRunSpeed,
ForceMode.VelocityChange);
if (jumpingTime > 0f)
{
m_JumpTargetPos =
new Vector3(m_AgentRb.position.x,
m_JumpStartingPos.y + m_WallJumpSettings.agentJumpHeight,
m_AgentRb.position.z) + dirToGo;
MoveTowards(m_JumpTargetPos, m_AgentRb, m_WallJumpSettings.agentJumpVelocity,
m_WallJumpSettings.agentJumpVelocityMaxChange);
}
if (!(jumpingTime > 0f) && !largeGrounded)
{
m_AgentRb.AddForce(
Vector3.down * fallingForce, ForceMode.Acceleration);
}
jumpingTime -= Time.fixedDeltaTime;
}
public override void OnActionReceived(ActionBuffers actionBuffers)
{
MoveAgent(actionBuffers.DiscreteActions);
if ((!Physics.Raycast(m_AgentRb.position, Vector3.down, 20))
|| (!Physics.Raycast(m_ShortBlockRb.position, Vector3.down, 20)))
{
SetReward(-1f);
EndEpisode();
ResetBlock(m_ShortBlockRb);
StartCoroutine(
GoalScoredSwapGroundMaterial(m_WallJumpSettings.failMaterial, .5f));
}
}
public override void Heuristic(in ActionBuffers actionsOut)
{
var discreteActionsOut = actionsOut.DiscreteActions;
discreteActionsOut.Clear();
if (Input.GetKey(KeyCode.D))
{
discreteActionsOut[1] = 2;
}
if (Input.GetKey(KeyCode.W))
{
discreteActionsOut[0] = 1;
}
if (Input.GetKey(KeyCode.A))
{
discreteActionsOut[1] = 1;
}
if (Input.GetKey(KeyCode.S))
{
discreteActionsOut[0] = 2;
}
discreteActionsOut[3] = Input.GetKey(KeyCode.Space) ? 1 : 0;
}
// Detect when the agent hits the goal
protected virtual void OnTriggerStay(Collider col)
{
if (col.gameObject.CompareTag("goal") && DoGroundCheck(true))
{
SetReward(1f);
EndEpisode();
StartCoroutine(
GoalScoredSwapGroundMaterial(m_WallJumpSettings.goalScoredMaterial, 2));
}
}
//Reset the orange block position
void ResetBlock(Rigidbody blockRb)
{
blockRb.transform.position = GetRandomSpawnPos();
blockRb.velocity = Vector3.zero;
blockRb.angularVelocity = Vector3.zero;
}
public override void OnEpisodeBegin()
{
ResetBlock(m_ShortBlockRb);
transform.localPosition = new Vector3(
18 * (Random.value - 0.5f), 1, -12);
m_Configuration = Random.Range(0, 5);
m_AgentRb.velocity = default(Vector3);
}
void FixedUpdate()
{
if (m_Configuration != -1)
{
ConfigureAgent(m_Configuration);
m_Configuration = -1;
}
}
/// <summary>
/// Configures the agent. Given an integer config, the wall will have
/// different height and a different brain will be assigned to the agent.
/// </summary>
/// <param name="config">Config.
/// If 0 : No wall and noWallBrain.
/// If 1: Small wall and smallWallBrain.
/// Other : Tall wall and BigWallBrain.
/// </param>
protected virtual void ConfigureAgent(int config)
{
var localScale = wall.transform.localScale;
if (config == 0)
{
localScale = new Vector3(
localScale.x,
m_ResetParams.GetWithDefault("no_wall_height", 0),
localScale.z);
wall.transform.localScale = localScale;
SetModel(m_NoWallBehaviorName, noWallBrain);
}
else if (config == 1)
{
localScale = new Vector3(
localScale.x,
m_ResetParams.GetWithDefault("small_wall_height", 4),
localScale.z);
wall.transform.localScale = localScale;
SetModel(m_SmallWallBehaviorName, smallWallBrain);
}
else
{
var height = m_ResetParams.GetWithDefault("big_wall_height", 8);
localScale = new Vector3(
localScale.x,
height,
localScale.z);
wall.transform.localScale = localScale;
SetModel(m_BigWallBehaviorName, bigWallBrain);
}
}
}