CodingTrain / CodingTrain/Suggestion-Box

Ray Casting in Box2D Processing

未关闭
#652 0 条评论 2 个 reaction 已指派 0 人 在 GitHub 查看
主要语言
没有语言数据
星标
570
派生
85
PR 合并指标
30 天内没有已合并 PR

描述

I've been scouring the internet for ages looking for a good example/tutorial on how to use the rayCastCallback() function in Box2DProcessing. The only ones that I have found are for C++ or jBox2d and thus I don't know what libraries to import or proper syntax. If you could make a tutorial on ray casting in Box2dProcessing I would really appreciate it.

Just some context (if you care to know):
I was really inspired but you tutorials on Neural Networks and Genetic algorithms that I decided attempt to make a Neuroevolution program to create cars that can avoid other cars and park themselves. I have been able to implement the neural network, and the math needed to support it. Each car has a neural network that takes in inputs (from sensors) and produces output (steering and throttle). A car receives this input from its 12 sensors: 8 collision aversion sensors tell the car if it is about to collide with something, and 4 other senses tell it how far it is from its parking spot. The 8 collisions aversion (C.A.) sensors have fd.isSensor = true. I took into account that sensors collide with their own car (since they originate from it's center). The issue is that sometimes the collision listener beginContact/endContact does not trigger when it overlaps with a Boundary and only triggers in response to a Car when it collides with the center of the car. Sometimes, the sensors collisions do not register until the car collides with something. This defeats the purpose of the sensor. If you wouldn't mind revisiting collision listening in Box2D I would really appreciate you help so I can get on to the fun part-the genetic algorithm. Thank you!
![2017-09-19 2](https://user-images.githubusercontent.com/32319404/30895605-566107b4-a319-11e7-8959-96da3cfba2ff.png)

贡献指南

这个仓库没有索引到贡献指南

调研方向

从 Box2DProcessing rayCastCallback() API 以及 issue 中描述的传感器和碰撞监听器行为开始。完成的贡献应提供一个教程,涵盖 ray casting 所需的 import 和语法,并解释报告中传感器与边界和汽车发生重叠的行为。

由索引模型根据 Issue 内容生成。

评估

领域
documentation, game-dev
Issue 类型
文档
难度
4/5
预计耗时
3-5 天
活跃度
停滞
描述清晰度
基本清楚
新手友好度
35/100

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。