AI Learns to Park – Deep Reinforcement Learning




[ad_1]

An AI learns to park a car in a parking lot in a 3D physics simulation. The simulation was implemented using Unity’s ML-Agents framework (https://unity3d.com/machine-learning). The AI consists of a deep Neural Network with 3 hidden layers of 128 neurons each. It is trained with the Proximal Policy Optimization (PPO) algorithm, which is a Reinforcement Learning approach.

Basically, the input of the Neural Network are the readings of eight depth sensors, the car’s current speed and position, as well as its relative position to the target. The outputs of the Neural Network are interpreted as engine force, braking force and turning force. These outputs can be seen at the top right corner of the zoomed out camera shots.

The AI starts off with random behaviour, i.e. the Neural Network is initialized with random weights. It then gradually learns to solve the task by reacting to environment feedback accordingly. The environment tells the AI whether it is doing good or bad with positive or negative reward signals.
In this project, the AI is rewarded with small positive signals for getting closer to the parking spot, which is outlined in red, and gets a larger reward when it actually reaches the parking spot and stops there. The final reward for reaching the parking spot is dependent on how parallel the car stops in relation to the actual parking position. If the car stops in a 90° angle to the actual parking direction for instance, the AI will only be rewarded a very small amount, relative to the amount it would get for stopping completely parallel to the actual direction.
The AI is penalized with a negative reward signal, when it either drives further away from the parking spot or if it crashes into any obstacles.

The training process shown in this video took about 23 hours on a computer with an i5 (7th or 8th gen) and a GTX 1070 with 100x simulation speed.

Subscribe for more content like this:
https://www.youtube.com/channel/UC_eerU4SleeptEbD2AA_nDw?sub_confirmation=1

Follow me on Twitter for more frequent updates on my projects:
https://twitter.com/SamuelArzt

Also check out my other videos related to this Project:

Two AI fight for the same Parking Spot:
https://www.youtube.com/watch?v=CqYKhbyHFtA

Neural Networks Explained in a Minute:
https://www.youtube.com/watch?v=rEDzUT3ymw4

Cars learn to maneuver Parcour with Genetic Algorithm:
https://www.youtube.com/watch?v=Aut32pR5PQA

Music from Bensound.com:
Timelapse Music: “The Elevator Bossa Nova”
Comedic Background: “Jazz Comedy”
Outro: “All That”

#ArtificialIntelligence #MachineLearning #ReinforcementLearning #AI #NeuralNetworks

Source


[ad_2]

Comment List

  • Samuel Arzt
    November 20, 2020

    Watch two AI Agents fight for the same parking spot:
    https://www.youtube.com/watch?v=CqYKhbyHFtA

  • Samuel Arzt
    November 20, 2020

    This is not the way to learn. This is like my Russian-made RC car in the past where it hit the wall first and then be able to change direction. In fact, those objects (light pole, sidewalk, other vehicles, etc) should be labeled in advance, so autonomous cars must learn how to avoid those obstacles in the first place. Then, time to learn how to park. Not randomly park like this.

  • Samuel Arzt
    November 20, 2020

    handicap parking permit granted

  • Samuel Arzt
    November 20, 2020

    This is exactly what my X would look like when trying to park. It's funny at the beginning but it turns into argument afterwards.

  • Samuel Arzt
    November 20, 2020

    It's like when spongebob tries to teach patrick to open a jar of pickles

  • Samuel Arzt
    November 20, 2020

    Blyada is learning to drive

  • Samuel Arzt
    November 20, 2020

    Wow…this amazing..
    I better understood this after reading this blog…
    https://martian07.github.io/intuitiveAI/car_ai.html

  • Samuel Arzt
    November 20, 2020

    i don't want to begin to imagine how long this took

  • Samuel Arzt
    November 20, 2020

    this agent is like my wife ,when she goes to supermarket

  • Samuel Arzt
    November 20, 2020

    looks like that car driven by a 5yrs old kid

  • Samuel Arzt
    November 20, 2020

    Круто 👍

  • Samuel Arzt
    November 20, 2020

    accurate representation of US citizens learning to drive

  • Samuel Arzt
    November 20, 2020

    What old people think autonomous driving looks like:

  • Samuel Arzt
    November 20, 2020

    Please more video

  • Samuel Arzt
    November 20, 2020

    Maybe I am an AI..

  • Samuel Arzt
    November 20, 2020

    It would be funny if a success would be parking in the spot in front of the red spot and just straight reverse into the red spot.

  • Samuel Arzt
    November 20, 2020

    Elon musk will like this.

  • Samuel Arzt
    November 20, 2020

    Well If It takes 20,000 Attempts to Park a Car THEN I BETTER NOT GET ONE

  • Samuel Arzt
    November 20, 2020

    Instead of measuring the distance in the form of lines, measure it in the form of concentric circles
    That's why your AI is taking more attempts to suceed

  • Samuel Arzt
    November 20, 2020

    This new tesla looks cool

  • Samuel Arzt
    November 20, 2020

    Time to turn my Stupid Car into a Smart Car.

  • Samuel Arzt
    November 20, 2020

    Sorry I'm late Susan, my car needs 20000 attempts before it can park somewhat decently

  • Samuel Arzt
    November 20, 2020

    I wanna see how it goes when its penalized for hitting the lines on the street.

  • Samuel Arzt
    November 20, 2020

    Over 300000 attempts and still driving like a drunk… I am not buying a Tesla till I can drive no more. LOL

  • Samuel Arzt
    November 20, 2020

    Yeah im not too worried to ai taking over the world yet

Write a comment