机器人如何实时确定他们的状态，并从带有噪声的传感器测量量获得周围环境的信息？在这个模块中，你将学习怎样让机器人把不确定性融入估计，并向动态和变化的世界进行学习。特殊专题包括用于定位和绘图的概率生成模型和贝叶斯滤波器。

Loading...

來自 University of Pennsylvania 的課程

机器人学：估计和学习

275 個評分

机器人如何实时确定他们的状态，并从带有噪声的传感器测量量获得周围环境的信息？在这个模块中，你将学习怎样让机器人把不确定性融入估计，并向动态和变化的世界进行学习。特殊专题包括用于定位和绘图的概率生成模型和贝叶斯滤波器。

從本節課中

Gaussian Model Learning

We will learn about the Gaussian distribution for parametric modeling in robotics. The Gaussian distribution is the most widely used continuous distribution and provides a useful way to estimate uncertainty and predict in the world. We will start by discussing the one-dimensional Gaussian distribution, and then move on to the multivariate Gaussian distribution. Finally, we will extend the concept to models that use Mixtures of Gaussians.

- Daniel LeeProfessor of Electrical and Systems Engineering

School of Engineering and Applied Science

[MUSIC]

Welcome to the Robotics Estimation and Learning Course.

This is one of the modules in this robotics specialization series.

This course will teach you how robots can estimate properties of the world

from observations over time, and learn from their prior experience.

Over the next four weeks, you will learn various methods to deal with noise and

uncertainty in real robots.

And how to implement probabilistic algorithms to account for

this uncertainty.

Let's get started.

Why do robots need to estimate and learn?

Consider the following robot soccer example.

Here our humanoid robot soccer team is playing in a match with another team

at the RoboCup competition.

The robots are complete autonomous and their onboard computers need to integrate

information from the inertial and vision sensors to perceive the world around them.

Plan their behaviors to either attack or defend, and send motor commands for

locomotion and to manipulate the orange soccer ball in various ways.

In this scenario, the attacking robot needs to estimate where the ball and

the goal are located in order to line up a kick.

Then as the ball approaches the goal,

the goalie robot needs to estimate the speed and direction of the ball.

In order to execute an appropriate dive to save the ball from going into the goal.

In order to accomplish this accurately and efficiently, the robots need to learn

the appropriate parameters during the many hours of practice before the match.

This course will teach you the underlying mathematical framework and

the computational algorithms that the robots are using to do these tests.

What do we mean by estimation and learning?

By estimation, we mean to estimate some aspect of the state of the world

from noisy, incomplete and uncertain data.

And by learning, we mean to use, have the robots use prior experience to improve

their performance under this uncertainty.

What are the sources of uncertainty in robotics?

One, there are a sensor noise.

That is that the sensors can provide inaccurate information.

Two, there could be a lack of knowledge about the world.

That is, things can be hidden from view or that the robots could not perceive.

And three, there might be dynamic changes in the motion and in the environment.

That is, that things maybe moving over time and

the robots do not know exactly where things are at the current instant.

So there are several ways to deal with this uncertainty.

In this course,

we'll focus on two different aspects of dealing with this uncertainty.

The first is probabilistic modeling.

That is, using probability distributions to account for this uncertainty.

And the second method is by using machine learning, to learn

from previous experience to be able to predict the future uncertain world.

The core structure is as follows,

over the next four weeks you'll be learning four different topics.

In the first week, we will be focused on Gaussian Model Learning.

That is using a Gaussian Model to represent the probability distribution

over potential states.

Then you will learn how to use this Gaussian Model

to do maximum likelihood learning.

In the second week, we will focus on Kalman Filtering.

That is how to model in a probabilistic manner a dynamic world.

And in the third week, you will be doing Robot Mapping, using probabilistic

techniques to map out the surrounding environment around the robot.

And in the fourth week you will learn about robot localization using

particle filters.

That will incorporate different aspects of sensor information

to keep track of the robot's pose over time.