Probabilistic and Statistical Estimation System

Course title

1M987600

Click to show questionnaire result at 2018

Course content

In most intelligent systems including robots, the systems decide their actions based on information obtained by sensors. However, since sensor data contains noise, we need to estimate optimal values to extract necessary information. To do this, statistical and probabilistic approaches have been used. This course covers theoretical fundamentals of estimation methods based on statistical signal processing and estimation theory. Some applications such as mobile robot localization and mapping are also introduced to better understand the methods.

Purpose of class

This course introduces students to the concept of estimation based on probability theory and statistics. It also aims to help student learn how to apply each estimation method through application examples.

Goals and objectives

To explain relationship among estimation methods
To obtain knowledge of characteristics of each estimation method
To solve problems using estimation methods

Language

Japanese(English accepted)

Class schedule

	Class schedule	HW assignments (Including preparation and review of the class.)	Amount of Time Required
1．	Introduction of estimation theory	Survey on estimation theory	60minutes
2．	Cramer-Rao lower bound	Survey on Cramer-Rao lower bound	180minutes
3．	Maximum likelihood estimation	Survey on maximum likelihood estimation	180minutes
4．	Least squares	Survey on least squares	180minutes
5．	Applications (1): Curve fitting	Review of maximum likelihood estimation and least squares Survey on curve fitting	240minutes
6．	Minimum mean square error estimation	Survey on minimum mean square error estimation	180minutes
7．	Maximum a posteriori estimation	Survey on maximum a posteriori estimation	180minutes
8．	Linear minimum mean square error estimation	Survey on linear minimum mean square error estimation	180minutes
9．	Kalman filter (1): Dynamical signal models	Survey on Gauss-Markov process	180minutes
10．	Kalman filter (2): Derivation of Kalman filter	Survey on Kalman filter	180minutes
11．	Extended Kalman filter	Survey on extended Kalman filter	180minutes
12．	Applications (2): Sensor fusion in sensor networks	Review on Kalman filter, survey on information filter	240minutes
13．	Applications (3): Mobile robot localization - Kalman filter approach and Monte Carlo localization	Review on extended Kalman filter, survey on particle filter	240minutes
14．	Applications (4): SLAM (Simultaneous Localization And Mapping)	Review on particle filter, survey on SLAM algorithm	250minutes
Total.	-	-	2650minutes

Relationship between 'Goals and Objectives' and 'Course Outcomes'

	Assignments	Total.
1.	30%	30%
2.	35%	35%
3.	35%	35%
Total.	100%	-

Evaluation method and criteria

Assignments (100%)
- 60% if students can solve example problems given in the lecture.

Textbooks and reference materials

Reference materials are introduced during lectures

Prerequisites

Knowledge of linear algebra, calculus, statistics, and probability theory is required

Office hours and How to contact professors for questions

I accept questions during the lecture or a break after the lecture in the classroom.

Relation to the environment

Non-environment-related course

Regionally-oriented

Non-regionally-oriented course

Development of social and professional independence

Course that cultivates an ability for utilizing knowledge

Active-learning course

N/A

Course by professor with work experience

Work experience	Work experience and relevance to the course content if applicatable
N/A	N/A

Last modified : Thu Mar 21 15:30:41 JST 2019