7a - Self Localisation Flashcards

1
Q

What is the key point of self-localisation?

A

So a robot can find its position and situation in the environment

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What can be used to approximate functions?

A

Particle sets

The more particles fall into an interval, the higher the probability of that interval

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is a particle?

A

An individual state estimate

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is a particle defined by?

A

Its state values that determine the robot’s pose in the plane

A weight that indicates its likelihood

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What do particle filters use?

A

Many particles to represent the belief state (where am I?)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What are the 4 stages to particle filters?

A

Initialisation, Prediction, Update, Resample

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is initialisation in particle filter localisation?

A

Sampling from initial distribution

No idea where robot is, so throw particles everywhere

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What are the 3 stages that loop in particle filters?

A

Prediction, Update, Resample

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What happens in the prediction step?

A

For each particle, sample and add random noisy values from the motion model

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What happens in the update step?

A

Each particle’s weight is the likelihood of getting the current sensor readings from that particle’s hypothesis (compared to the predicted readings from the map)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What happens in the resample step?

A

A new set of particles is chosen such that each particle survives in proportion to its weight

Survival of the fittest: replace unlikely samples by more likely ones

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What can particle filters represent?

A

Arbitrary probability density functions (distributions) using samples

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What can particle filters solve?

A

The kidnapped robot problem and handle perceptually aliased environments

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is Kalman filter localisation?

A

Instead of an arbitrary density function, KF uses Gaussians for robot belief, motion and measurement models

Only mean and covariance needs to be updated - efficient computation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What does Kalman filter localisation address?

A

Addresses position tracking, not global localisation or kidnapped robot problem

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What are the 5 steps to Kalman Filter localisation?

A

Prediction (ACT) based on previous estimate and odometry

Observation (SEE) with on-board sensors

Measurement prediction based on prediction and map

Matching of observation and map

Estimation -> position update

17
Q

Does does Markov Localisation do?

A

Uses an explicitly specified probability distribution across all possible robot positions

18
Q

What are the pros of Markov localisation?

A

Allows for localisation starting from any unknown position

Can recover from ambiguous situations

19
Q

What are the cons of Markov localisation?

A

Requires a discrete representation of the space, such as a geometric grid of topological graph

Requires memory and computational power

20
Q

What are the pros of Kalman filter localisation?

A

Tracks the robot from an initially known position

Precise and efficient

Can be used in continuous world representations

21
Q

What are cons to Kalman filter localisation?

A

If uncertainty of robot becomes large (collision with object), it can fail to capture the multitude of possible robot positions and become irrevocably lost