EM3: Driver Assistance Systems at Karlsruher Institut Für Technologie | Flashcards & Summaries

Select your language

Suggested languages for you:
Log In Start studying!

Lernmaterialien für EM3: Driver Assistance Systems an der Karlsruher Institut für Technologie

Greife auf kostenlose Karteikarten, Zusammenfassungen, Übungsaufgaben und Altklausuren für deinen EM3: Driver Assistance Systems Kurs an der Karlsruher Institut für Technologie zu.

TESTE DEIN WISSEN

Name some duties of driver assistance systems

Lösung anzeigen
TESTE DEIN WISSEN

- Minimize driver stress
􏰕

- Eliminate deficit in the perception and processing of relevant driver information


- Claim the driver as little as possible
􏰕

- Do not give false warnings
􏰕

- Act as much as possible as the driver himself (during vehicle interventions)


􏰕- Help to avoid mistakes of the driver
􏰕

- Mitigate the consequences of driving errors that occur nevertheless


- Make traffic more efficient, more economical and more environmentally friendly


Lösung ausblenden
TESTE DEIN WISSEN

Describe how a camera is working

Lösung anzeigen
TESTE DEIN WISSEN

- Operating in visible spectral range of light

- Each photodiode forms a pixel

- The larger the pixel the more light can be captured

- RGB color filters allow reconstruction of color in an image


- Each camera is slightly different ➔ Calibration is important

- Rolling shutter effect can cause issues

Lösung ausblenden
TESTE DEIN WISSEN

Name advantages and disadvantages of cameras

Lösung anzeigen
TESTE DEIN WISSEN

Advantages

- High angular resolution

- Allows color vision (e.g. detection of yellow lines)

- Good for object classification

- Comprehensive data for humans

- Cheap

Disadvantages

- Complex distance estimation with mono cameras

- Susceptible to glare

- Restriction at night, rain, fog, 

- Limited range <100m

- Susceptible to poolution

- Demanding data processing

Lösung ausblenden
TESTE DEIN WISSEN

Describe how time of flight sensors are working

Lösung anzeigen
TESTE DEIN WISSEN

- Emit signal and measure time until echo is received

- Distance calculated based on runtime and speed of wave


Challanges

- Echo may never reach sensor (stealth aircraft)

- Minimum object size and distance limited by physical properties

- Limited resolution

- Many disturbances that must be filtered

Lösung ausblenden
TESTE DEIN WISSEN

Describe properties and working methods of ultrasonic sensors

Lösung anzeigen
TESTE DEIN WISSEN

- Piezo material generates impulses and receives echo

- Triangulation with multiple sensors

- 0.25 - 4m range

- Wide horizontal angle

- Use in parking and low speed use cases

- Wind as disturbance factor

Lösung ausblenden
TESTE DEIN WISSEN

Name advantages and disadvantages of ultrasonic sensors

Lösung anzeigen
TESTE DEIN WISSEN

Advantages

- Insensitive to poor visibility

- Detection largely independent of material and surface

- Insensitive to contamination (self cleaning)

- Small design and highly integrated

Disadvantages

- Limited range

- Limited resolution

- Error due do misdirected reflections possible

- Influenced by other sound sources

- Does not read color

Lösung ausblenden
TESTE DEIN WISSEN

Describe the properties and working principles of radar sensors

Lösung anzeigen
TESTE DEIN WISSEN

- Pulse - echo principle with electromagnetic waves

- Continuous wave (relative speed measurement)

- Frequency modulated continuous wave

- Phased array to control measurement direction

- Angle information in horizontal and vertical plane

- Object distance

- Object speed based on phase shift

Lösung ausblenden
TESTE DEIN WISSEN

Name advantages and disadvantages of radar sensors

Lösung anzeigen
TESTE DEIN WISSEN

Advantages

- Independent of weather and time

- Large sensor range (250m)

- Measurement though housing or below other vehicles

Disadvantages

- Changing reflection centres leading to errors

- Limited angle resolution

- Overlapping with other radar sensors

- No color

- Challenging evaluation

Lösung ausblenden
TESTE DEIN WISSEN

Name advantages and disadvantages of lidar sensors

Lösung anzeigen
TESTE DEIN WISSEN

Advantages

- Very good distance resolution

- Detection of smaller objects

- Exact angular resolution

- Immune to other lidar sensors

- daylight independent

- can determine weather condition

Disadvantages

- Problem with glare

- Surface dependency (absorbant materials)

- Not suitable for bad weather

- Mechanical scanner fragile

- Non mechanical scanner very expensive

Lösung ausblenden
TESTE DEIN WISSEN

Name different localisation approaches and their positive and negative aspects

Lösung anzeigen
TESTE DEIN WISSEN
  • GNSS based localization
    • based on satellite system
    • + Cheap
    • + Fusion allows increased accuracy
    • - Problems in urban canyons
    • - No indoor driving
    • - Low accuracy
  • Map based localization
    • Match sensor data with map data to get current position
    • Fusion over time using probabilistic filters
    • + Very accurate
    • + No stellites
    • - Huge amount of data needed
    • - Localisation needs existing map
    • - Dependant on daytime and weather
  • Simultaneous localization and mapping
    • Map generated on the fly
    • + No prior map needed
    • + No satellites needed
    • - Large amount of data
    • - No additional semantic information in map
    • - Accuracy not given during creation of map
    • - High computational effort
Lösung ausblenden
TESTE DEIN WISSEN

Common definition of machine learning

Lösung anzeigen
TESTE DEIN WISSEN

System learns from experience with respect to a class of tasks and a performance measure when its performance in tasks increases through experience.

It generates one or more solution hypotheses to solve the tasks.

Lösung ausblenden
TESTE DEIN WISSEN

Explain principle of supervised learning

Lösung anzeigen
TESTE DEIN WISSEN

Estimation of the best function parameters (hypothesis) using known examples (training data) generated by a unknown system.

Problem is defined by X x Y:

X ➔ True, false: Concept learning

X ➔ Set of classes: Classification

X ➔ R: Numerical regression


Estimate empirical error of given parameter set based on given sample data. Error calculated based on integrated loss function or probability of incorrect result.

Learn Data ➔ learn error

Verification Data ➔ verification error

Test Data ➔ generalisation error

Independent and identical distributed data sets needed!


Calculate best parameters by minimising error function. Gradient descent is one possible method. Parameters are changed based on the derivative of the error function and a constant learning rate.

Lösung ausblenden
  • 160585 Karteikarten
  • 2734 Studierende
  • 204 Lernmaterialien

Beispielhafte Karteikarten für deinen EM3: Driver Assistance Systems Kurs an der Karlsruher Institut für Technologie - von Kommilitonen auf StudySmarter erstellt!

Q:

Name some duties of driver assistance systems

A:

- Minimize driver stress
􏰕

- Eliminate deficit in the perception and processing of relevant driver information


- Claim the driver as little as possible
􏰕

- Do not give false warnings
􏰕

- Act as much as possible as the driver himself (during vehicle interventions)


􏰕- Help to avoid mistakes of the driver
􏰕

- Mitigate the consequences of driving errors that occur nevertheless


- Make traffic more efficient, more economical and more environmentally friendly


Q:

Describe how a camera is working

A:

- Operating in visible spectral range of light

- Each photodiode forms a pixel

- The larger the pixel the more light can be captured

- RGB color filters allow reconstruction of color in an image


- Each camera is slightly different ➔ Calibration is important

- Rolling shutter effect can cause issues

Q:

Name advantages and disadvantages of cameras

A:

Advantages

- High angular resolution

- Allows color vision (e.g. detection of yellow lines)

- Good for object classification

- Comprehensive data for humans

- Cheap

Disadvantages

- Complex distance estimation with mono cameras

- Susceptible to glare

- Restriction at night, rain, fog, 

- Limited range <100m

- Susceptible to poolution

- Demanding data processing

Q:

Describe how time of flight sensors are working

A:

- Emit signal and measure time until echo is received

- Distance calculated based on runtime and speed of wave


Challanges

- Echo may never reach sensor (stealth aircraft)

- Minimum object size and distance limited by physical properties

- Limited resolution

- Many disturbances that must be filtered

Q:

Describe properties and working methods of ultrasonic sensors

A:

- Piezo material generates impulses and receives echo

- Triangulation with multiple sensors

- 0.25 - 4m range

- Wide horizontal angle

- Use in parking and low speed use cases

- Wind as disturbance factor

Mehr Karteikarten anzeigen
Q:

Name advantages and disadvantages of ultrasonic sensors

A:

Advantages

- Insensitive to poor visibility

- Detection largely independent of material and surface

- Insensitive to contamination (self cleaning)

- Small design and highly integrated

Disadvantages

- Limited range

- Limited resolution

- Error due do misdirected reflections possible

- Influenced by other sound sources

- Does not read color

Q:

Describe the properties and working principles of radar sensors

A:

- Pulse - echo principle with electromagnetic waves

- Continuous wave (relative speed measurement)

- Frequency modulated continuous wave

- Phased array to control measurement direction

- Angle information in horizontal and vertical plane

- Object distance

- Object speed based on phase shift

Q:

Name advantages and disadvantages of radar sensors

A:

Advantages

- Independent of weather and time

- Large sensor range (250m)

- Measurement though housing or below other vehicles

Disadvantages

- Changing reflection centres leading to errors

- Limited angle resolution

- Overlapping with other radar sensors

- No color

- Challenging evaluation

Q:

Name advantages and disadvantages of lidar sensors

A:

Advantages

- Very good distance resolution

- Detection of smaller objects

- Exact angular resolution

- Immune to other lidar sensors

- daylight independent

- can determine weather condition

Disadvantages

- Problem with glare

- Surface dependency (absorbant materials)

- Not suitable for bad weather

- Mechanical scanner fragile

- Non mechanical scanner very expensive

Q:

Name different localisation approaches and their positive and negative aspects

A:
  • GNSS based localization
    • based on satellite system
    • + Cheap
    • + Fusion allows increased accuracy
    • - Problems in urban canyons
    • - No indoor driving
    • - Low accuracy
  • Map based localization
    • Match sensor data with map data to get current position
    • Fusion over time using probabilistic filters
    • + Very accurate
    • + No stellites
    • - Huge amount of data needed
    • - Localisation needs existing map
    • - Dependant on daytime and weather
  • Simultaneous localization and mapping
    • Map generated on the fly
    • + No prior map needed
    • + No satellites needed
    • - Large amount of data
    • - No additional semantic information in map
    • - Accuracy not given during creation of map
    • - High computational effort
Q:

Common definition of machine learning

A:

System learns from experience with respect to a class of tasks and a performance measure when its performance in tasks increases through experience.

It generates one or more solution hypotheses to solve the tasks.

Q:

Explain principle of supervised learning

A:

Estimation of the best function parameters (hypothesis) using known examples (training data) generated by a unknown system.

Problem is defined by X x Y:

X ➔ True, false: Concept learning

X ➔ Set of classes: Classification

X ➔ R: Numerical regression


Estimate empirical error of given parameter set based on given sample data. Error calculated based on integrated loss function or probability of incorrect result.

Learn Data ➔ learn error

Verification Data ➔ verification error

Test Data ➔ generalisation error

Independent and identical distributed data sets needed!


Calculate best parameters by minimising error function. Gradient descent is one possible method. Parameters are changed based on the derivative of the error function and a constant learning rate.

EM3: Driver Assistance Systems

Erstelle und finde Lernmaterialien auf StudySmarter.

Greife kostenlos auf tausende geteilte Karteikarten, Zusammenfassungen, Altklausuren und mehr zu.

Jetzt loslegen

Das sind die beliebtesten StudySmarter Kurse für deinen Studiengang EM3: Driver Assistance Systems an der Karlsruher Institut für Technologie

Für deinen Studiengang EM3: Driver Assistance Systems an der Karlsruher Institut für Technologie gibt es bereits viele Kurse, die von deinen Kommilitonen auf StudySmarter erstellt wurden. Karteikarten, Zusammenfassungen, Altklausuren, Übungsaufgaben und mehr warten auf dich!

Mehr Karteikarten anzeigen

Das sind die beliebtesten EM3: Driver Assistance Systems Kurse im gesamten StudySmarter Universum

ERP-Systems

Frankfurt University of Applied Sciences

Zum Kurs
Science - Systems

University College of the Cariboo

Zum Kurs
Embedded Systems

University of Durham

Zum Kurs
M03 Asset maintenance systems

University of Manchester

Zum Kurs

Die all-in-one Lernapp für Studierende

Greife auf Millionen geteilter Lernmaterialien der StudySmarter Community zu
Kostenlos anmelden EM3: Driver Assistance Systems
Erstelle Karteikarten und Zusammenfassungen mit den StudySmarter Tools
Kostenlos loslegen EM3: Driver Assistance Systems