CS-7638 FINAL REVIEW PAPER 2026
COMPLETE QUESTIONS AND SOLUTIONS
GRADED A+
◉What additional equipment is used in self-driving cars for
localization? Answer: A stereo camera system and GPS antennas.
◉What is the main purpose of tracking in self-driving cars? Answer:
To understand the position and speed of other vehicles to avoid
collisions.
◉How do Kalman filters differ from Monte Carlo localization?
Answer: Kalman filters estimate a continuous state, while Monte
Carlo localization uses discrete places.
◉What type of distribution do Kalman filters produce? Answer: A
unimodal distribution.
◉What is the goal of using a Kalman filter in tracking? Answer: To
estimate future locations and velocities based on noisy and
uncertain data.
,◉What is a Gaussian in the context of Kalman filters? Answer: A
continuous function characterized by a mean (μ) and variance (σ²)
that represents the probability distribution of the state.
◉What does the area under a Gaussian curve represent? Answer: It
sums up to 1, indicating the total probability.
◉What are the two parameters that characterize a Gaussian?
Answer: The mean (μ) and the variance (σ²).
◉What does a larger variance (σ²) indicate about a distribution?
Answer: It indicates greater uncertainty about the actual state.
◉What is the relationship between covariance and the spread of a
Gaussian function? Answer: Larger covariance results in a wider
spread of the function.
◉What does the term 'unimodal' refer to in Gaussian distributions?
Answer: It refers to distributions that have a single peak.
◉What is the significance of the quadratic function in the Gaussian
formula? Answer: It helps to determine the shape of the Gaussian
distribution based on the distance from the mean.
,◉What is the expected output of the Kalman filter when given noisy
measurements? Answer: An estimate of future locations and
velocities that accounts for uncertainty.
◉What is the purpose of normalization in the Gaussian formula?
Answer: To ensure that the area under the curve sums to 1.
◉What is the main focus of the Kalman filter class mentioned in the
notes? Answer: To teach how to write software that estimates future
locations and velocities using sensor data.
◉What is the role of the Google self-driving car in the context of
Kalman filters? Answer: It uses methods like Kalman filters to
understand the position of other traffic based on radar and laser-
range data.
◉How does the Kalman filter handle uncertainty in measurements?
Answer: By maintaining estimates of the mean and variance to
represent the state of the system.
◉What is the significance of the exponential function in the
Gaussian distribution? Answer: It describes how the probability
decreases as you move away from the mean.
, ◉What is the expected behavior of an object moving with constant
velocity in a Kalman filter? Answer: The filter predicts future
positions based on past measurements and assumed constant
velocity.
◉How does the Kalman filter improve over time? Answer: By
continuously updating its estimates based on new measurements.
◉What is the relationship between the Kalman filter and particle
filters? Answer: Both are techniques for estimating state, but
particle filters can handle multimodal distributions.
◉What is the importance of understanding sensor data in self-
driving cars? Answer: It is crucial for making assessments about the
environment and ensuring safe navigation.
◉What is the preferred Gaussian when tracking another car with a
self-driving car? Answer: The third Gaussian, as it is the most
certain and minimizes the chance of an accident.
◉What are the characteristics of a Gaussian distribution? Answer:
Gaussians are unimodal distributions that are symmetrical.
◉How do you evaluate a Gaussian with μ = 10, σ² = 4, and x = 8?
Answer: The approximate answer is 0.12.
COMPLETE QUESTIONS AND SOLUTIONS
GRADED A+
◉What additional equipment is used in self-driving cars for
localization? Answer: A stereo camera system and GPS antennas.
◉What is the main purpose of tracking in self-driving cars? Answer:
To understand the position and speed of other vehicles to avoid
collisions.
◉How do Kalman filters differ from Monte Carlo localization?
Answer: Kalman filters estimate a continuous state, while Monte
Carlo localization uses discrete places.
◉What type of distribution do Kalman filters produce? Answer: A
unimodal distribution.
◉What is the goal of using a Kalman filter in tracking? Answer: To
estimate future locations and velocities based on noisy and
uncertain data.
,◉What is a Gaussian in the context of Kalman filters? Answer: A
continuous function characterized by a mean (μ) and variance (σ²)
that represents the probability distribution of the state.
◉What does the area under a Gaussian curve represent? Answer: It
sums up to 1, indicating the total probability.
◉What are the two parameters that characterize a Gaussian?
Answer: The mean (μ) and the variance (σ²).
◉What does a larger variance (σ²) indicate about a distribution?
Answer: It indicates greater uncertainty about the actual state.
◉What is the relationship between covariance and the spread of a
Gaussian function? Answer: Larger covariance results in a wider
spread of the function.
◉What does the term 'unimodal' refer to in Gaussian distributions?
Answer: It refers to distributions that have a single peak.
◉What is the significance of the quadratic function in the Gaussian
formula? Answer: It helps to determine the shape of the Gaussian
distribution based on the distance from the mean.
,◉What is the expected output of the Kalman filter when given noisy
measurements? Answer: An estimate of future locations and
velocities that accounts for uncertainty.
◉What is the purpose of normalization in the Gaussian formula?
Answer: To ensure that the area under the curve sums to 1.
◉What is the main focus of the Kalman filter class mentioned in the
notes? Answer: To teach how to write software that estimates future
locations and velocities using sensor data.
◉What is the role of the Google self-driving car in the context of
Kalman filters? Answer: It uses methods like Kalman filters to
understand the position of other traffic based on radar and laser-
range data.
◉How does the Kalman filter handle uncertainty in measurements?
Answer: By maintaining estimates of the mean and variance to
represent the state of the system.
◉What is the significance of the exponential function in the
Gaussian distribution? Answer: It describes how the probability
decreases as you move away from the mean.
, ◉What is the expected behavior of an object moving with constant
velocity in a Kalman filter? Answer: The filter predicts future
positions based on past measurements and assumed constant
velocity.
◉How does the Kalman filter improve over time? Answer: By
continuously updating its estimates based on new measurements.
◉What is the relationship between the Kalman filter and particle
filters? Answer: Both are techniques for estimating state, but
particle filters can handle multimodal distributions.
◉What is the importance of understanding sensor data in self-
driving cars? Answer: It is crucial for making assessments about the
environment and ensuring safe navigation.
◉What is the preferred Gaussian when tracking another car with a
self-driving car? Answer: The third Gaussian, as it is the most
certain and minimizes the chance of an accident.
◉What are the characteristics of a Gaussian distribution? Answer:
Gaussians are unimodal distributions that are symmetrical.
◉How do you evaluate a Gaussian with μ = 10, σ² = 4, and x = 8?
Answer: The approximate answer is 0.12.
