Probabilistic Robotics (MATLAB)
Budget: $30 – $250 USD
Task 1: Robot model (10%)
With reference to the robot description files for Stage and the MATLAB robot controller files provided, generate a specification sheet that contains the following details about the platform, sensors and environment:
i) Robot chassis dimensions ii) Pose of each range finder (robot coordinate frame) iii) Max. range of range finders and fiducial detector iv) Field of view of range finders and fiducial detector v) Robot motion error parameters vi) Variance in range and bearing measurements from fiducial detector vii) Size of warehouse
viii)Number of fiducials in the warehouse
ix) List of coordinates of each fiducial (world coordinate frame) x) Initial pose of the robot (world coordinate frame)
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
Task 2: Mapping (20%)
Generate 2 figures using MATLAB and the data set provided. The data was captured using the robot model and controller interacting with the Stage simulator via ROS. You should refer to the robot controller MATLAB files to determine the content and format of each file in the data set.
Figure 1:
i) Plot the ground truth path taken by the ED 208 in black.
ii) Overlay this plot with a plot of the noisy pose estimates in red.
iii) Mark the ground truth locations of the fiducials as blue circles on the plot
Figure 2:
Generate an occupancy grid map using the ground truth pose and range finder readings from the data set provided
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
Task 3: Extended Kalman Filter based localisation (35%)
Implement an Extended Kalman Filter (EKF) based localisation algorithm using MATLAB.
Document your EKF algorithm using clearly commented MATLAB code.
Demonstrate performance of your EKF using the data set provided by producing a figure that overlays ground truth pose, noisy pose estimate and best estimated pose generated by your EKF for each data point in the set.
In a second figure, overlay a plot of the L2Norm distance between ground truth pose and noisy estimated pose against the L2Norm between ground truth pose and EKF best estimated pose for each data point in the set.
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
Task 4: Particle Filter based localisation (35%)
Implement a Particle Filter (PF) based localisation algorithm using MATLAB.
Document your PF algorithm using clearly commented MATLAB code
Demonstrate performance of your PF using the data set provided by producing a figure that overlays ground truth pose, noisy pose estimate and best estimated pose generated from PF for each of the data points in the set
In a second figure, overlay a plot of the L2Norm distance between ground truth pose and noisy estimated pose against the L2Norm between ground truth pose and PF best estimated pose for each data point
With reference to the robot description files for Stage and the MATLAB robot controller files provided, generate a specification sheet that contains the following details about the platform, sensors and environment:
i) Robot chassis dimensions ii) Pose of each range finder (robot coordinate frame) iii) Max. range of range finders and fiducial detector iv) Field of view of range finders and fiducial detector v) Robot motion error parameters vi) Variance in range and bearing measurements from fiducial detector vii) Size of warehouse
viii)Number of fiducials in the warehouse
ix) List of coordinates of each fiducial (world coordinate frame) x) Initial pose of the robot (world coordinate frame)
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
Task 2: Mapping (20%)
Generate 2 figures using MATLAB and the data set provided. The data was captured using the robot model and controller interacting with the Stage simulator via ROS. You should refer to the robot controller MATLAB files to determine the content and format of each file in the data set.
Figure 1:
i) Plot the ground truth path taken by the ED 208 in black.
ii) Overlay this plot with a plot of the noisy pose estimates in red.
iii) Mark the ground truth locations of the fiducials as blue circles on the plot
Figure 2:
Generate an occupancy grid map using the ground truth pose and range finder readings from the data set provided
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
Task 3: Extended Kalman Filter based localisation (35%)
Implement an Extended Kalman Filter (EKF) based localisation algorithm using MATLAB.
Document your EKF algorithm using clearly commented MATLAB code.
Demonstrate performance of your EKF using the data set provided by producing a figure that overlays ground truth pose, noisy pose estimate and best estimated pose generated by your EKF for each data point in the set.
In a second figure, overlay a plot of the L2Norm distance between ground truth pose and noisy estimated pose against the L2Norm between ground truth pose and EKF best estimated pose for each data point in the set.
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
Task 4: Particle Filter based localisation (35%)
Implement a Particle Filter (PF) based localisation algorithm using MATLAB.
Document your PF algorithm using clearly commented MATLAB code
Demonstrate performance of your PF using the data set provided by producing a figure that overlays ground truth pose, noisy pose estimate and best estimated pose generated from PF for each of the data points in the set
In a second figure, overlay a plot of the L2Norm distance between ground truth pose and noisy estimated pose against the L2Norm between ground truth pose and PF best estimated pose for each data point