Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

A LEVEL 4 MODULE, AUTUMN SEMESTER 2021-2022

COMP4034  Autonomous Robotic Systems

Question 1:

A mobile robot is deployed as a security patrol robot in an office environment mainly consisting of corridors, with many spots that appear similar to the robot due to perceptual aliasing and the limitations of the robot’s sensors. The robot is equipped with odometry sensors (wheel encoders) and a forward-facing Kinect sensor, which is capable of providing both colour vision and depth information. Assume that the robot has already acquired an accurate and up-to-date 2D metric map of the environment using the same sensors and an appropriate mapping algorithm.

Your task is to describe an appropriate self-localisation algorithm for effectively tracking the 2D pose of the robot during operation.

[overall <30> marks]

a.  Which state values will be estimated by your algorithm?

[<3> Marks]

b.  Name a tracking algorithm we studied in the lectures, which is appropriate for this          problem, explain its main steps of using equations and pseudocode, taking care to          properly describe the inputs and resulting outputs of each step. You may use diagrams as appropriate.

[<17> Marks]

c.   Outdoor mobile robots may also access measurements from Global Positioning System (GPS) satellites to help localise themselves. Explain how measurements from GPS can be integrated into your localisation framework in addition to the odometry and Kinect measurements. Explicitly indicate at which step of the algorithm you will use the GPS data and for what purpose.

[<10> Marks]

Question 2:

In your assignment, you worked on developing a mobile patrol robot, which searched for           predefined objects in a scene. Let’s assume that your robot is kidnapped and placed randomly   at a key location in the scene. Assume that these key locations are already saved in your robot’s memory as “snapshot signatures”, which are scans of range measurements obtained by firing    the laser scanner over 360 degrees around the robot, at each of these locations. Your task is to perform global localisation to find out where your robot is.

Provide the pseudo-code for a ROS program, which allows the robot to recognise which of the snapshot locations it is in, regardless of what orientation it is placed at. Your code should also be able to detect if the robot is not at a saved snapshot location. Please write down your assumptions, such as the topic names and message types.

[overall <25> marks]

Question 3:

You are asked to implement a simple program for a differential drive robot. Your robot has a Lidar mounted on top, similar to the Turtlebot3 Waffle you used in your lab sessions. Your program should enable the robot to explore underground tunnels, which have uneven widths, however are always wide enough for your robot to pass through.

Unfortunately, the robot can be driven through controlling wheel velocities only, so you cannot simply publish a Twist message to drive the robot as you have practiced in the minitasks. Instead, the   controller   node   responsible   for   setting   the   motion   listens   to   a   topic   named /wheel_speed_cmd of message type std_msgs/Float32MultiArray, which contains the left and right wheel velocities in array format: [v_left, v_right]. Passing the data [0.2, 0] will rotate only the left wheel with a speed of 0.2 m/s; [0, 0.2] will rotate only the right wheel with 0.2 m/s; [0.2, 0.2] will rotate both wheels with 0.2 m/s, enabling a straight forward motion.

The    Lidar   data    is    published   over   a   topic    named   /scan,   of   message   type

sensor_msgs/LaserScan.

Provide the pseudo-code for a ROS program, which steers the robot safely in the tunnels with a constant forward velocity of 0.5 m/s, but also maintaining the same distance to both walls.

[overall <25> marks]

Question 4:

An exciting use of mobile robots in the the consumer market is the robotic telepresence robots, similar to the Ohmni robot we used in the first few physical lab sessions. These robots consist of a mobile base and a tablet that enables video connection to a remote user. The robot can be controlled  by  the  remote  user through  a  graphical  user  interface,  to  select  waypoints  to autonomously navigate the robot, as well as to control the height and orientation of the tablet.   Considering their use in an educational scenario, where the robot is used by a remote teaching assistant to provide lab support within a crowded human space, answer the following questions:

[overall <20> marks]

a.  What problems could there be with usability and/or user experience? Please indicate two specific issues for the students and two specific issues for the remote teacher?

[<8> Marks]

b.  Briefly describe how you can critically evaluate the performance of a telepresence robot in a teaching environment. Suggest a set of metrics to be used in this evaluation.

[<12> Marks]