Intelligent Autonomous Robotics Practicals

Semester 1 2016/2017

Note that the first session is in week 1 which is when you will be assigned a robot. All sessions are Tuesdays 4-6pm in G.A11 Forrest Hill.

Lecturer: Barbara Webb

This page describes the assignments and deadlines for the IAR course. More information, such as how to use the robots, answers to questions arising etc. can be found at the IAR practicals wiki

Demonstrator: Your first contact for problems with the practicals is Dylan Ross.

Summary of deadlines

Week 3 Task 1: Robot avoids obstacles and follows walls. First report due Thursday 6th October Demo due at practical week 3
Week 5 Task 2: Robot can do odometry. Second report due Thursday 20th October Demo due at practical week 5
Week 8-9 Task 3: Robot collects maximum food in minimum time. Final Report due Thursday 17th November Final competition at practical week 8

You should submit just one joint report per group. If there is any reason you do not think the marks for the report should be distributed evenly amongst the group members, please put a note on the report stating what percentage split you have agreed. Please also make sure that the front page of your report has your names and student numbers on it.

Submission should be an electronic document in pdf format, submitted via the on-line submit system by 4pm:

submit iar 1 Your_surnames_report1.pdf
(For later reports replace "1" with 2 or 3 as appropriate)

Feedback

The first two tasks are assessed, but worth fewer marks (2% for the demo, 8% for the report), as their purpose to make sure you are making steady progress at the start of semester, and to give you feedback as you go on how to write a good report. But note, the demo is required; reports will not be marked unless we have actually seen your robot attempt the task during the designated practical session. As detailed below, you will get back your marked report with written comments at the practical following your submission, and will have the opportunity to discuss it at that time with the marker to clarify any issues. You will also get formative feedback from the tutor at the practicals preceding submission and from the demo about your approach to solving the task and whether this will meet the assignment criteria.

Further guidance on report writing here

Good scholarly practice: please remember the University requirement that material submitted for assessment is entirely the student's own work. Details about this can be found here and here .

Task 1

First, your robot should move around without hitting obstacles or getting stuck in corners or dead-ends. Second, it should tend to follow long walls, keeping a consistent distance away from the wall.

In implementing this task you should particularly explore different ways of using the eight IR sensors: for example, reactively using one pair of IR for avoiding obstacles and another pair for wall following; or combining all the readings to try to identify the shape around the robot before deciding how to act. You should also explore whether particular ways of moving are more or less effective: for example on-the-spot turns vs. curved paths. In your report you could discuss whether you think a change in the physical design of the robot could have made the task easier.

Demonstration (2 marks) due at practical week 3.

Report (8 marks) on Thursday week 3 (6th October) at 4pm.

Submission should be an electronic document in pdf format, submitted via the on-line submit system by 4pm:

submit iar 1 Your_surnames_report1.pdf
Marked reports will be returned at the practical in week 4.

Task 2

The aim of this task is to implement basic odometry using the wheel encoders on your robot (see IVR practical 5) so that while avoiding obstacles and following walls, it maintains an estimate of where it is relative to its starting point. You could use this to draw a track of where the robot was during a test (this will get you one mark in the demo) and see if it is able to get back to its starting point either by retracing its outward route or more directly. The second mark in the demo will be given if your robot can be started at an arbitrary location, moves around the arena (at a reasonable speed, avoiding obstacles) for 30 seconds, and then returns to within 10cm of its start position. If you feel really ambitious, you could to build a basic map of the arena by having the robot record, along with its current position, what walls or obstacles were detected around it in that position, and redrawing these as well. Note for this task you should not assume a fixed starting point or fixed location of walls and obstacles in the arena.

Demonstration (2 marks) due at practical week 5.

Report (8 marks) on Thursday week 5 (20th October) at 4pm.

Submission should be an electronic document in pdf format, submitted via the on-line submit system by 4pm:

submit iar 2 Your_surnames_report2.pdf
Marked reports will be returned at the practical in week 6.

Task 3

This task is inspired by ant navigation capabilities, but you can use any algorithm you like to solve it (i.e. it does not have to be a bio-inspired algorithm).

The robot will start from a particular location, the 'nest' (marked by small cross on floor). The robot needs to explore the arena looking for 'food' (small circles on the floor). Your robot should also be monitoring for a keypress, and during the test, one of the demonstrators will press the key when the robot has found (i.e. driven over) the food. The robot should stop when this happens to indicate that it has 'picked up' the food item. That food location will now be considered empty until the robot has visited the nest location.

To gain points, the robot has to deposit as much 'food' as possible at the nest within 5 minutes. Food is deposited by the robot returning to the nest location and indicating it is dropping off food by flashing its LEDs three times (we will allow getting back to within 10cm radius of the nest location to count as success).

You can decide whether you want the robot to go home every time it picks up food; whether it should keep going back to the same place to get more, or should explore for new sources; or if it should try to collect multiple food items from different places before going home. However, we suggest you start by implementing a simple but reliable random search and direct return strategy based on odometry before spending too much time devising more complex solutions, which may not work.

The layout of the arena (walls and obstacles) will be fixed and you can programme the robot with a map if you want. However, food locations will not be known in advance. Here is an example layout from last year: Khepera arena

Competition at practical week 8. Note this is more than a week earlier than the due date of the report so that you have time to gather data on your robot's performance after the competition.

Final Report (30 marks) on Thursday week 9 (17th November) at 4pm.

Submission should be an electronic document in pdf format, submitted via the on-line submit system by 4pm:

submit iar 3 Your_surnames_report3.pdf
Marked reports will be returned via the ITO by week 10.

Last update: August 25, 2016.


Home : Teaching : Courses : Iar 

Informatics Forum, 10 Crichton Street, Edinburgh, EH8 9AB, Scotland, UK
Tel: +44 131 651 5661, Fax: +44 131 651 1426, E-mail: school-office@inf.ed.ac.uk
Please contact our webadmin with any comments or corrections. Logging and Cookies
Unless explicitly stated otherwise, all material is copyright © The University of Edinburgh