ASSIGNMENT DETAILS

This assessment covers the development in Jason of a multi-agent system for a “search & rescue” mission. This assignment counts for 50% of the final grade for the module. Note that you will be working with your assignment 2 group, which is different to the group you worked with in assignment 1.

The scenario is as follows: a doctor agent has information about 5 possible locations of 3 victims in a given area.  Both agents have a map of the arena and know of the location of possible obstacles.  The doctor directs the scout agent to look for the victims in these locations with the smallest number of moves possible.  Once three victims have been found (and the severity of their injuries has been assessed), the mission ends, at which point the doctor reports back the location of the victims.  However, a complicating factor is that the scout doesn’t know its starting position, and thus needs to figure this out. 

This is the detailed specification of the assignment:

1. •The multi-agent system consists of two Jason agents: a scout and a doctor.
   

2. •The scout agent is responsible for controlling the EV3 robot, and is responsible for getting percepts from the robot, as well as directing the robots movements around the arena.
   

3. •The doctor agent is responsible for planning the scout’s actions and directing the scout agent.
   

4. •Both agents will be given a set of beliefs which correspond to the location of obstacles on a map of 6x7 cells.
   

5. •Each obstacle will be about 21cm width, 33 cm depth, and 21cm high (we will use empty photocopier paper boxes). They are intended to be easily detectable.
   

6. •The doctor agent will also possess a set of five beliefs of possible locations of the victims.  Three will be correct, but two will be incorrect.
   

7. •Victims will be designated through colored A4 papers placed on the floor of the arena.
   

8. •Different colors indicate different levels of severity of injury that victims are subjected to:
   

9. •red = critical
   

10. •blue = serious
    

11. •green = minor
    

12. •To assist with correcting your odometry, you can use the cell boundaries (i.e., the coloured tape), the walls and obstacles, all of which can be detected by your robot.  However, you should not use the coloured tape to determine your localisation.
    

13. •You will demonstrate your final solution on an arena with cells marked out in black tape.
    

14. •The accuracy of the proximity sensors when sensing the walls may degrade the further away you are from the walls.  This is intentional!
    

15. •You c
    

1. •The doctor agent
   

2. •The doctor agent provides in real time on the screen the status of the rescue mission.  This should include details of which of the 5 victim locations has been checked, the location of any victims found (and their severity) and the location (or possible locations) of the scout.  A count of the cells visited should also be recorded.
   

3. •The real-time details can be presented textually, but extra marks will be awarded for presenting this information graphically.
   

4. •Not all of the victim locations need to be visited; once all three victims have been found, there will be no need to check the other locations.
   

5. •The doctor can make use of any path planning algorithm, which can be implemented as an action (in java) in the Jason environment.
   

6. •Communication between the scout and the doctor should be conducted using agent communication (as discussed in the lectures).
   

1. •The scout agent
   

2. •Communication between the scout and the EV3 robot should be through the exchange of messages using sockets (through bluetooth, etc).
   

3. •The scout agent will start anywhere on the arena.  However, it be placed in the middle of a cell, aligned with the walls of the cell.
   

4. •The scout agent should not do any path planning, but should be responsible for maintaining a set of beliefs regarding its location.
   

5. •The scout agent should inform the doctor agent where it believes it is - if there are several places, then the scout agent can inform the doctor agent of all of these places.  It should also inform the doctor agent when it no longer believes it is in a location (or locations).
   

6. •The EV3 robot should display the status of the map, the location of the known victims and possible locations, and information about where the scout believes it is on the LCD screen. When the robot finishes the demo, the final map should be displayed until a final button (e.g. ESCAPE) is pressed, so that the map can be inspected.
   

1. •Finally:
   

2. •You can use code from any of your group member’s assignment 1 solutions.
   

3. •You can use line maps and waypoints, or occupancy grids to model the maps.
   

4. •Examples of the beliefs of the arena will be published shortly.
   

Obtaining robots

1. •Robots are available from the HelpDesk during lab hours.  If the lab is not open, then you can try asking Dr Payne (Terry) in 218, or Mr Alcock (Josh) in 213 to get access to the Robots.
   

2. •One robot will be allocated to each group; please make sure that you always use the same robot (this should be the same robot id as your group id.
   

3. •Robots cannot be taken out of the building (e.g. over night or over the weekend).
   

SUBMISSION INSTRUCTIONS

You should work with your assigned groups.  Your solution should be submitted ELECTRONICALLY as follows:

1. •Each member of the group is responsible for submitting a zip file containing:
   

1. 1.A log of their individual activity as a pdf file; and
   

2. 2.A filled in peer-evaluation form.
   

3. •Furthermore, each group should nominate a single member of the group who will be responsible for submitting:
   

3. 3.The source code for the group solution; and
   

4. 4.A (joint) complete log of the meetings and minutes for your group.
   

To summarise: in each group, two members will submit just their individual peer review form and individual activity report.  These should not be shared with other students.  The third member of the group will submit a zip file containing the source code and a joint log in addition to their individual peer review form and individual activity report.

Your solutions must be submitted to the departmental electronic submission system, and select ‘Assignment 2’ from the drop-down menu.  You then locate the zip file containing your submission, check the box stating that you have read and understood the university’s policy on plagiarism and collusion, then click the ‘Upload File’ button.

Each group will have the choice of either demonstrating their solution during the last week of term (week 12) or the first week of next term (week 13).  Partial marks will be awarded for partial solutions.  As an insurance, groups should consider preparing a video in case there are problems with the demonstration.

MARKING SCHEME - see opposite.

Individual marks will be based on the team mark adjusted according to the peer-evaluation form.

Purpose of Assessment & Learning Outcomes

The purpose of the assessment is to provide the student with experience programming an autonomous robot, in particular in the exploration and mapping of a space, and to assess the learning outcomes 1, 2, 4, 5, and 6 (in bold/italic) from the following stated outcomes for the module:

1. 1.explain the notion of an agent, how agents are distinct from other software paradigms (e.g., objects), and judge the characteristics of applications that lend themselves to an agent-oriented solution;
   

2. 2.identify the key issues associated with constructing agents capable of intelligent autonomous action;
   

3. 3.describe the main approaches taken to developing such agents;
   

4. 4.use a contemporary agent programming platform (e.g., AgentSpeak) for developing significant software or hardware-based agents;
   

5. 5.identify key issues involved in building agents that must sense and act within the physical world;
   

6. 6.program and deploy autonomous robots for specific tasks.