Project 3 -- Deliberating and Acting

NOTE: The hard copies of the documents for this assignment are due at the beginning of the class period on the date given above. This means that if you are even a minute late, you lose 20%. If you are worried about potentially being late, turn in these documents ahead of time. Do this by submitting them to me during office hours or by sliding them under my office door. Do not leave them in my departmental mail box.

NOTE: Electronic copies of all of the documents for this assignment are due within 24 hours of the time that the hard copies are due. Note that this is 24 hours from the start of the class period. This is not 24 hours from the end of the day in which the hard copy was due.

One problem with purely reactive systems is that they are limited to acting based on their current sensory information. This is a problem because that information may not be sufficient for the robot to efficiently accomplish the task given to it. For example, if you want a robot to get from point A to point B but the robot is unable to sense B from A, then there must be some other way for the robot to know which way to head next or the robot will be likely to take a very inefficient route, if it manages to find B at all. Because it is impractical to create a set of reactive behaviors that will overcome this problem in the general case, robotics researchers often add a deliberative component to the system. This deliberative component can plan an appropriate sequence of actions to carry out based on information not directly available to the robot through its sensors. Such information may be given to the robot from an outside source or recorded by the robot during its previous activities.

This project asks you to design, build, program, and demonstrate a robot that accomplishes a task somewhat different than the one from the previous project. To accomplish this task you will again need to have the robot autonomously carry out several different activities at appropriate times and in appropriate ways based on its sensing. However, in the present project, you will be graded in part on how efficiently your robot carries out the task. Use of a hybrid deliberative/reactive robotic paradigm is suggested, but not required, for this project.

Goals of this project:

• To give you experience engineering a robot that can makes use of information that is not directly sensed to improve the efficiency of its performance.
• To give you additional experience engineering a robot to carry out a task in an uncertain environment. The uncertainty in this environment will come from "noise" in both sensing and acting as well as a dynamic object that will be present in the test environment.
• To give you additional experience engineering a robot to carry out a task that is complex enough that it will need to engage in several different activities at appropriate and not necessarily sequential times.
• To improve and refine your team structure and method of operation.

You will design, build, program, and demonstrate an autonomous robotic system that carries out the following task: To efficiently find and move a set of target objects in the environment to one or more locations within the environment and, if possible, to find and disable a moving dynamic object. You will also turn in written material regarding team structure and organization, timelines and milestones, and design and implementation of the robot and its software. You will complete online evaluations of team members and other teams. One or more team members will present information to the class about their team's robot.

Details of the task (many are similar to those discussed in class) follow:

• Each target object to be found will be a two inch by two inch by two inch orange cube. These objects are likely to be flimsy so your robot will need to handle them delicately. In particular, your robot should not attempt to pick up these objects. Rather, it should push them gently across the floor to one of the destination locations. There will be anywhere from zero to sixteen of these objects in the arena at any given time.
• The destination locations will be marked on the floor using black electrical tape. Each destination will be a one foot by one foot square. There will be anywhere from one to four of these locations marked on the floor during a demonstration.
• Your robot should move exactly one target object to within a destination location at a time. When your robot has moved a target object within a destination location it should give a clearly recognizable signal (such as a sequence of tones) to indicate that it accomplished this goal. When your robot gives this signal, its performance will be scored as follows:
  • If there is exactly one target object within the destination locations, the robot will earn ten points.
  • If the robot has moved this target object from the location at which is was found to the closest destination location, the robot will earn ten additional points.
  • If there are no target objects within the destination locations, the robot will lose one point.
  • If there is more than one target object within the destination locations, the robot will lose one point for each target object within the destination locations.
  Immediately after adjusting the robot's score, all target objects within all destination locations will be removed from the arena.
• At the start of each demonstration, I will provide the team with a list of the coordinates of target objects and destination locations that will be placed into the enclosure (see below). Once the list has been given to the robot, the team will signal the robotic system that the test is to begin immediately.
• The format of the list of coordinates will be:

      x1,y1,x2,y2,...x21,y21
  

  where:
  • each xi,yi is one set of coordinates,
  • x1,y1 through x16,y16 are the coordinates of up to sixteen target objects,
  • x17,y17 through x20,y20 are the coordinates of up to four destination locations, and
  • x21,y21 is the coordinate pair of robot's starting position.

  The coordinates will be measured in feet with a resolution of one half foot. The x coordinate will measure distance from the north edge of the enclosure. The y coordinate will measure distance from the east end of the enclosure.

  As an example, one coordinate pair might be 3,6.5, indicating the object is 3 feet from the north edge and 6.5 feet from the east end of the enclosure. You would not get coordinates such as 7.3 or 2.25, which are not multiples of six inches.

  The special coordinates -1,-1 will indicate that no information is known about that object. For example, I might only provide the coordinates of ten target objects. If I choose to do this, I would provide these coordinates in x1,y1 through x10,y10. Then for x11,y11 through x16,y16 I would enter -1,-1 for each one. It might be the case that there are only ten target objects in the arena at the start of the demonstration or there might be fewer or more such objects (see below). The number of -1,-1 pairs given should not be taken to imply the number of actual objects found in the arena.

  I would prefer to enter these coordinates directly into your laptop in this format. If you would like your system to receive the coordinates in some other way, you will need to have this approved in advance.

• Note that the list of coordinates given may be inaccurate in two ways.
  1. The coordinates given for the objects may be slightly off. For example, an object said to be at 7.5,9 might really be closer to 7,9.5. Note, however, that the locations given for the objects will be accurate to within six inches in each dimension.
  2. The list may include a few spurious coordinates of objects or fail to include the coordinates of a few objects. For example, even though the list may include the coordinates 1,1, there may not be any object in the environment that corresponds with that coordinate pair. Similarly, there might be an object at 7,11, even though there is no corresponding object in the list. Remember, part of your grade is based on how efficiently your robot retrieves a target object. So, if it would be more efficient to move a discovered target object to a destination location, rather than continuing on to a target given in the coordinate list, you would get ten bonus points for retrieving the discovered one. Similarly, if it would be more efficient to move a target object to a discovered destination location, rather than continuing on to a destination location given in the coordinate list, you would get ten bonus points for moving the target object to the discovered destination location.
• The dynamic object will be a simple robot based around the RCX brick. This robot will move around the environment attempting not to hit any objects, including target objects, the walls of the arena, and your robot. However, it may not be perfect in this objective. It will move in simple relatively straight lines. When it detects an object in front, it will back up and turn to one side.
• The dynamic object will have a bumper covered in blue paper on each side. If you make contact with one of these bumpers and depress its contact switch, the dynamic object will be disabled (turn itself off). If your robot succeeds in disabling the dynamic object, your robot earns fifty points.
• Your robot will start each demonstration with zero points.
The remaining items in this list are the same as those from the previous assignment, unless italicized.
• You will construct a robot to complete this task. It may be of any dimension but its primary components must be from the kits provided with limited supplementary items (see below). The robot may be tethered to a remote device outside the arena for additional power or communication with a laptop computer but not for movement, or other purposes. Other than entry of the coordinates before the start of a demonstration, you may not have any contact with a laptop connected to your robot during a demonstration.
• The robot will operate within an enclosure (arena) made from 2x4 lumber. The area enclosed is roughly 8 feet by 12 feet. The bottom of the enclosure will be the tile floor of the lab. There will also be target objects, destination locations, and a dynamic object within the enclosure. Together, the enclosure and these items will constitute the test environment.
• The robotic system will be demonstrated in a series of tests. Prior to the start of each test there will be target objects and a dynamic object placed within the enclosure and destination locations marked on the floor. At the start of each test, the team will place the robot into the enclosure at a location I specify at test time. At that time, the team will signal the robotic system that the test is to begin immediately.
• Team members may not modify the environment prior to, during, or after each test of the robot on this task, except to clean up and restore the environment to its original state if it is modified by the robot.
• Team members may not modify the robot or its code, nor communicate with the robot during each test of the robot on this task with one exception: If the team so chooses, the robot may be returned to its original starting location and reset. Restarting the test in such a manner does not restart the timer. That is, the time already taken by the robot will be deducted from a team's test time.
• Each test will be 20 minutes long. A team must be ready to test its robot when its time starts. The 20 minute period includes time spent by team members while their robot is inactive, as well as time spent by the robot itself.
• You will sign up for up to two test periods, each 30 minutes long. The extra 10 minutes will be used to return the arena to its correct starting conditions and switch between teams. Your score will be the greatest score achieved by your robot on either test.
• Team members may modify the robot and/or the code and may communicate with the robot between tests of the robot on this task.
• The robot may make non-destructive, non-permanent changes to the environment during each test of the robot on this task. Team members must clean up and restore the environment to its original state if it is modified by the robot.
• The robot should primarily use the sensors, actuators, electronics, and building materials provided in the kits. You may augment these components slightly with such materials as tape, rubber bands, wires, etc., but you will need prior approval for any such materials, including what they are and how they will be used. You may not damage or destroy any of the components from the kits.

Team Organization and Task Allocation Proposal.

You will turn in a proposal for how your team will be organized, what tasks need to be carried out, and who will carry out each one. This should include reasons for carrying each task as well as for assigning them to particular people, with an eye towards balancing the amount of work each team member has to complete. You should note here any changes from your previous team organization and method of task allocation with indications for why you are making these changes and where your new ideas originate. Keep in mind that after this there will be no more projects where team members will switch tasks.

This should be from 2 to 2.5 pages in length (roughly 80 characters per line, 50 lines per page). This is due Friday, April 11.

Timeline with Milestones and Fallback Plan.

You will turn in a list describing milestones to be accomplished and an associated timeline showing when each milestone is to be completed. You will also include a fallback plan describing what will be done in the event that a milestone is not reached on time.

This should be from 1.5 to 2 pages in length (roughly 80 characters per line, 50 lines per page). This is due Friday, April 11.

Robot Design

You will turn in a document describing the physical robot design, including its body, suspension, gearing, motors, sensors, and the layout of these components. This does not need to be detailed enough for an exact replica to be made (that is, you don't need to get to the level of the individual Lego pieces). Instead, it should give a high-level description of the parts used and how they are put together. This is the level of description I am looking for (taken with modifications from a description in my Ph.D. thesis of a real robot, constructed for a different task):

This robot consists of two parts, a truck cab unit and a trailer unit. The cab is an Ackerman-steering 4-wheeled vehicle with rear-wheel differential drive. The front of the cab has a binary bumper switch that lets the robot know when it has contacted something from the front. The trailer is attached to the cab by a hitch constructed from a single-turn potentiometer. A servo with a set of photoresistors (CdS cells) is mounted on the back of the trailer so that the robot can keep track of the goal (a light bulb). The trailer also has a binary bumper switch on its rear to tell the robot when it is in contact with the goal.

The potentiometer on the hitch measures the angle that the trailer makes with the cab. Valid angles that the trailer can make are within a 180 degree arc due to the physical constraints of the system. The light tracking servo on the trailer is also limited to a 180 degree arc.

This should be from 1.5 to 2 pages in length (roughly 80 characters per line, 50 lines per page). This is due Wednesday, April 30.

Robot Code

You will turn in the actual code you have written to control this robot. This will be written in Interactive C. Your source code should be well structured and well commented. It should conform to good coding standards (e.g., no memory leaks).

This is due Wednesday, April 30.

Robot Code Documentation

You will turn in a document explaining the data structures and algorithms used in your code.

This should be from 1.5 to 2 pages in length (roughly 80 characters per line, 50 lines per page). This is due Wednesday, April 30.

Presentation Materials

If you choose to use a computer or overhead projector for your presentation, you should turn in a copy of your slides. If you choose to use the board or simply talk, you should turn in an outline of your talk and any notes you use during your presentation.

This is due Wednesday, April 30.

Team Organization Evaluation and Plans

You will turn in a document evaluating how well your team organization worked and giving a plan of what you would do the same and differently if you could do it over again. The plan should be agreed upon by all members. The evaluation can either be a consensus document or individual team members may give independent evaluations.

This should be from 1.5 to 2 pages in length (roughly 80 characters per line, 50 lines per page). This is due Friday, May 2.

Individual Evaluations

You will be given access to web-based evaluation forms for your team members and for the other teams. These are due before class time on Friday, May 2.

Notes on this assignment

You will turn in both a hard copy and an electronic copy of all documentation, including code. Electronic copies should be mailed to hougen@ou.edu as mime enclosures. The electronic copies should be in a format that is stable and easily readable for people using software on a wide variety of operating systems. PostScript and PDF are good candidates for this. MS Word and PowerPoint are not.

You will test your robot during a 30 minute period between 9:00 am and 5:30 pm on Friday, April 25 and/or Monday, April 28 (or on an earlier date, if you so choose). You may schedule up to two test periods, one on each day. The thirty minutes will include set up and clean up times. You must be ready to test your robot when your scheduled time arrives.

Your demonstration will be video-taped for later use, including your use in your presentation if you so choose.

You will also have one or more group members present to the class the key features of your robot, including both its hardware and its software. This presentation will last 8 minutes and will be followed by a 1 minute question and answer period. Your group will be selected for presentation in a random order. You must be ready to present when your group number is called. If you will be using a laptop for your presentation, you should ensure that it works with the classroom projector before April 30.

You will be graded on the following issues:

• 5%: How you organize your team. (One grade for entire team.)
• 5%: How you divide up the tasks. (Either individual or team grades.)
• 5%: How you manage deadlines. (Both individual and team grades.)
• 50%: How you complete the tasks which you have agreed to complete. (Individual grades.)
• 30%: How well the robot accomplishes its task. (One grade for entire team.)
• 5%: How you evaluate the other members of your team. (Individual grades.)

The page limitations given above do not include figures, which are encouraged and may take up any amount of space.

Other

You may write your programs from scratch or may start from programs for which the source code is freely available (such as on the web or from friends or student organizations). If you do not start from scratch, you must give a complete and accurate accounting of where all of your code came from and indicate which parts are original or changed, and which you got from which other source. Failure to give credit where credit is due is academic fraud and will be dealt with accordingly.