Midterm Competition Post Mortem

Transcription

1 Midterm Competition Post Mortem Why didn t my robot work? 1

2 Class Suggestions? More sensors Calibrate better for the turns Use fresh batteries (duracell), varying power affects the performance Use the colour red, not blue Competition sets the max time taken per obstacle, not up to the team lab/lecture crossover better General understanding of how the system works Going faster was better than being accurate Keeping breadboard clean and tidy (colour coding for wire) IR and Ping sensors drew lots of power Robot freaked out at very low power Constant power supply would have been useful Bring multimeters!! 2

3 What were common problems during the midterm competition? When trying to turn, robot went back and forth to the original position onthe tape and couldn t complete turn. Use feedback to know that the turn was completed. Running open loop isn t reliable. eg e.g. another tape sensor to know you ve turned 90deg Robot went the wrong way around can, got stuck Robot missed tape or went into walls (sensors didn t respond) Check sensor values during the motor loop! Chances are good that you might miss a sensor input. To avoid using motor loops entirely, use the ServoPal module (NB: the manual isn t easy to follow, might not be worth the effort to use the system). 3

4 Robot was often lost in the code (hard to tell which hpart of the code was running at any given moment). Create a status/state variable to keep track of the current state of the robot. Report the current status/state to the user while hl the robot us running Use red/green LEDs and audio beeps to indicate the current state or position in the code, or during transitions. We have access to high speed camera for filming the bot, so you can potentially see lights flashing very quickly. Use the Debug Terminal to log data back to the PC through the USB port (see the QTI Follower example code). You can print out ANY VARIABLE to the Debug Terminal usingsimilarcommandssimilar commands. However, usingdebug might be SLOW, so only print out variables periodically Use the Bluetooth module to log data back wirelessly to the PC through the Debug Terminal, as described here: php?t=

5 Make your sensors rigid. Use wire ties / zap straps to hold wires in place Custom brackets from sheet metal and plastic Jon will bring material and hand tools to class Use lock washers to hold screws in place 5

6 Tools for organizing software Flowcharts and State Diagrams Diagrams are useful tools for planning out software and order of events, and to keep organized. Design your entire system using a diagram and keep your diagrams up to date! (make changes on the diagram before doing any coding) Flowcharts show you step by step operations and key decision making points, useful lfor State Diagrams and State Variables for knowing exactly where you are in your code. 6

7 7

8 8

9 9

10 Proposed Final Competition Home A B C 10

11 A B H Rules (very preliminary...) Will be discussed during next 2 C lab periods Objective: Pickup and deliver the highest value of cans back to Home. Teams of 4 robots (8 students). 1 robots may be tele operated operated. Competition is 8 min long. Cans initially placed at each site A, B, and C, each with different values (e.g. A = 5, B = 8, C = 9) Teams will pick up and deliver cans to Home Every 30 seconds, 1 new can introduced to the table at only 1 location. The location of the can (at A, B, or C) will be scheduled by the team before the run begins. If there is still a can already at the site, no new can will be placed onto the surface at that time. Walls will be on the surface as well. Would like to have more walls, 11 so work in midterm is still usable here.

12 H A B Design Decisions i for Teams C Should the tele operated robot be a stationary controller/pit boss, or be another moving bot? What is the simplest way to coordinate each robot? Make each independent or centrally coordinated? How to send signals to the autonomous bots? (colours/lines, sounds?) How to optimize can pickup schedule for scoring? How to incorporate robot avoidance with gripper? 12

13 Next 2 lab sessions: a.) Work with gripper arms and see how easy/difficult it is to incorporate whiskers or IR sensors b.) Determine final rules for Final Competition I 13

14 Marking Scheme 75% of mark on performance 25% of mark on peer evaluation 20% of final course mark 14

15 Groups Group 1 The Big Easy (Elan, Andrew) Go Muck Yourself(Chris, Greg M) OPRAH (Niko, Carter) Merlin (Aaron, Kyle) Group 4 CuBe (Billy, Amy) Liquid id Courage (Jason, Graeme) Sensory Overload (Choong Eun, Jati ) V.M (Ahsan, James ) Group 2 Optimus Crime (Jake, Jonathan) SHEILA Glenn, Peter YOU MUST NOT BE NAMED Greg N, Liz D.E.N.N.I.S Rory, Jesse Group 5 Big Bird (Brad S, Jaylem ) Dilbert (Roben,Trevor,Steven ) BOB (Jetzen, Brad L ) Neil Armstrong (Devon, Ben ) Group 3 Clyde (Vanessa, Gajan ) AB RIPPER X (Kurtis, Adam ) PRINCESS (Paul, Vincent ) Cookie Monster (Olga, Mohsen) 15

16 Proposed Lab Schedule Week# Dates Wed Lab Other Gripper Kit hand out 8 Oct. 26 / 27 Meet with ih8 person team 9 Nov 2 / 3 Finalize Rules for Competition Midterm Reports Due Tues (20% of grade) 10 Nov 9 / 10 Work Day 11 Nov. 16 / 17 Work Day 12 Nov. 23 / 24 Final Competition 13 Dec 2 / 3 Last week of classes (no lab) Hand in for assignments.... Dec 16 Final Exam, 830 8:30am 16