4. Evaluation4.1. Overview of Testing RequirementsThe tests described in this section ensure that the robot meets every design constraint. The tests performed on the hardware and software portions of the robot guarantee that each performs satisfactorily. Table 4.1 below shows the division of design constraints and provides an overview of how each was tested. Table 4.1 Testing Set summary categorized by Design ConstraintsDesign Constraint Test Set DescriptionDimensions The completely assembled robot was measured to ensure its dimensions do not exceed contest restrictions. Navigation The robot’s ability to navigate the playing field correctly was tested. This required testing the following hardware components: stepper motors, line sensors, and sonar sensors. This testing also required the certification of the navigational software used for maneuvering, range finding, and motor control.Speed The robot’s ability to meet speed requirements was tested by performing stepper motor testing and navigational software testing. Package Identification and ManipulationThe robot’s ability to accurately identify, collect, and store all 12 packages was tested. This required testing the following hardware components: servos, stepper motors, storage chutes, turntable, barcode scanner, and gripper. This testing also required the certification of the collection softwareused for servo and motor control, barcode reading, PIC communication, andpackage extraction.Weight The completely assembled and loaded robot was weighted to ensure it met the 20-pound weight limit. Power Each power supply was tested for correct voltages and currents. Each supply will also be tested for adequate battery life. The printed circuit board (PCB) was also test for correct power trace routes. 4.2 Detailed Description of Each Subsystem Test4.2.1 Software4.2.1.1 Barcode ScannerThe packages used during the competition are labeled with a barcode to indicate the ultimatedestination, Plane A, B, or C, of each package. Software must be able to enable and disablepower to the scanner and correctly read the data bytes of a barcode to determine the proper actionof the loading and unloading mechanism. The objective of this test is to verify that softwareenables and disables power to the scanner, recognizes each barcode appropriately as eitherPackage A, Package B, Package C, or Unrecognized, and in the event that no barcode is present,exits the function. 4.2.1.1.1 Barcode Scanner Test ProcedureThe primary microprocessor was programmed to utilize the Scan_Barcode() function 12 times,because there are 12 packages in total, and to transmit the results of each scan serially as shownin Table 4.2.1.1.1.Table 4.2.1.1.1 Serial Communication for Each Return Integer of Scan_Barcode():ReturnIntegerSerial Communication1 Package A2 Package B3 Package C4 Unrecognized Scan5 Invalid Barcode6 No Barcode Present< 1 or > 6 ErrorIn the event that an incoming data byte is corrupted a software reset will trigger and either “Error:Start bit is 1” or “Error: Stop bit is 0” will be transmitted serially. To monitor the results, anRS232 interface was used to connect serially from the robot’s primary microprocessor to acomputer; the program HyperTerminal was used to monitor the serial transmissions. During thetest, the results were monitored for when all three known barcodes were scanned, for when anunknown barcode is scanned, and for the event in which no barcode is present. 4.2.1.1.2 Barcode Scanner Test CertificationInitial software tests, as outlined in Section 4.1.1.1.1, were unsuccessful. The error message“Error: Start bit is 1” was continuously transmitted. Additional coding was added to instruct theprogram to serially transmit any and all data bytes received. The additional coding revealed thatthe first frame of data bytes had been received and that the error was occurring as the secondframe of bytes was being received. This indicated that the problem could be a timing issue withthe software. Perhaps the program was spending too much time reading the first frame of bytesand missing all or part of the second frame. To validate this theory, an oscilloscope was used todetermine the time delay between frames. However, not only was it discovered that there issufficient delay between each frame for the microprocessor to receive all the data, but also thatthe problem went away. Because the oscilloscope provides some capacitance to settle noise, itwas determined that the error condition was occurring due to some noise on the clock line inbetween data frames. To solve this problem, a small capacitance, on the order of 2.2 nF, wasadded to the clock line of the scanner interface, as shown in Figure 4.2.1.1.2, to settle any noisepresent.Figure 4.2.1.1.2: Wiring Diagram of Scanner Interface with Capacitor4.2.1.2 NavigationUnder the software category of navigation, there are two major features with regard to stepcalculation that need to be addressed and tested: traveling distance and angle turning. Tocalculate the number of steps necessary for each motor to turn, it is imperative to know the radiusof the wheels being used and the number of steps per revolution of the motor.4.2.1.2.1 Traveling DistanceIn order to test the accuracy of the step calculations, a program titled StepTest.c was created.This program accepts a distance in inches from the user and then calculates and outputs thenumber of steps needed for each motor to travel the given distance. 4.2.1.2.1.1 Distance Test ProcedureA program, StepCount.c, was written that calculated the number of steps that each stepper motormust travel to cover a given distance. The PIC4620 was then programmed with StepCount.c.Using the HyperTerminal application, a distance was entered in inches, and the number of stepsreturned by the PIC4620 was noted. This number was then compared to the steps calculated byhand. With this data, the accuracy of the program was determined.4.2.1.2.1.2 Distance Test CertificationThe accuracy of the program’s step calculation for distance is near 100%. Since the motorscannot move a fraction of a step, the step count is calculated as an integer. As a result, anyfloating point ending is truncated from the final result. This is why the accuracy cannot reachabsolute 100%. The data from Figure 4.2.1.2.1.2.b shows, however, that the step count is alwaysat least 98% accurate. Figure
View Full Document
Unlocking...