Test Plan. Robot Soccer. ECEn Senior Project. Real Madrid. Daniel Gardner Warren Kemmerer Brandon Williams TJ Schramm Steven Deshazer

Transcription

1 Test Plan Robot Soccer ECEn Senior Project Real Madrid Daniel Gardner Warren Kemmerer Brandon Williams TJ Schramm Steven Deshazer

2 CONTENTS Introduction... 3 Skill Tests Determining Robot Position... Error! Bookmark not defined. Ball Tracking... 4 Spin in Circle... 5 Move in Rectangular Path... Error! Bookmark not defined. Goal Defense... 7 Conclusion

3 INTRODUCTION The following document contains the test created by the Real Madrid Robot Soccer Team in order to test critical areas of functionality for our soccer playing robot. We will utilize these tests in order to confirm that our robot functions according to the critical design specifications. Specifically, this test plan contains tests that will focus on the following areas: Determining the position of the robot Tracking the ball and determining its position Controlling circular rotational movement of the robot Controlling linear motion control through following a rectangular path Defending the goal from an oncoming ball DETERMINING ROBOT POSITION Being able to locate the robots position is very important playing soccer. From a robotic perspective, it demonstrates the success of image processing. Image processing first locates significant land marks using tracking software available as part of the OpenCV suite. In combination with relative object size, view angle, and distance feedback, it determines the location of fixed objects relative to the location of the robot. These fixed objects can then be used to calculate the robot s locations in the field coordinate frame. The Wi-Fi module on the robot can then broadcast a data packet containing this locational information. A base station listening to the Wi-Fi packets broadcast from the robot can be used to record the calculated coordinates. Through previously measuring the location of the robot using a tape measure, we can verify the accuracy of the calculated position. List of Equipment: Robot under test Standardized playing field Computer base station with Wi-Fi module Tape measure Test Description: 1. Robot will be placed on the field in a random location 2. Robot position will be hand measured using a tape measure 3. Robot will determine its relative position using image processing 4. Field position will be broadcast over Wi-Fi and verified for accuracy at computer base station. 5. Difference between estimated and actual position will be measured and recorded. 6. Process is repeated at a new location 3

4 Expected Results: We expect that using the above procedure we will be able to quantify the accuracy of our image processing in measuring fixed landmarks and determining robot position relative to these landmarks. Initial preliminary results should be able to determine the robot position within 2 inches of its measured location. The image processing can then be tuned to increase the accuracy. Upon tuning maximum deviation should be limited to about 0.5 inches. BALL TRACKING Being able to locate the robots position is very important playing soccer. From a robotic perspective, it demonstrates the success of image processing. Image processing first locates significant land marks using tracking software available as part of the OpenCV suite. Through using a combination of the relative ball size, bearing of ball center, and distance measurement feedback, we determine the location of the ball relative to the location of the robot. Through converting to the field s coordinate frame, we can verify the measured results using a tape measure. As in the test performed for calculating the robot s position, the Wi-Fi module on the robot can then broadcast a data packet containing this locational information. A base station listening to the Wi-Fi packets broadcast from the robot can be used to record the calculated coordinates. Through previously measuring the location of the robot using a tape measure, we can verify the accuracy of the calculated position. List of Equipment Robot under test Standardized playing field Computer base station with Wi-Fi module Tape measure Test Description 1. Robot will be placed on the field in a random location 2. Robot will determine its relative position, and then await for the ball to be placed within its field of view. 3. Ball will be statically placed on the field in a random location within the field of view of the robot 4. Actual ball position will be measured using the tape measure 5. Robot will determine the position of the ball 6. Ball position will be broadcast over Wi-Fi and verified for accuracy at computer base station 7. Difference between estimated and actual position will be measured and recorded 8. Process is repeated keeping the robot s location constant while repositioning the ball Expected Results We expect that using the above procedure we will be able to quantify the accuracy of our image processing in measuring the ball position relative to the robot s position. Initial preliminary results should be able to determine the robot position within 2 inches of its measured location. The image processing can then be tuned to increase the accuracy. Upon tuning maximum deviation should be limited to about 0.5 inches. We also anticipate the locational accuracy to decrease with increased distance between the robot and the ball. 4

5 CIRCULAR ROTATION In soccer, players need to get to the ball, take the ball to the opponent s goal, and shoot. Though if the ball go past a player, the player needs keep the ball in his sight. So the player could do two things: back pedal till player can see ball in field of vision, or spin around till ball is in sight. Problem with back peddling till robot sees the ball in front of it is that the robot could accidentally run into the ball and score on own goal. Solution to this is by having the robot spin till it sees the ball and then it could go toward the ball. With this solution there is restrictions on what the robot can do. Robot cannot move in x or y coordinate direction when spinning. Also the robot needs to finish spinning with the ball in the sight of the camera. An additional thing is that it needs to accomplish this test by spinning in both directions. The thought of spinning to know where the robot is in reference to the playing field was taken out of consideration due to the fact an implementation of a Kalman filter will replace the need of needing to spin to figure out where the robot is. List of Equipment Robot under test Standard playing field Computer ROS Wi-Fi Communication Measuring Tape Colored Stickers Yellow standard golf ball Test Description This test will allow us to observe any deviations from the desired behavior. Such as if the robot moves in an x and/or y direction then we can adjust our code to reduce that behavior. If robot spins beyond golf ball we can adjust problems/errors with vision sensing. 1. Place robot in standard playing field with minimum of 3 inches away from wall. 2. Place colored stickers on playing field under the perimeter of robot. 3. Place yellow standard golf ball out of sight of camera. 4. Turn on computer and start up ROS. 5. Start Wi-Fi Communication channel between robot and computer. 6. Start spin program on robot. 7. Observe Robot to see if robot stops with yellow golf ball in sight. 8. Read ROS communication to see if yellow golf ball is observed by camera. 9. Measure with measuring tape the distance robot is away from a given colored sticker. 10. Repeat procedure with spin program making robot spin opposite direction. 5

6 Expected Results We expect that the robot will have the ball slightly off centered from robots front. Also there may be little deviation in x or y coordinates from initial position. Even with a 5% error in these results the robot will still perform to optimal performance for a soccer game. MOVE IN RECTANGULAR PATH Traveling from a given location to a target destination is critical to implementing plays within a game of robot soccer. It requires the coordination of vision image processing with motion control. However, simply moving from one location to another is insufficient; rather, the robot must follow a predefined path. For the purposes of our test, we have chosen a rectangular path which the robot will be required to follow. List of Equipment Robot under test Standard playing field Computer Wi-Fi Communication Test Description The test is to supply the robot with an x and y coordinate on the field and then see if it can navigate to that point in a rectangular trajectory. The key thing we want to test is whether the robot can accomplish this without rotating. This will be a key test of the omni-directional capabilities of our tri-wheeled robot. The ability to move laterally in both directions without turning will make the robot able to move about the field much quicker. We will measure the robot s average speed over the rectangular distance. 1. Robot will be placed on field in random location 2. We will provide a coordinate on the field as a destination 3. The robot will move on two lines to form the rectangle to that coordinate 4. We will measure the time it takes to travel the distance Expected Results We would ideally like the robot to move laterally in one direction, and then laterally in the perpendicular direction, with no turning involved at all. We also expect it to be able to cover this distance in a similar time to what it would take if it were traveling in a straight line. 6

7 GOAL DEFENSE The ability of our robot to defend the goal is critical to our success as a team. If our robot leaves the goal completely unprotected, or is unable to track the ball as it approaches the goal, then the other team will score against us with ease. The purpose of this test then is to ensure that our defensive robot has, at the very least, the basic capabilities of a goalie in order to prevent embarrassing loss. List of Equipment Robot under test Standard playing field (with color markers in each corner) Regulation (Robot Soccer) yellow golf ball Test Description Our robot's ability to succeed as a goalie will easily be measured with the following test procedure: 1. Begin by placing the robot (with vision sensor) in a random position on the field. Measure the amount of time it takes the robot to identify its relative position on the field. 2. Measure the amount of time that it take for the robot to position itself 2 inches in front of its own goal 3. Place the yellow golf ball on the field in a random location between 10cm and 1.3m away from the robot. 4. Measure the amount of time it takes for the robot to orient itself to face the ball. 5. Roll the ball towards the goal at different speeds in order to test the robot's response time. There should be at least one fast, medium and slow speed. 6. Attempt each speed at least ten times. Record the number of times the robot is able to keep itself positioned between the ball and the goal until the ball is within 10cm of the robot. Expected Results It is expected that the robot will be able to identify its position on the field within 5 seconds of being placed in a random location. This means that after 5 seconds, it should have already started moving in the direction of its own goal. The amount of time that it takes to get within 2 inches of its own goal will vary depending on placement position on the field. However, it should be obvious that the robot is not wasting time on superfluous movement as it relocates. Also, the amount of time it takes to reposition itself should be relatively consistent when placed at relatively the same distance from its goal. It is also expected that the robot will be able to maintain its position between the ball and goal 100% of the time with slow rolls, 85% of the time with medium-speed rolls, and greater than 60% of the time with fast rolls. 7

8 CONCLUSION The tests discussed above will ensure that our robot is functioning within design specifications. Additionally, these tests will serve as guide in helping us to determine areas in which improvements can be made. By focusing our attention on critical areas of improvement, we will be able to enhance the performance of our robots, ultimately achieving better results within the soccer tournament. After having passed these tests, our robots will be ready for any advanced tests that we may create to maximize our success on the soccer field. 8