Increasing Robot Accuracy
For precise movement on a game table, there's a key principle to remember: The robot must always know its position. This means the robot needs to be aware of its current location on the table and where it needs to go next. This is crucial when the robot must reach a specific spot, such as to pick up an astronaut or drop off a pom.
This is easier said than done, especially since external reference points like GPS aren't allowed.
X-Y Approach
The X-Y approach offers a straightforward method to ensure the robot consistently knows its x and y coordinates when needed. This is achieved through various squaring-up techniques.
This graphic illustrates the X-Y approach, showing how the robot can square up with the table's edges to determine its position. This image is from jonathan Harrison's video, mentioned below.
Back-EMF
Back-EMF measures the ticks of the robot’s motors to estimate its position. However, this method can be inaccurate, as the measured ticks may deviate from the actual ticks, especially over longer distances. It works best for short movements.
The graph above illustrates the deviation when using the Back-EMF method, showing significant inaccuracy over longer distances. This data is from the paperr Enhancement of Accuracy in Botball Navigation, presented at GCER 2024.
Aligning with the Walls
This method involves aligning the robot by gently driving it against the PVC pipes along the game table's edges. This ensures that the robot's x or y position is accurate and can also correct its rotation.
Line Up
Similar to the wall alignment method, this approach uses line-follow sensors to align the robot with black lines on the table. This method ensures the robot knows its x or y position.
For more details, check out this video by Jonathan Harrison on writing a line-up method: Square Ups
Line Following
Line following is another straightforward way to navigate the game table. The robot follows a black line to track its position, which is especially useful when it needs to reach a specific spot.
Jonathan Harrison explains how to implement a line-following method in this video: Line Following
Driving Straight
Driving perfectly straight is critical for ensuring the robot can travel long distances accurately without deviation. Small inaccuracies can lead to significant errors when covering a large area, especially if the robot is relying solely on dead reckoning (estimating its position by measuring motor ticks as example). Here are a few strategies to help the robot maintain a straight path:
Back-EMF: The most common method for straight driving is to use Back-EMF. The ticks allow the robot to adjust its speed by comparing the ticks of the left and right motors. If one motor is slightly ahead, the robot can slow it down until the other catches up.
Gyroscope: A gyroscope sensor can detect angular deviations and correct them. If the robot starts to veer off-course, the gyroscope will sense this, and the control algorithm can adjust the motor speeds to bring the robot back in line. Read IMU Sensor for more information about gyroscope sensors.
Calibrating Motors: If the motors themselves have different strengths or frictions, it can cause the robot to veer. Regular calibration can help ensure both motors are equally powered, minimizing the chance of drifting.
Rotating Angles
Precise rotation is key to reorienting the robot when it needs to turn and drive toward a new direction. Here are common techniques for achieving accurate rotation:
Back-EMF: Similar to straight driving, Back-EMF can also be used for turning by counting the number of motor ticks. To rotate the robot, the motors on each side move in opposite directions, and by monitoring the encoder ticks, the robot can calculate when it has reached the desired angle. However, this method is less reliable over time, as slippage and friction can introduce errors.
Gyroscope: The most reliable way to achieve precise turns is by using a gyroscope sensor. The gyroscope tracks the robot’s rotation in real-time, and the control algorithm can stop the motors as soon as the robot has turned the correct number of degrees. Read IMU Sensor for more information about gyroscope sensors.
Botball Analysis
To see the X-Y approach in action, I recommend watching Botball Analysis by Jonathan Harrison. In this video series, Harrison analyzes various Botball runs, highlighting what was done well and what could be improved. It's a great resource for anyone looking to enhance their Botball strategies, and I've learned a lot from it.
Authors
Tobias Madlberger