Distance Measurement. Figure 1: Internals of an IR electro-optical distance sensor

Transcription

1 Distance Measurement The Sharp GP2D12 Infrared Distance Sensor is an electro-optical device that emits an infrared (IR) beam from an LED and has a position sensitive detector (PSD) that receives reflected IR; see Figure 1. A lens is positioned in front of the PSD in order to focus the reflection before it reaches the sensor. The PSD sensor is an array of IR detectors. The distance of an object can be determined through optical triangulation, as depicted in Figure 2. The location of the focused reflection on the PSD is translated to a voltage that corresponds with the measured distance. The IR distance sensor is designed to measure distances from 9-80+cm. Figure 1: Internals of an IR electro-optical distance sensor 1

2 Figure 2: Top: Optical triangulation gauging distance of a near object. Bottom: Optical triangulation gauging range of a more distant object. The datasheet for the GP2D12 IR distance sensor on page 8 has a response curve with voltage on the y-axis and distance on the x-axis. The output of the device is a continuous voltage ranging from 0-2.6V. In order for the cybot to have range finding abilities, an A/D conversion of the sensors' voltage output must be performed, followed by a translation of the quantized (digital) value to distance. As you can see, the sensor's range finding output curve is non-linear with distance. This means that translating the values to distances is not as simple as using the slope of a line for the response curve (as in the lecture slides). There are several methods to address this programmatically. A couple examples are: Find an equation that approximates the curve and use this equation in your program to translate quantization (digital) values to distance. Complicated mathematical computations (e.g., non-integer, non-algebraic functions) are a challenge for an MCU. Depending on design criteria, the computational cost may not be acceptable. Create a static lookup table consisting of quantized values and corresponding distances. A simple table scheme for 12-bit quantization has 4096 rows. This application is a good example of what is called a time/memory tradeoff. One method takes more time, another method takes more memory, and your selection of a method may depend on design constraints and criteria. 2

3 Note also that the distance range from 0-9 cm will cause the sensor to produce voltage values consistent with distances farther away. Ideally the sensor is never applied in situations where the distances are that short. In our mobile environment, this ideal is not practical, so we will need to address this issue in future labs. TM4C123 ADC Basics Here are a few ADC facts for you to keep in mind: 1. The voltage on ADC0 will decrease as the distance increases. 2. The A/D quantized (digital) value will increase with voltage. The maximum digital value represents a voltage at VREF, and the lowest value represents a voltage at GND. 3. The number of quantization levels depends on the number of bits of resolution used in the A/D process. 4. There is a tradeoff between A/D sampling speed and resolution. How fast a sample is taken depends on the ADC conversion time, which depends on the number of bits of resolution. 5. The VREF we need is 2.56V, which is internally provided. This is because the voltage from the sensor will range from 0 to 2.6V. One could choose higher VREF values, but with a fixed number of quantization values, one sacrifices resolution in order to support a higher voltage range. 6. When configuring A/D conversion, keep in mind that register settings are necessary to enable A/D conversion and initiate conversion. 7. How do you know the conversion result is ready for use? There are two methods: 1) poll the Raw Interrupt Status register, or 2) enable an interrupt handler. Using an interrupt in this lab is optional; no extra points will be given. If you want to try using an interrupt, refer to textbook Table 5.10 (vectors and NVIC definitions) to find the name of the ADC interrupt handler (ISR name). Handler names can also be found in the TM4C123.s startup file. 3

4 IR Sensor Connection Make sure the IR sensor is connected to the correct pins on the cybot board. See the picture. The A/D pins of the TM4C123G are accessible on the peripherals set of pins on the cybot board. The pins labeled SNSR1 on the cybot board are connected to POWER, GROUND, and a connection to PB4 (AIN10). The pins are marked with 3 rows: signal, 5V (power), and ground. You need to orient the sensor connector so the white wire connects to signal and black connects to ground. The voltage supply wire is in the center position (5V Row). Part 1. Initial Distance Measurement Requirement - Write a program that will sample the IR distance sensor's analog output (voltage between 0-2.6V) with the ADC, compute the distance, and display both the ADC quantization value and the distance value (in centimeters). As noted above, due to the nonlinear sensor response, you should implement a tabular technique to map quantization values to distance values. Accurately display a distance value for an object 9-50 cm away. The output format should be the same for Part 1 and Part 2. The format should be: Quantization Value, Distance cm The displayed quantization values and distance values will be dynamic based on values produced by the ADC. The comma and the cm abbreviation should be present on the display as values update in order to make the output more understandable. You will notice quite a bit of jitter in the quantized value even when the device and object are stationary. You may want to slow down how fast samples are taken as well as how frequently the display is updated. 4

5 Part 2. Calibrate Distance Measurement Requirements Reduce the jitter seen in Part 1 by averaging multiple samples using the hardware averaging (Tiva datasheet section ) or software averaging to collect and average 16 samples to get a more stable sensor value. Also experimentally calibrate your IR sensor to accurately read distances from 9 to 50 cm. Continue to display the output as: Quantization Value, Distance cm The effect of the jitter observed in Part 1 can be reduced by averaging multiple samples and treating the average value as the current distance value. Use an averaging mechanism in your program. Beyond the averaging mechanism, work on calibrating the distance measurement so that it is accurate to one centimeter from 9-50 cm. To complete this activity, use a ruler provided and a flat reflective surface to determine the quantization value for the distance between the object and the sensor. Note this mapping between quantization value and distance and move to the next distance. There are potentially a number of quantization values between centimeter gradations. More than one quantization value may have the same distance value assigned to it. As described in the lecture slides, there are a couple of approaches to empirically calibrate the distance: 1) measure 50 points and create a table of quantization values for each cm, or 2) measure 5 points and use Excel to get a trend line. You will need to provide accurate distance values for distances beyond 50 cm for use in future labs, but you are not required to support distances beyond 50 cm for this lab. Demo the program execution on the board to your TA. 5