Chapter 2 Development of a MATLAB-based Graphical User Interface Environment for PIC Microcontroller Projects 2.1. Introduction Peripheral Interface Controllers (PICs), developed and marketed by Microchip Technology, Inc. , are inexpensive microcontroller units that include a central processing unit and peripherals such as memory, timers, and input/output (I/O) functions on an integrated circuit (IC). There are more than 100 varieties of PIC microcontrollers available, each providing functionality for different types of applications , making PICs one of the most popular microcontrollers for educational, hobby, and industrial applications. Similar to other microcontrollers, PICs are usually not designed to interface with human beings; instead they are directly embedded into automated products/processes. Thus, graphical user interface (GUI) capabilities, which have become a mainstay of many personal computer (PC) applications, are nonexistent for PICs. Endowing PIC-based projects with GUI tools can speed the development process in data driven applications such as feedback control, smart sensors, etc. Microchip Technology’s emulator and debugger products (e.g., MPLAB IDE, MPLAB-ICE) are very helpful in debugging PIC source code and emulating user-written programs. However, these tools do not provide data co-processing and advanced data visualization capabilities. Fortunately, PIC microcontrollers include serial communication functionality to facilitate data communication with external devices such as analog to digital converters (A2D), 1-wire sensors, etc. Similarly, MATLAB, a commercially available interactive mathematical programming software, also provides serial data communication functionality on PCs. In addition, Simulink, MATLAB’s interactive icon-based programming environment, enables users to simulate and analyze dynamic system models. Finally, the Dials and Gauges Blockset of Simulink 7 8 allows users to embed control objects (e.g., sliders, knobs) and display objects (e.g., graphs, gauges) in Simulink models to develop an interactive GUI environment. In this dissertation, we exploit the serial communication functionality of MATLAB to enable a PC to communicate with PIC microcontrollers to transmit control commands and receive sensory data. In addition, we utilize MATLAB, Simulink, and Dials and Gauges Blockset to develop an interactive GUI environment for PIC projects, allowing enhanced data processing and visualization. In this dissertation, we use a PIC16F74, 40-pin, 8-bit CMOS FLASH dual inline package IC. To facilitate serial communication between PIC and PC, we interface a RS232 driver/receiver with the PIC16F74. The effectiveness of our MATLAB-based GUI environment to interact with PIC microcontroller projects is demonstrated by using three examples: i) export user commands from a Simulink GUI to an actuator interfaced to the PIC; ii) import signals from a sensor interfaced to the PIC into a Simulink GUI; and iii) use Simulink GUI to export user commands to the PIC and import sensory data from the PIC to control a device and monitor its status. 2.2. Hardware Environment The hardware environment for this dissertation consists of a PIC microcontroller, a PC, a RS232 driver/receiver, and a DB-9 serial cable. The PIC microcontroller is interfaced with external devices such as sensors (e.g., photoresistors) and actuators (e.g., servomotors). In addition, the PIC microcontroller performs embedded computing. The PC is used to write user specified embedded programs to be executed by the PIC microcontroller. Furthermore, the PC hosts an interactive GUI for the user to manipulate control variables and visualize sensory data. The PIC microcontroller and the PC communicate using a serial interface. A PIC development board (see section 2.2.4) and a light refraction experiment test bed (see section 2.2.5) are used to illustrate our PIC-based data acquisition and control approach. 2.2.1. Peripheral Interface Controller PIC microcontrollers are small, low-cost controllers that include a processor and a variety of peripherals. PICs are significantly easier to use vis-` a-vis embedded microprocessors. As an 9 example, users can assign desired functionality (e.g., A2D, universal synchronous/asynchronous receiver and transmitter (USART)) to I/O pins of PICs. PICs can be operated at various clock speeds (32 kHz to 20 MHz). PIC’s memory architecture separates its data memory from its program memory with the program memory available as One-Time Programmable (OTP), Erasable Programmable Read-Only Memory (EPROM), or FLASH. PICs are programmed in the PIC assembly language using a 35 single-word instruction set. See  for more details on hardware and software features of PIC microcontrollers. The user specified embedded PIC program is written on the PC and downloaded from the PC to the PIC microcontroller using the DB-9 serial cable connection between the PC and a PIC Development Programmer on which the PIC microcontroller is installed. Commonly available PIC Development Programmers include PICSTART Plus  from Microchip, Inc., and PIC-PG2B, a handy, low-cost programmer  from Olimex Ltd., among others. In this dissertation, we use the PICSTART Plus programmer that requires MPLAB Integrated Development Environment, a free software available on the Microchip website, for programming PICs. In this dissertation, we employ a PIC16F74, a 40-pin CMOS FLASH-based, 8-bit, midrange (14-bit instruction word length) microcontroller (see Figure 2.1). PIC16F74 has 4 Kbytes of FLASH program memory and 192 bytes of data memory. Furthermore, it has 33 digital I/O pins organized in 5 groups of I/O ports that can be assigned as 8-bit A2D, Capture/Compare/PWM (CCP), the 3-wire Serial Peripheral Interface (SPI), the 2-wire InterIntegrated Circuit (I2 C) bus, USART ports, etc. We use an external 20 MHz high-speed crystal oscillator to supply operating clock cycles. The PIC16F74 can be powered using a wide range of voltage sources, e.g., 2-volt direct current (VDC) to 5.5VDC, and each of its I/O pin can sink or source up to 25mA of current. It is ideal not only for laboratory data acquisition (the application considered in this dissertation), but also for automotive, industrial, and consumer applications. 10 (a) (b) Figure 2.1: (a) PIC16F74 (b) Pin diagram of PIC16F74 2.2.2. Personal Computer In this dissertation, an IBM-compatible Pentium 3 PC running Microsoft Windows NT 4.0 operating system is used. As previously mentioned, the PC is used to write, debug, and download embedded PIC programs. One of the serial ports on the PC is reserved for serial communication with the PIC microcontroller. MPLAB, MATLAB (version 6.1), Simulink, and Dials and Gauges Blockset are installed on the PC. Control variables are manipulated via the PC by interacting with control panels embedded in the Simulink program. In addition, all experimental data is collected and displayed on the PC in display panels embedded in the Simulink program. 2.2.3. RS232 Driver/Receiver MAX232 (see Figure 2.2) is a 2-channel, RS232 driver and receiver manufactured by Maxim Integrated Products, Inc. It requires a 5VDC power supply and converts voltage levels between PC-based logic and PIC microcontroller-based logic. Specifically, whereas the voltage levels of logic high and logic low for the PC correspond to -12VDC and 12VDC, respectively, like many other microcontrollers the logic high and low for the PICs correspond to 5VDC and 0VDC, respectively. The MAX232 is used with five 1µF capacitors to adjust the voltage level differences between the PC-based logic and the PIC-based logic. See  for more details of the MAX232 hardware features. 11 (a) (b) Figure 2.2: (a) MAX232 (b) Pin diagram of MAX232 2.2.4. PIC Development Board The PIC development board (see Figure 2.3) consists of a sensor (photoresistor), a 3pin header for a servomotor connection, a 20MHz crystal oscillator, a MAX232 with five 1µF capacitors, a PIC16F74 microcontroller, a breadboard, and two DB-9 connectors. The photoresistor sensor provides light intensity measurement and is interfaced to a pin allocated as an 8-bit A2D in port A of the PIC16F74 microcontroller. The circuit diagram of Figure 2.4 illustrates how various sensors and actuators of the light refraction experiment test bed (see section 2.2.5) are interfaced to the PIC microcontroller. The PIC transmits/receives sensory data to/from the PC via the MAX232. A red reset button is connected to the Master Clear (MCLR) pin of the microcontroller. Pentium class PC Reset button DB-9 serial cable PIC DB-9 connector to light refraction test bed Photoresistor 3-pin servomotor connector MAX232 (a) (b) Figure 2.3: (a) PC and PIC development board (b) Larger view of the PIC development board 12 5V 5V 10 kΩ 10 kΩ 1kΩ Reset button 1 2 Linear potentiometer 3 Limit switches 4 Photoresistor 15pF 20MHz 39 H-Bridge 38 Laser 34 Servomotor 32 31 11 0.1μF 26 25 14 1μF 1μF 12 13 15pF H-Bridge 5V 5V 0.1μF 40 RX TX 5V 1μF 1 3 1μF 16 2 6 1μF 4 5 12 13 RS232 output 11 14 RS232 input 15 MAX232 PIC16F74 Figure 2.4: Circuit diagram of the PIC development board 2.2.5. Light Refraction Test Bed The light refraction test bed (see Figure 2.5) is a mechatronics-aided physics experiment developed under a National Science Foundation (NSF) sponsored Science and Mechatronics Aided Research for Teachers (SMART) program  at Polytechnic Institute of New York University. This experiment is designed to demonstrate the law of light refraction. It consists of a light source, a light sensor, a linear potentiometer, two limit switches, a servomotor, a DC motor, a liquid reservoir, and necessary circuitry. A liquid reservoir on the top of the test bed can store various liquid media whose index of refraction needs to be determined. For simplicity, in this dissertation, we use water from a water fountain as the test liquid. Servomotor Light source Reservoir H-Bridge Light sensor DC motor Limit switches (a) Linear potentiometer (b) Figure 2.5: (a) Light refraction experiment test bed (b) Light source mounted on the servomotor 13 On one side of the tank, as shown in Figure 2.5 (b), a laser pointer, used as the light source, is mounted on the arm of the servomotor that sets the angular position of the light source to the incidence angle specified by the user. On the other side of the tank, a general Cadmium Sulfide (CdS) photoresistor, used as the light sensor, is mounted on the wiper of the linear potentiometer. It monitors the refracted light coming out from the liquid reservoir (see Figure 2.6). A DC motor drives the light sensor along the linear potentiometer by turning a motor shaft connected to a brass screw rod thereby transforming rotary motion into linear motion. Limit switches at each end of the linear potentiometer indicate sensor travel limit. The photoresistor and the linear potentiometer output analog voltage signals between 0VDC and 5VDC. Figure 2.6: Detailed view of the light sensor traveling along the linear potentiometer 2.3. Software Environment The software environment for this dissertation consists of the PIC assembly language, MATLAB, Simulink, and Dials and Gauges Blockset. The PIC assembly language is a primitive programming language consisting of a 35 single-word instruction set. MATLAB is an interactive technical computing software. Simulink is MATLAB’s model-based, system-level, visual programming environment that is widely used to simulate and analyze dynamic system models using icon-based tools. Finally, the Dials and Gauges Blockset of Simulink provides an ability to embed visual, realistic-looking, virtual instrumentations in Simulink models. In this dissertation, these software tools are judiciously synthesized to produce an effective, interactive GUI environment. In the sequel, we summarize key instructions of the PIC assembly language 14 and MATLAB that enable serial communication between PIC microcontroller and MATLAB GUI running on the PC. 2.3.1. PIC Assembly Program As indicated above, the PIC assembly language consists of a 35 single-word instruction set (see datasheets  for details). The PIC data memory is partitioned into several banks (e.g., 5 banks for PIC16F74) that contain the general-purpose registers and the special-function registers. The special-function registers are used to set up special operations (e.g., A2D, USART, and PWM) and to watch the status of the special operations (e.g., the availability of transmission or reception of the USART). Below, we review key PIC instructions and special function registers used for serial communication functionality. 1) Key PIC Instructions BCF: Bit clear f Syntax: [label] BCF f, b BCF literally means that the bth bit in the register ‘f’ is cleared. BCF sets the bth bit in the register ‘f’ to zero, logic low. BSF: Bit set f Syntax: [label] BSF f, b BSF instruction does the opposite of BCF, i.e., it sets the bth bit in the register ‘f’ to one, logic high. MOVLW: Move literal to w Syntax: [label] MOVLW k The literal ‘k’ is loaded into the working register. The literal ‘k’ can be expressed in terms of an 8-bit binary, decimal, or hexadecimal number. For example, b’00101111’ in 8-bit binary is equivalent to 0x2F in hexadecimal. Note that the prefixes b, 0x, and d declare the data type to be binary, hexadecimal, and decimal, respectively. MOVWF: Move w to f Syntax: [label] MOVWF f 15 MOVWF transfers data from the working register to the specified register ‘f.’ Since the literal ‘k’ cannot be directly assigned into the specified register ‘f,’ the literal ‘k’ is first assigned to the working register (e.g., MOVLW k) and then moved into the register ‘f’ (e.g., MOVWF f). BTFSS: Bit test f, skip if set Syntax: [label] BTFSS f, b BTFSS checks the bth bit in the specified register ‘f,’ and executes the next instruction if this bit is zero. Alternatively, if the bit is one, the next instruction is skipped, and the following instruction is executed. 2) Special Function Registers Used for Serial Communication Functionality MOVLW d’value’ MOVWF SPBRG The special function register ‘SPBRG’ contains the user-specified baud rate for serial communication. In particular, the command MOVLW d’129’ places 129 in the working register. Next, the command MOVWF SPBRG moves the content of the working register to the special function register ‘SPBRG.’ The placement of ‘value’ 129 in the ‘SPBRG’ register sets the baud rate to 9,600. MOVLW b’clock source select bit, 9-bit transmit enable bit, transmit enable bit, usart mode select bit, unimplemented, high baud rate select bit, transmit shift register status bit, 9th bit of transmit data’ MOVWF TXSTA The special function register ‘TXSTA’ contains information for the data-transmit status and control in an 8-bit binary expression. In particular, the use of commands MOVLW b’00100100’ and MOVWF TXSTA, sets up the ‘TXSTA’ register to enable 8-bit, high speed asynchronous serial data transmission. MOVLW b’serial port enable bit, 9-bit receive enable bit, single receive enable bit, continuous receive enable bit, unimplemented, framing error bit, overrun error bit, 9th bit of received data’ MOVWF RCSTA The special function register ‘RCSTA’ contains information for the data-receive status and con- 16 trol in an 8-bit binary expression. In particular, the use of commands MOVLW b’10010000’ and MOVWF RCSTA, sets up the ‘RCSTA’ register to enable 8-bit, continuous asynchronous serial data reception. 2.3.2. MATLAB Program MATLAB is a commercially available, widely used, interactive, technical computing software. MATLAB’s versions 6.1 and higher provide serial communication functionality. To serially communicate with an external device from MATLAB, the following steps need to be performed. First, create a serial port object to identify the specific serial port of the PC connected to the external device. In addition, specify how this serial port is to be configured (i.e., baud rate, number of data bits, etc.). Second, connect the serial port object created above to the external device. Third, send command signals to the external device and receive data from the external device. Fourth, disconnect serial communication connection from the external device and close the serial port object. Finally, release control of the serial port. Next, we list the key MATLAB instructions used for serial communication. See  for further details. serial (the PC serial port, the baud rate, the number of data bits) This command is used to create a new serial port object. In addition, it configures the serial port properties. In this dissertation, we used the COM2 serial port of the PC with 9,600 baud rate. fopen (object) This command opens the serial port object just created and connects the PC to the external device for actual serial communication. fread/fwrite (object, size, precision) The fread command enables the PC to read binary data from the external device. Alternatively, the fwrite command enables the PC to send control data in binary format to the external device. fclose (object) This command closes the serial port object, thereby disconnecting serial communication be- 17 tween MATLAB and the external device. freeserial (port) Once MATLAB establishes a data link with the serial port, it assumes complete control of the serial port. The freeserial command is used, after closing the port object using the fclose command, to force MATLAB to relinquish control of the serial port. The command takes on one argument, the port that was used for data communication. This command is executed from the MATLAB command line after the termination of experiment. 2.3.3. Simulink Simulink is MATLAB’s interactive, icon-based programming environment . It enables users to build block diagrams to simulate and analyze dynamic system models. Designers can effortlessly transfer paper designs of dynamic systems into Simulink block diagrams. Simulink block diagrams can be modified as easily as paper models of dynamic systems. In addition, Simulink allows for detailed monitoring of dynamic system outputs at any point in the block diagram using various tools (e.g., Scope, Display, etc.). Finally, data processing tasks such as signal scaling, filtering, etc., can be easily performed in Simulink. 2.3.4. Dials and Gauges Blockset The Dials and Gauges Blockset  provides enriched views of graphical, 3-D instruments called virtual instruments. It has various templates that can be customized to create realistic virtual instruments for electrical, aerospace, automotive, medical, and process control systems. The virtual instruments created using the Dials and Gauges Blockset dynamically interact with MATLAB and Simulink, thus providing an interactive interface for users to enter command inputs and visualize sensory outputs. 2.4. 2.4.1. Examples of Serial Communication between PIC and PC Serial Communication from PC to PIC: Servomotor Position Control This example illustrates one-directional serial communication from the PC to the PIC microcontroller. In particular, it demonstrates that the user commands from a Simulink block 18 diagram can be exported to an actuator interfaced to the PIC microcontroller. The example focuses on servomotor position control. The Simulink block diagram for this example is shown in Figure 2.7. It consists of a dial, from the Dials and Gauges Blockset, denoted as the servo angle knob. The user interacts with the dial to enter servomotor position control command. The dial has a range from 0 to 90 degrees with one-degree resolution. The value of the angle commanded by the user is shown in the middle of the knob. The MATLAB m-function block next to the knob contains a MATLAB m-file to perform serial communication from the PC to the PIC. The user specified servomotor position control command is transmitted to the PIC via a serial cable connection between the PC and the PIC. When the PIC receives the command angle, it assigns the angle to a variable in the PIC code. Next, the PIC utilizes the command angle to compute, generate, and apply pulse trains for servomotor position control. In this example, we used a 6VDC standard servomotor that is interfaced to the 3-pin servomotor connection header on the PIC development board (see Figure 2.3). The PIC assembly code corresponding to this example is available in Appendix A. Table 2.1: M-function for PC to PIC serial communication %MATLAB function serial out.m for serial communication from PC to PIC function serial out(angle) %serial out function defined ser obj=serial(‘COM2’,‘baudrate’,9600); %create and configure a serial port object fopen(ser obj); %connect the serial port object to the device ServoCommand=round(angle+107.3); %input for servomotor where 107.3 refers to offset fwrite(ser obj,[ServoCommand],‘async’); %send user command, i.e., dial input, to the PIC pause(1); fclose(ser obj); %disconnect the serial port object from the device 2.4.2. Serial Communication from PIC to PC: Data Acquisition, Processing, and Plotting This example illustrates one-directional serial communication from the PIC microcontroller to the PC. In particular, it demonstrates that a Simulink block diagram can be designed to acquire measurement from a sensor that is interfaced to the PIC. The example focuses on 19 Figure 2.7: Simulink block diagram for PC to PIC serial communication acquiring measurements from a photoresistor that senses light intensity. Referring to Figure 2.4, a light sensor is constructed by connecting a 10 KΩ resistor and a photoresistor in a voltage divider circuit. The output of the light sensor varies depending on the light intensity incident upon the photoresistor; here the light sensor output refers to the voltage at the junction of the 10KΩ resistor and the photoresistor. This output is connected to I/O pin 2 of the PIC16F74. The I/O pin 2 is configured as an A2D in the PIC assembly code. Each time, the PIC assembly code tasks the PIC to measure the light sensor output, the PIC16F74 converts the analog voltage signal at the voltage divider output into a corresponding 8-bit digital value. Thus, when the photoresistor is placed in dark condition, the 8-bit A2D returns a value close to 255. Alternatively, when the photoresistor is exposed to bright light, the A2D returns a value close to 0. The Simulink block diagram for this example is shown in Figure 2.8, where a MATLAB mfunction is used to acquire the digitized output of the sensor using serial communication. The Simulink block diagram of Figure 2.8 also processes and plots the sensory data. In particular, the top scope in Figure 2.8 plots the light intensity measurement (in terms of digitized output of the voltage divider circuit) versus time, where the measurements are filtered using a low-pass filter. The middle scope plots the unfiltered light intensity measurement. Finally, the bottom 20 scope plots the light intensity measurement in terms of voltages by processing the 8-bit digital value of the voltage divider circuit through a gain factor. Table 2.2: M-function for PIC to PC serial communication %MATLAB function serial in.m for serial communication from PIC to PC function v=serial in(dmyin) %serial in function defined ser obj=serial(‘COM2’,‘baudrate’,9600);%create and configure a serial port object ser obj.ReadAsyncMode = ‘manual’; %specify an asynchronous read operation fopen(ser obj); %connect the serial port object to the device LightSensOut=fread(ser obj,1,‘uint8’); %read the light sensor output fclose(ser obj); %disconnect the serial port object from the device v=LightSensOut; %8-bit representation of the light sensor output Scopes Figure 2.8: Simulink block diagram for PIC to PC serial communication An experiment was conducted in which the light intensity was abruptly altered at several time instances. The response plots acquired and processed using the Simulink block diagram of Figure 2.8 are shown in Figure 2.9. The filtered output response in Figure 2.9(b) is much smoother than the unfiltered response in Figure 2.9(a). Thus, Figure 2.9 demonstrates the efficacy of signal co-processing using MATLAB and Simulink for PIC-based projects. The PIC assembly code corresponding to this example is available in Appendix B. Light sensor O/P (8-bit) Light sensor O/P (8-bit) 21 Time (Sec) (a) Time (Sec) (b) Figure 2.9: (a) Unfiltered plot of A2D and (b) Filtered plot of A2D 2.4.3. Bi-directional Serial Communication between PIC and PC In this example, the light refraction test bed is used to demonstrate the advantage of exploiting bi-directional serial communication between PIC and MATLAB-based GUI executing on the PC. A Simulink-based interactive GUI for the light refraction test bed is shown in Figure 2.10. The user interacts with the dial object to command the angle of incidence of light source. The MATLAB m-function block next to the knob contains a MATLAB m-file that transmits the user command input to the PIC serially. The PIC stores the user input in a variable and uses it to compute, generate, and apply pulse trains to control servomotor position. This positions the light source, mounted on the servomotor arm, at the commanded angle of incidence. Next, the PIC turns on the light source and performs the following tasks: drive the light sensor along the linear potentiometer by turning the DC motor, measure the position of the light sensor along the linear potentiometer and the corresponding output of the light sensor, and transmit the position and light sensor measurements to the PC. The MATLAB m-function block shown in Figure 2.10 enables receipt of the position and light sensor measurements from the PIC serially. Simulink blocks following the m-file function block are used for various data processing tasks, e.g., conversion of position measurement to the refraction angle and computation of index of light refraction. Finally, a generic numeric LED display, from the Dials and Gauges Blockset, is used to indicate the calculated value of index of light 22 refraction for the experimental liquid. Table 2.3: M-function for bi-directional serial communication between PIC and PC %MATLAB function serial inout.m for bi-directional serial communication bet. PIC & PC %serial inout function defined function V=serial inout(angle) ser obj=serial(‘COM2’,‘baudrate’,9600);%create and configure a serial port object ser obj.ReadAsyncMode = ‘manual’; %specify an asynchronous read operation fopen(ser obj); %connect the serial port object to the device ServoCommand =round(angle+107.3); %input for the servomotor with offset of 107.3 fwrite(ser obj,[ServoCommand],‘async’); %send user command, i.e., dial input, to the PIC LightSensOut =fread(ser obj,1,‘uint8’); %read the light sensor output from the PIC Position=fread(ser obj,1,‘uint8’)+9; %read the linear pot. output from the PIC fclose(ser obj); %disconnect the serial port object from the device V=[LightSensOut;Position]; %output in matrix form Figure 2.10: Simulink block diagram for bi-directional serial communication between PIC and PC Figure 2.11 shows the block diagram of index of light refraction subsystem of Figure 2.10. The block diagram of Figure 2.11 is used to generate a plot of angle of refraction versus the light sensor output. Figure 2.12 shows the plots of angle of refraction versus the light sensor output for two commanded values of incidence angle, namely, 40◦ and 20◦. Note that for each incidence angle, the index of refraction is computed from the angle of refraction corresponding to the smallest output returned by the light sensor. Thus, the block diagram of Figure 2.11 is 23 also used to calculate the index of light refraction. The MATLAB m-function in this subsystem monitors and captures the angle data corresponding to the smallest measurement returned by the light sensor. Note that the light sensor output is smallest when the intensity of refracted light focused on the light sensor is highest. Next, the angle data is used to compute the index of light refraction. The PIC assembly code corresponding to this example is available in Appendix C. Table 2.4: M-function for calculating the index of light refraction %MATLAB function PickMin.m for capturing the angle of refraction when max. %Light is on the light sensor. function y=PickMin(minangle,minLSO,LSO,angle) %PickMin function defined if LSO <= minLSO %loop for updating the minangle minangle=angle; %update minangle wrt minLSO end y=minangle; %angle at the most light intensity XY Graph block Figure 2.11: Simulink block diagram for calculating the index of light refraction 2.5. Conclusion In this dissertation, we developed and presented MATLAB-based GUIs for PIC microcontroller projects by exploiting Simulink, Dials and Gauges Blockset, and serial communication capabilities of MATLAB and PIC. Three examples were presented to illustrate the productivity enhancement potential of the MATLAB-based GUI environment when developing PIC 24 (a) (b) Figure 2.12: Angle of refraction vs. light sensor output for incidence angle (a) 40◦ and (b) 20◦ microcontroller projects. The GUIs designed using framework of this dissertation allow the user to: i) vary control commands; ii) acquire sensory data; iii) perform on-line data processing; and iv) visualize data using realistic looking virtual instruments. Note that the framework of this dissertation allows the use of microcontroller as a low-cost, stand-alone data acquisition and control board (DACB). Whereas PC-based DACBs typically cost several hundred to over thousand dollars, a PIC microcontroller costs only a few dollars. Thus, the use of PIC microcontrollers with the proposed MATLAB-based GUI environment provides a low-cost DACB solution that can be particularly beneficial to educators. Chapter 3 Development of a MATLAB Data Acquisition and Control Toolbox for PIC Microcontrollers 3.1. Introduction Data acquisition and control boards are essential for interfacing sensors/actuators with decision making devices such as a PC. Thus, data acquisition and control boards are used in monitoring/instrumentation applications involving machinery, process, environment, etc., and in automatic control applications. Even though a variety of data acquisition and control boards have become widely available in the last 15 years, the systems that target the educational sector and provide support for icon-based programming environments, such as LabVIEW  and Simulink , tend to be quite expensive (over $500 to several thousand dollars). Moreover, instructional labs generally may not require the intrinsic high-performance features of many of the commercially available data acquisition and control boards (e.g., high sampling rates, high resolution analog to digital converters, etc.) for the typical electro-mechanical laboratory experiments. This dissertation proposes a microcontroller-based data acquisition and control system that is particularly suitable for educators interested in developing lab experiments that do not require high-cost, high-performance data acquisition hardware and yet can benefit from the icon-based programming environment of Simulink. Several recent papers have focused on interfacing low-cost microcontrollers (such as BASIC Stamp 2 (BS2) and PIC) with icon-based programming environments such as LabVIEW and Simulink. Specifically, Refs. [45–47] concentrated primarily on endowing microcontrollers with graphical user interface (GUI) capability by exploiting the GUI tools of LabVIEW and Simulink. However, the methodology of Refs. [45–47] requires manually programming the microcontroller for all sensing, control computation, and actuation tasks and for serial communication with the GUI running on the PC. To program a PIC microcontroller or a BS2 microcontroller using PIC assembly programming language or PBASIC programming language, 25 26 respectively, requires knowledge and experience with the syntax of these languages and is often tedious. This dissertation proposes a PIC microcontroller-based low-cost data acquisition and control system that exploits MATLAB and Simulink as the key software components for implementing data acquisition and control algorithms using a block-diagram format. Specifically, the dissertation exploits a newly developed library of PIC functions for Simulink and the serial communication capability of both the PIC microcontroller and MATLAB to produce a seamless integration between them. The framework of this dissertation completely obviates the need to manually program the PIC microcontroller by creating a library of PIC microcontroller functions for Simulink. Specifically, the data acquisition and control toolbox of this dissertation facilitates: i) automatic generation of proper PIC assembly codes for a variety of sensors and actuators; ii) automatic programming of the PIC microcontroller; and iii) data communication between the PIC microcontroller and MATLAB. In an instructional laboratory, this approach allows instructors and students to focus on hardware-in-the-loop implementation, experimental validation, and industry-style rapid control prototyping. Finally, this dissertation is in the spirit of Ref. , which provided a MATLAB data acquisition and control toolbox for the BS2 microcontrollers. However, whereas the BS2 microcontroller costs over $45 and includes only digital input/output (I/O) functionality, thus requiring external analog to digital converters (A2Ds) to interface with analog sensors, the PIC16F74 microcontroller, used in this dissertation, costs under $5 and includes built-in A2D functionality. This dissertation is organized as follows. Section 3.2 describes the PIC microcontroller and the related development hardware. Section 3.3 describes the software environment used in this dissertation. Section 3.4 gives details concerning the software integration of Simulink with the PIC microcontroller. Section 3.5 illustrates the functionality and capability of the data acquisition and control hardware and software of this dissertation by performing position control of a DC motor. Finally, Section 3.6 provides some concluding remarks. 27 3.2. Hardware Environment The main components of the data acquisition and control hardware of this dissertation are a PIC microcontroller, a PIC-PG2C programmer, and a PIC development board. A DB9 serial cable is used to interface the programmer/development board to a PC which hosts the MATLAB data acquisition and control toolbox. Specifically, the DB-9 cable allows i) programming the PIC microcontroller from the PC and ii) data communication between the PIC and the PC. In this dissertation, an IBM-compatible Pentium 4 PC running Microsoft Windows XP operating system is used. See Figure 3.1 for a pictorial representation of the aforementioned hardware environment. PC PIC development board PIC microcontroller DB-9 serial cable PIC-PG2C programmer Figure 3.1: Hardware environment 3.2.1. Peripheral Interface Controller (PIC) PIC microcontrollers, developed, manufactured, and marketed by Microchip, Inc. , are small, inexpensive controllers that include a processor and a variety of peripherals such as memory, timers, and I/O functions on an integrated circuit (IC). PIC microcontrollers are widely popular among educational, hobby, and industrial users who can select from more than 100 varieties of PICs one that suits their application and functional needs. In contrast to many other microcontrollers, PICs are quite versatile since their I/O pins can be assigned desired functionality (e.g., A2D, universal synchronous/asynchronous receiver and transmitter (USART)) under program control. Moreover, using an appropriate crystal oscillator, PIC microcontrollers can be operated at clock speeds of 32 kHz—20 MHz. The PIC assembly lan- 28 guage, consisting of a 35 single-word instruction set, is used to program PIC microcontrollers. See Ref.  for more details on hardware and software features of PIC microcontrollers. The data acquisition and control platform of this dissertation uses a PIC16F74 , a 40pin CMOS FLASH-based, 8-bit, mid-range (14-bit instruction word length) microcontroller. Pertinent specifications of PIC16F74 include: 2—5.5 Volt direct current (VDC) voltage input; 25mA current sink/source capability at each I/O pin; 4 Kbytes of FLASH program memory; 192 bytes of data memory; and 33 digital I/O pins organized in 5 ports (A—E) of I/Os that can be assigned as 8-bit A2Ds, Capture/Compare/PWMs (CCPs), 3-wire Serial Peripheral Interfaces (SPIs), 2-wire Inter-Integrated Circuit (I2 C) buses, USART ports, etc. In this dissertation, five of the six I/O pins of port A and three I/O pins of port E are reserved for eight 8-bit A2Ds, eight I/O pins of port B are reserved for eight digital inputs, two of the eight I/O pins of port C are reserved for two PWM outputs, and eight I/O pins of port D are reserved for eight digital outputs. Finally, an external 20 MHz high-speed crystal oscillator is used to supply operating clock cycles to the PIC. 3.2.2. PIC-PG2C Programmer The user specified PIC program, which is created on a PC, is downloaded from the PC to a PIC microcontroller by serial communication. Serial communication between the PC and the PIC microcontroller is enabled by using a DB-9 serial connection between the PC and a PIC development programmer that hosts the PIC microcontroller. Two widely used PIC development programmers are Microchip’s PICSTART Plus and Olimex’s PIC-PG2C . In this dissertation, the handy and low-cost PIC-PG2C programmer (see Figure 3.2) is used. In contrast to other PIC programmers, the PIC-PG2C programmer receives power from the PC’s serial port thus obviating the need for any additional power supply. Finally, the PIC-PG2C programmer requires IC-Prog , a freely available software, to download PIC HEX code to the PIC microcontroller. Note that the PIC HEX code is obtained from the PIC assembly code by using the MPASM assembler , also available for free. 29 (a) (b) Figure 3.2: (a) PIC-PG2C programmer and a PIC (b) PIC-PG2C programmer with a PIC mounted 3.2.3. PIC Development Board The PIC development board of this dissertation is created on a breadboard and consist of: i) a PIC16F74 microcontroller; ii) a 20MHz crystal oscillator to supply operating clock cycles to the PIC microcontroller; iii) the RS232 driver/receiver circuitry for serial data communication with the PC; iv) a DB-9 connector; and v) a breadboard area for custom circuits and easy connectivity between the PIC microcontroller and sensors/actuators. Note that Maxim’s MAX232 IC  with five 1µF capacitors serves as the RS232 driver/receiver to transform voltage levels between PC-based logic (±12VDC) and the PIC microcontroller-based logic (0VDC and 5VDC). 3.3. Software Environment The software environment for this dissertation consists of MATLAB version 6.5, Simulink version 5.0, the PIC assembly language, a newly developed Simulink library for the PIC microcontroller, MPASM, and IC-Prog. As previously discussed, the PIC assembly language is a primitive programming language consisting of an instruction set of 35 single-words. MATLAB is an interactive technical computing software and Simulink is MATLAB’s icon-based programming environment. The MATLAB toolbox for the PIC microcontroller consists of a Simulink library of PIC microcontroller functions such that based on the user selected configuration of individual I/O pins of the PIC microcontroller, Simulink automatically produces 30 and downloads the proper PIC assembly code to the microcontroller. Moreover, the MATLAB toolbox also allows data communication between the PIC microcontroller and MATLAB. Thus, the MATLAB toolbox for the PIC microcontroller completely obviates the need to manually program the PIC microcontroller. Note that the MATLAB toolbox automatically executes the assembler program MPASM and the download program IC-Prog, both of which usually require command line execution. See Refs. [52, 53] for details on programming the PIC microcontroller in command line via serial communication. The MATLAB toolbox for the PIC microcontrollers has two main components: i) a Simulink model file named Template.mdl and ii) a block library named PICLibrary. 3.3.1. Template.mdl The Template.mdl model file (see Figure 3.3) is a predesigned Simulink model file which must be used to design Simulink block diagrams for interaction with the PIC microcontroller. A function named TotalCompile has been embedded within the callback parameters of the Template.mdl so that the TotalCompile function executes at the beginning of each Simulink block diagram cycle, before the block diagram actually runs. Details of various tasks performed by the TotalCompile function are provided in a later subsection. Finally, note that renaming the Template.mdl file still preserves the callback property embedded in the file, whereas opening a new Simulink model file does not. Template.mdl model properties are modified to call the TotalCompile function at the beginning of each Simulink block diagram cycle Figure 3.3: Template and model properties 31 3.3.2. PICLibrary The PICLibrary is a custom library of Simulink blocks (in the form of s-functions) that interface with sensors and actuators connected to the PIC microcontroller. The following blocks are currently included in the PICLibrary: A2D, PinStateIn, PWM, and PinStateOut. Moreover, the library includes a block labeled IOBlock that is required in all user-designed Simulink diagrams to enable serial communication between the PIC microcontroller and MATLAB. Hardware settings and parameter requirements of each block are detailed below. 1) A2D block (see Figure 3.4) configures the analog to digital conversion module of the PIC microcontroller. Note that five of the six I/O pins of port A and three I/O pins of port E of the PIC16F74 microcontroller can be configured as eight 8-bit A2Ds. Thus, analog sensors can be directly interfaced to any of the eight A2D pins and the corresponding pin number can be passed as the parameter required by the A2D block. Figure 3.4: A2D block and parameter 2) PinStateIn block (see Figure 3.5) configures the I/O pins of port B of the PIC16F74 microcontroller to serve as digital inputs. Specifically, each of the eight pins of port B can serve as a digital input by passing the corresponding pin number as a parameter to the PinStateIn block. 3) PWM block (see Figure 3.6) configures PWM modules of the PIC microcontroller. Specifically, two of the eight I/O pins of port C of the PIC16F74 microcontroller can be configured as PWM outputs. Since the PIC16F74 microcontroller does not include a digital to analog converter, in this dissertation, we use the PWM outputs to produce the required analog volt- 32 Figure 3.5: PinStateIn block and parameter age output by varying the duty cycle of the PWM signal. Thus, two analog actuators can be interfaced to the two I/O pins of port C that produce PWM outputs and the corresponding pin numbers are passed as the parameter required by the PWM block. Figure 3.6: PWM block and parameter 4) PinStateOut block (see Figure 3.7) configures the I/O pins of port D of the PIC16F74 microcontroller to serve as digital outputs. Specifically, each of the eight pins of port D can serve as a digital output by passing the corresponding pin number as a parameter to the PinStateOut block. 5) IOBlock is necessary for every Simulink block diagram that requires interaction with the PIC microcontroller. It performs the following tasks: i) initiate serial communication between MATLAB and the PIC microcontroller when the Simulink block diagram is initially executed; ii) transmit and receive data between MATLAB and the PIC microcontroller while 33 Figure 3.7: PinStateOut block and parameter the Simulink block diagram is running; and iii) terminate serial communication between MATLAB and the PIC microcontroller when the Simulink block diagram is stopped. The callback function properties of the IOBlock include start serial and stop serial functions that initiate and terminate serial communication, respectively. In the Simulink block diagram, the IOBlock is programmed to have the first priority for execution. This ensures that all sensor and actuator data in MATLAB are first received and sent, respectively, which then is used by the corresponding sensor and actuator blocks in the Simulink block diagram. Figure 3.8: IOBlock and parameters Several Simulink blocks, such as an integrator block, require the knowledge of sampling period for their proper use in a given Simulink block diagram. In this dissertation, the IOBlock is used to determine, experimentally, the sampling period of the Simulink block diagram. Here, 34 sampling period is defined as the time required to execute one entire cycle of the Simulink block. The IOBlock determines the sampling period by averaging the time taken to run a user-specified number of cycles of the Simulink block diagram. An averaged sampling period is not expected to provide the exact sampling period for each Simulink block cycle and its use is not recommended when hard real-time constraints are to be enforced. 3.4. Integration of Simulink and PIC When blocks from the PICLibrary are used in the Template.mdl model file, a sequence of operations specified by the TotalCompile function are performed before the Simulink block diagram begins to run. The main role of the TotalCompile function is to program the PIC microcontroller and to facilitate serial communication between MATLAB and the PIC microcontroller. As seen in Figure 3.3, the TotalCompile function is set as a “Simulation start function” of “Callbacks” option in the “Model properties” of Template.mdl. The TotalCompile function performs the following sequence of tasks. First, global variables are declared and used to share data with Simulink blocks of PICLibrary. Second, sensor and actuator blocks used in the Simulink diagram are matched with the corresponding Simulink blocks in the PICLibrary. Furthermore, each block is categorized as a sensor or an actuator and its name is stored in an array of sensor/actuator structures with the specified block properties. The sensor/actuator array information is also used when data is serially communicated. Third, using the sensor/actuator block information gathered in the previous step, a PIC assembly code is generated. This step is facilitated by the fact that for each sensor/actuator block in the PICLibrary the corresponding PIC assembly code has already been created and saved as an mfile. Fourth, a portion of the IOBlock MATLAB code is generated to allow serial communication between MATLAB and the PIC microcontroller. This MATLAB code sends and receives the same amount of data that the PIC microcontroller receives and sends, respectively. Fifth, the PIC microcontroller is programmed in two steps: i) using the MPASM assembler the PIC assembly code, generated in step 3 above, is converted to the corresponding PIC HEX code and ii) using the IC-Prog the PIC HEX code is downloaded to a PIC microcontroller installed on a PIC-PG2C programmer. Figure 3.9 shows a flow diagram of the three steps involved in 35 programming the PIC microcontroller. PIC assembly code is generated by TotalCompile PIC assembly code is converted to PIC HEX code by MPASM PIC HEX code is downloaded by IC-Prog via the serial port Figure 3.9: Flow diagram of programming the PIC microcontroller After the TotalCompile function completes its sequence of tasks, the Simulink block diagram begins to execute when the user confirms that the PIC microcontroller has been removed from the PIC-PG2C programmer and properly installed onto the PIC development board. At this stage, serial communication between the PIC microcontroller and MATLAB also begins. If Simulink is stopped and needs to be run again, without any changes to the configuration of the PIC microcontroller I/O pins, then the PIC microcontroller need not be reprogrammed. Once the Simulink block diagram begins to execute, the PIC and PC exchange sensory feedback and actuator commands via serial data communication. Specifically, special function 8-bit PIC registers are used for the serial communication of sensor/actuator data [42, 46]. The IOBlock receives/transmits data from/to the PIC and stores the data in sensor/actuator global variables. 3.5. Example—DC Motor Control To illustrate the functionality and capability of the data acquisition and control hardware and software of this dissertation, position control of a DC motor is performed. Specifically, a DC motor test-bed is interfaced with a PIC-based data acquisition and control board and a control algorithm is implemented using Simulink and the MATLAB toolbox for the PIC microcontroller. The DC motor test-bed, shown in Figure 3.10, consists of an armature controlled DC motor, instrumented with a continuous rotation potentiometer, and a power module. The potentiometer output is used to obtain the necessary feedback signals and to provide a real-time 36 display of the angular position of the motor. To control the angular position of the DC motor, the PIC microcontroller applies a controlled voltage signal produced by a control algorithm running on Simulink. Digitized sensor data from PIC Analog sensor data Analog control output DC motor and power module Control output from Simulink PIC development board Simulink block diagram Figure 3.10: Hardware layer schematic In this dissertation, a proportional-integral-derivative (PID) controller  is used for the angular position control of the DC motor. The functionality of various Simulink blocks used in Figure 3.10 (see Figure 3.11 for an exploded view) is as follows. The A2D Pot block serves as an A2D block to convert the analog output of the potentiometer into an 8-bit digital data. The PID Controller block encapsulates the standard PID control algorithm. The inputs to the PID Controller block are: i) the desired angular position of the DC motor and ii) the potentiometer signal (the digitized output of A2D Pot block). The output of the PID Controller block is a controlled voltage signal to be applied to the motor. In Figure 3.11, the PID Controller block output is processed by the PWM Motor block which serves as a PWM block. Note that the power amplifier module of the DC motor shown in Figure 3.10 requires a ±5VDC to drive the DC motor. Accordingly, the PID control algorithm outputs a ±5VDC control signal. However, the PIC microcontroller can output only 0—5 VDC using the PWM output. Thus, the ±5VDC output of the PID controller is appropriately transformed to command the PWM Motor block with a 0—5 VDC command signal. Finally, by processing the PWM output from the PIC microcontroller using a simple operational amplifier based circuitry, the 0—5 VDC PWM output is converted into a ±5VDC signal for input to the power amplifier module. An analytical model of the DC motor under PID control, yielding a third-order closed-loop transfer function , is used to determine numerical values of the PID controller gains. Specifically, by requiring the closed-loop transfer function to have: i) a pair of complex-conjugate 37 Figure 3.11: Simulink block diagram: Exploded view poles with a damping ratio and natural frequency of 0.69 and 1.93, respectively, and ii) a third real pole at -80, the PID control gains are computed to be KP=1.43, KI=1.97, and KD =0.5. This analog PID controller is implemented using Simulink’s ODE4 (Runge-Kutta) algorithm with a sampling period of 0.13 second. For the computed gains, the experimental response exhibits an average 2 percent settling time of 5.9sec and a percentage overshoot of 18.25%. For comparison, the theoretical values for the 2 percent settling time and the percentage overshoot are 3sec and 14%, respectively. Figure 3.12 shows a sample experimental response of the DC motor angular position under PID control implemented using the framework of this dissertation. Figure 3.12: DC motor angular position tracking response 38 3.6. Conclusion This dissertation provided an overview of a low-cost data acquisition and control toolbox that consists of the newly developed Simulink library for PIC microcontrollers. Serial communication capabilities of the PIC microcontroller and MATLAB allowed programming of the PIC microcontroller from MATLAB and exchange of sensory data and actuation signals between the PIC microcontroller and MATLAB. The capabilities of this low-cost data acquisition and control system were illustrated through a DC motor angular position control experiment.