README.txt Steven Rofrano 2014-07-28 This file describes the basic operation of the MouselabRig software package. It includes an index of files and basic functionality. INSTALLATION To install the system, one needs the software as well as hardware components. The hardware consists of (1) the Arduino, (2) a completed and functional MouselabShield (along with its necessary hardware), and (3) an internet-equipped Raspberry Pi (Pi). The software includes the Arduino interface software (ArduinoCode) and and the main conditional reward program (PiCode). HARDWARE A general overview of the hardware layout follows. Emulate the design decisions of an existing copy of MouselabRig for specific instructions. CONNECTIONS (1) USB connection from Pi to Arduino (2) Ethernet from Pi to imec network (3) Ribbon cable from MouselabShield to Rig (4) 12V power adapter from MouselabShield to outlet (5) Power over Mini USB from Pi to outlet With these connections in place, it is possible to install, set up, and run the software system. SOFTWARE The following lists the procedure for installing the software and setting it up for the first time. ARDUINO (1) Dowload and install the Arduino IDE software to a computer of your choosing (http://arduino.cc/en/main/software) (2) Clone the repository for MouselabRig onto your computer (https://github.com/srofrano/MouselabRig) Further information on this process can be found on github.com (3) Open MouselabRig/ArduinoCode/run.ino in the Arduino IDE by double-clicking the file (4) Plug the Arduino into the computer and compile/upload the program (5) The Arduino is now ready to use RASPBERRY PI (1) SSH into the Pi and install Pip (http://pip.readthedocs.org/en/latest/installing.html) (2) Use pip to install pySerial (http://pyserial.sourceforge.net/pyserial.html#installation) (3) Clone the MouselabRig repository into the home directory USAGE Running the system is fairly straightforward. After correct installation, simply modify parameters.py with your desired settings (consult in-file comments and section below) and run the program with the command pi$ python main.py from the command line on th Pi. The program will ask you if you want to begin. Enter 'y' when ready, and the program as determined by the settings in parameters.py will begin to execute. OPERATION The main program will accomplish five tasks simultaneously. (1) Messaging - The program establishes communications with the Arduino, which monitors the state of the sensors and controls the hardware. Communication between the devices is one-to-one, in that each device sends information only after the previous message from the other device has been received. When the program is started, it sends a reset command to the Arduino. The Arduino reboots, sends the first message, and they alternate thereafter. (2) Processing - After a message is received, the device will parse out the data and process it in real time. This means that processing occurs before new data can be received/transmitted. The Pi will take the state of the sensors and generate position data (speed, distance travelled, etc.). The Arduino will look for commands from the Pi to make some modification to hardware. (3) Logging - The Pi will log all the data as it is processed in a txt file at a location determined in parameters.py. It logs all relevant information on a single line, to be parsed in post-processing in MATLAB. (4) Displaying - The Pi will display the current time, position, and other relevant data on the command line. This notifys the user of the current state of the experiment. (5) Reward testing - The Pi will constanltly check if the reward conditions have been met. When they have, the Pi will issue a command to the Arduino to open the valve for a predetermined amount of time. These functions are elaborated upon in the discussion below. METHODOLOGY In the case of our lab, an experiment is simply a set of environmental rules that determine the reward for the mouse. Any experiment answers the question, "Under what conditions is the mouse given the reward?" The goal of the MouselabRig system is to allow researchers to design and implement experiments easily and effectively. To control an experiment, a researcher merely has to write a short Python module specifying their desire conditions using pre-defined variables and paramters, and the MouselabRig will take care of the rest. There are two main aspects of the system that help control the hardware so the researcher doesn't have to. The sensors on the rig will be controlled by and interface directly with the Arduino microcontroller. The Arduino subsequently interfaces with the Raspberry Pi (Pi) microcomputer, which determines how the hardware should behave. In this system, the Arduino is a dummy board--its role is merely to translate the requests from the Pi into a fomrat understood by the hardware, and relate any information from the sensors back to the Pi. Once at the Pi, the data from the sensors is logged, processed, and fed into the running experiment. The Pi then sends new orders back to the Arduino should the experiment request that hardware be manipulated. The researchers role in all of this is to provide the conditions under which the Pi sends orders back to the Arduino. From three lines to one hundred, an experiment simply runs code based on universal parameters without changing the communication or data processing on the back end that's used to make the system actually run. For example, import rig exp = rig.run() while true: if exp.numLaps() - exp.oldLaps() == 1: exp.valve(500) is all it takes for a researcher to run an experiment that opens the water valve for 500 ms every time the mouse completes a new lap. Of course, there is much more going on in the background that allows all this to happen, and these other processes (although they need not be changed often) are described below. COMMUNICATION The Arduino and Pi communicate through a serial data connection. They are physically connected with a USB cable and using Arduino's Serial module and Python's pySerial extension, they can talk back and forth using serial data communication. The commands used for the Arduino are Serial.print() and Serial.readBytesUntil(). The Pi uses serial.readline() and serial.write() to read and write to the serial bus. The following is a summary of the functions and parameters of the serial transfer methods for both devices. On the Arduino, Serial.print(<'message'>) prints the string <'message'> to the serial port Serial.readBytesUntil(<'endline'>, <buffer>, <characters>) reads at most <characters> number of ASCII characters, stopping at the character<'endline'>, and writes them to a character array <buffer>. On the Pi, serial.readline() reads the incoming serial data stream until it reaches the newline ('\n') character. serial.write(<'message'>) writes the string <'message'> to the serial port. Once at the other device, the string must be split and the values from the stream extracted. This is easily accomplished if we specify a communication protocal for sending information between the two devices. The transfer protocol is simple--a header, a trailer, and a comma seperated list of keywords and values make up the transfer string: "SENDING_DEVICE,KEYWORD_1,VALUE_1,KEYWORD_2,VALUE_2,...,\n" This takes the following form with a message sent from Arduino to Pi: "ARD,MILLIS,1345,PHOTO_STATE,1,ENC_STREAM_1,1,ENC_STREAM_2,0,\n" Because the read methods for both devices include a provision for the last character of a line, all communications shall end in '\n' and start with a recognizable header. This allows usto easily process an incoming stream with a simple algorithm that can be used by the Arduino, the Pi, or any MATLAB script used for processing the log files afterward. Here is the basic algorithm demonstrated in Python code: import serial msg = serial.readline.split(",") # Identify important keywords and values if msg[0] == 'SENDING_DEVICE' : for i in range(1, len(msg) - 1) : # no need to count the last vale '\n' if msg[i] == 'KEYWORD_1' : # identify the keyword keyWord1 = msg[i + 1] # assign the next list entry to an appropriate variable elif msg[i] == 'KEYWORD_2' : # keep going until you have all desired values keyWord2 = msg[i + 1] elif ... This bit of code is efficient and can be easily ported to other languages and used for other purposes. An obvious drawback is that it doesn't self-identify the incoming values--that is, it has to be expecting a certain keyword in order to read its corresponding value. However, in this type of system this poses no real problem because we have to be expecting the value anyway, because the rest of the code depends on that value being assigned to a specific variable. To add additional values, simply follow this pattern and nothing should break.