simplify maze electronics and control

[amchagas]

Currently the maze is running three microcontrollers and a region of interest detection with a closed loop camera system. This is certainly an overkill for the tasks we are running, and complicated as at the moment there is a mix of Micropython, C++, and Python running on the system:

• One microncontroller to run four servos for food dispensing.
• One microcontroller to run four servos for controlling gratings orientations.
• One microcontroller to detect beam breaks in two infrared led and detector pairs at the maze entrance
• Camera uses python and OpenCV to detect if an animal is within one of four regions of interest.

Much simpler solution would be one microcontroller to run:

• running an add on board to control up to 16 servos
• detectiting the beam breaks at maze entrance
• detecting beam breaks at every decision point and/or at target areas (maybe even proximity sensors). This would avoid the need for a camera closed loop system

Let's break all that down into tasks so that we can step by step update the system:

• [ ] get the details of what a task consists of and generate "protocode" to address all the requirements based on the idea that everything will run on a single microcontroller.
• [ ] use micropython (as opposed to circuit python) to control the add on board and the servos (https://www.adafruit.com/product/815). Here is a candidate library https://micropython-pca9685.readthedocs.io/en/stable/servo.html
• [ ] create the IR beam and detector pairs and decide where to place them (under each grating, close to the ROIs?)
• [ ] Put all things together and write code to control a task on the microcontroller
• [ ] create code that will retrieve data from the microcontroller and plot/show learning etc.
• [ ] Create tutorial to show users how everything is running and how they can create their own tasks, analysis, etc

[amchagas]

more specifics on what needs finishing:

(main code is here https://github.com/Sussex-Neuroscience/mice-maze/blob/main/code/simplermaze/simplerCode.py)

• [ ] add code to record video (proto code already available on the main code file)
• [ ] add code that sends command to microcontroller (some of the locations where the code needs to be inserted are already on the main code file)
• [ ] code to generate trial sessions based on csv spreadsheet (https://github.com/Sussex-Neuroscience/mice-maze/blob/main/code/simplermaze/reward%20sequences%20.xlsx)
• [ ] build and test pellet dispenser (include IR beams and test them too)
• [ ] once above is done, make 3 more pellet dispensers.
• [ ] test arduino/ESP32 code (already written and loaded on GH)

---

not needed right away:

• [ ] add code that controls the probability of the animal getting a reward, even if it did a completely correct trial
• [ ] add code that controls the system in case the session needs to consider wrong locations (that is, if the animal visits the wrong location first in a trial, it does not get rewarded)
• [ ] code to measure amount of time spent in each area.
• [ ] code to record session data (number of correct trials, number of trials, etc)

[shahdalbalushi]

• [ ] add code that log entries into reward rois (timestamps, number and duration of entries, sequence/order of entries) in each trial.
• [ ] add code to set habituation gratings to horizontal rather than diagonal