{"id":8741,"date":"2023-11-20T15:47:46","date_gmt":"2023-11-20T14:47:46","guid":{"rendered":"https:\/\/dronesonen.usn.no\/?p=8741"},"modified":"2023-11-20T15:58:57","modified_gmt":"2023-11-20T14:58:57","slug":"group-1-week-12-13","status":"publish","type":"post","link":"https:\/\/dronesonen.usn.no\/?p=8741","title":{"rendered":"Group 1 – AutonoMouse – Week 12 & 13"},"content":{"rendered":"\n

Github<\/h2>\n\n\n\n

https:\/\/mavbs@dev.azure.com\/mavbs\/MicroMouse%202023\/_git\/MicroMouse%202023<\/a><\/p>\n\n\n\n

Andreas<\/h3>\n\n\n\n

Week 12<\/h4>\n\n\n\n

This week has been spent trying to ensure we have a finished product within the delivery date. This entailed fixing and decluttering sensors that were half-hazardously soldered together a month and a half ago, soldering the rotary encoder circuitry,  and preparing to model the motors.  <\/p>\n\n\n\n

Since we wanted to switch the motors and add rotary encoders to the robot, we decided that we would have to change up our entire design. This meant we had to disassemble the whole robot and transplant it into the new, 3D-printed chassis. During this process of disassembly, we realised that some of the sensors had been broken and there was a lot of clutter in the wiring. This led to me spending the week attempting to save the components and soldering new sensors so that we could assemble them in the new chassis. <\/p>\n\n\n\n

Before this, though, I performed tests so that in a perfect world, I would be able to model the new motors in MATLAB and hopefully be able to simulate the intended motor controller, though this was set on the back burner when I realised a lot of work would have to be put into the circuitry of the robot. The tests were intended to find the inductance and resistance of the motor by varying its frequency. This would enable me to use the general motor transfer function so that I could simulate a rough equivalent of the motor and create a controller, sadly we will most likely not have the time available to do this. <\/p>\n\n\n\n

Since the next week is the last one, I will mostly be spending it fixing the previous bits to assemble the bot and assist the computer engineers as much as I can so that we can hopefully be able to finish a labyrinth in however much time it takes <\/p>\n\n\n\n

Week 13<\/h4>\n\n\n\n

As predicted in last week I have spent the week fixing the electrical circuits and doing some final documentation on some of the circuits we have slightly changed. Luckily, since we decided to do the filter digitally instead of analogue, we simplified the circuits to such a degree that simulating the circuits was overkill. <\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n
\"\"<\/figure>\n\n\n\n
\"\"<\/figure>\n\n\n\n

On the programming side, I have been improving the filtering and ensuring that my MATLAB code is readable. I have also assisted Bendik in making a PID controller to the best of my ability, though we moved away from it as we had little to no time actually figuring it out. <\/p>\n\n\n\n

\"\"
The frequency spectra in dronesonen, downsampled, compared to the filtered signal.<\/figcaption><\/figure>\n\n\n\n

I also assisted in the logic behind the centring function of the robot, ensuring that the robot stays in the middle of the track when driving in a straight line. This worked basically by checking the side sensors to see if there was a significant difference, then turning to a specific angle towards the middle, driving for a calculated distance and straightening it to run parallel to the wall again. This function is mostly theoretical as of writing this. <\/p>\n\n\n\n

Practically, I have mostly been fixing broken circuits and ensuring that they all fit within the robot and function as intended. One of the motors was broken in shipping from Marte’s home to the school and needed replacing, as well as one of the sensors giving off some weird sensor data. I will be spending some further time attempting to fix this.<\/p>\n\n\n\n

Bendik<\/h3>\n\n\n\n

Week 12<\/h4>\n\n\n\n

This week, while I have been waiting for the hardware to be ready, I have created a main project in ADA to include all the different pieces of software for the Micromouse. This includes a PiD controller and the IR signal filtering algorithm. When Vendel finished implementing her path finding algorithm I will also implement that into the project. The idea is to use tasks to make a concurrent program. I also helped out with soldering the wires for the new motors on the MicroMouse. <\/p>\n\n\n\n

Week 13<\/h4>\n\n\n\n

This week I tested my PID controller and implemented the wheel encoders into code. I wasn\u2019t able to get the PID controller to work, sadly. As time was of the essence, I decided to comment out the PID controller, and implement my own speed controller to try to get the mouse to drive straight. <\/p>\n\n\n\n

Using the wheel encoders, I made code that adjusted the motor speeds via PWM (pulse width modulation). For every pulse of the encoder, I added to the \u201cDistance Travelled\u201d variables for both the wheels. If one of the wheel travelled further than the other in a certain amount of time, then the motor speeds were adjusted to compensate. This worked decently well but still needs calibration. <\/p>\n\n\n\n

I also made a \u201ccentering\u201d procedure, that takes in the direction and distance that needs to be driven towards the center, given that the mouse turns with a 31.5 degree angle towards the center. This has yet to be tested properly. I also mase a procedure to turning the mouse 90 degrees that also needs to be tested. When the front IR sensor senses that a wall is nearing, it will turn 90 degrees. <\/p>\n\n\n\n

The last minutes before the deadline will be used to test the mouse in the maze and make a short video, brushing up the code in the process. The week towards to presentation will also be used on coding to integrate everything together, the algorithm, wheel encoders, IR sensors and hopefully get it working properly. <\/p>\n\n\n\n

Marte<\/h3>\n\n\n\n

Week 12 & 13<\/h4>\n\n\n\n

De siste par ukene har v\u00e6rt hektiske for \u00e5 bli ferdig med mikromusa. Disse to ukene har jeg laga to design, produsert deler og ordnet sammenstillinger. Jeg har hatt mye kvalitetstid med 3D printeren jeg fikk for et par uker siden, og i begynnelsen brukte jeg mye tid p\u00e5 \u00e5 f\u00e5 deler til \u00e5 passe sammen mtp toleranser og krymping, men det har g\u00e5tt litt bedre etter hvert som jeg har f\u00e5tt mer erfaring og kjenner printeren bedre.
Planen var \u00e5 teste med forskjellige materialer, men vanskelighetsniv\u00e5et til PLA har v\u00e6rt h\u00f8yt nok for meg n\u00e5 i starten.<\/p>\n\n\n\n

Det var dessverre f\u00f8rst n\u00e5 i uke 13 at jeg fikk hardware helt ferdig p\u00e5 den nye varianten, s\u00e5 data og elektro har dessverre blitt hindra fra \u00e5 kunne gj\u00f8re noen av oppgavene sine og testing pga dette.<\/p>\n\n\n\n

Jeg lagde f\u00f8rst et design med de nye motorene, SolidWorks fila har dessverre kr\u00e6sja p\u00e5 denne f\u00f8rste varianten, men her er bilde av ferdig produkt av det f\u00f8rste designet:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n
\"\"<\/figure>\n\n\n\n

Jeg valgte \u00e5 flippe Kitroniken og Mikrobiten \u00abp\u00e5 hode\u00bb etter godkjenning fra de andre. Dette for \u00e5 redusere h\u00f8yden til mikromusa, og at Mikrobiten er bedre beskytta p\u00e5 denne m\u00e5ten.<\/p>\n\n\n\n

Et 9V batteri er festa ved at den ligger p\u00e5 en \u00abhylle\u00bb p\u00e5 innsiden av veggen, og er festa med dobbelsidig borrel\u00e5steip (takk for spons av denne til gruppe 4).<\/p>\n\n\n\n

P\u00e5 drivakselen sitter en motorencoder, og videre en cylinder med tannhjul. Girforholdet her ble 20\/40 = 0.5. Girforhold 0.5:1. Det viste seg at motorene roterte mye treigere enn antatt, og jeg valgte \u00e5 g\u00e5 for et nytt design for \u00e5 endre p\u00e5 girforholdet.<\/p>\n\n\n\n

Motorencoder:<\/p>\n\n\n\n

Jeg hadde tenkt til \u00e5 produsere disse i b\u00e5de PLA og kryssfin\u00e8r, men jeg har kun f\u00e5tt tid til \u00e5 teste i PLA.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Motorencoder i PLA. Ser en del “stringing” fra plasten, men vi fikk bort det meste f\u00f8r bruk. Den satt godt p\u00e5 drivakselen (etter en del modifiserte prints), og ga tilstrekkelig m\u00e5linger. <\/p>\n\n\n\n

Hvis jeg hadde hatt bedre tid hadde jeg fors\u00f8kt \u00e5 laserkutte en i kryssfin\u00e8r ogs\u00e5, da jeg tror n\u00f8yaktigheten og presisjonen p\u00e5 hullene ville blitt enda bedre. <\/p>\n\n\n\n

Hjulene:<\/p>\n\n\n\n

Jeg gikk til innkj\u00f8p av fire sm\u00e5 kulelagere med indre diameter p\u00e5 3mm. Det var overraskende vanskelig \u00e5 f\u00e5 tak i en aksling i metall i denne dimensjonen, s\u00e5 jeg gikk til innkj\u00f8p av en \u00e8n meter lang M4 gjenga st\u00e5lstang.  <\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Jeg markerte hvor jeg trengte aksling til kulelagerne, og lengde p\u00e5 akselen. For \u00e5 slipe ned gjengene og oppn\u00e5 en diameter p\u00e5 3mm brukte jeg en Dremel og slipte det ned.<\/p>\n\n\n\n

N\u00e5r fire akslinger var ferdig la jeg de i fryseren noen timer, og deretter presstilpassa de p\u00e5 kulelagerne. S\u00e5 la jeg disse tilbake i fryseren, og presstilpassa kulelager med aksel inn i hjulene:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Dette var tidkrevende, men resultatet ble veldig bra, og hjulene roterer bra n\u00e5r de er p\u00e5montert.<\/p>\n\n\n\n