Robotic Swarm (Rover) Project - Thoughts before going in.

[AlwaysAbia]

I am a fourth year Bachelor's student in ECE currently planning my senior project and I would like feedback from people with experience going in.

I am planning to build a swarm of rovers (about 6 of them as a proof of concept because of financial limitations). Each robot will have a hexagonal surface plate on top (probably supported by springs or something similar), and the concept is that the rovers will dock together in different shapes and formations to transport objects of different shapes and sizes. (Side-note, I want to model the rovers as turtles)

Here's the plan: each robot will work on a ESP32 MCU, will feature 3 omni wheels (Oriented 120 Degrees from each other) for omnidirectional movement. On each of its six sides, the rover will have a ToF sensor to measure the distance to the next robot, an IR LED, to transmit its unique IR ID, and an IR Receiver to receive a unique ID from an adjacent robot.

the ESP32-s will run microROS and have 4 nodes, a movement node, a sensor node, a P2P communication node, and a rover to hub communication node.

I will have a separate laptop as a hub running its own ROS2 node that will transmit high-level instructions to the robots, which they will complete based on the positions of adjacent robots.

The rover-hub communication will be established via wi-fi, while the P2P communication will be established by the ESP-NOW protocol.

Besides the formation forming, I want the robots to function as a unified and mostly synchronized (which I understand, will be a problem if I do decide to use dc motors, which is why I'm also considering using steppers, as speed isn't really my concern in this case) RC car.

I'm also thinking of implementing some kind of latching/docking system on the hexagonal plates on the top, probably mechanical, to keep things simple (I've also thought magnes).

I only have a rough idea of how the actual formation forming algorithms will work, and I want to implement at least a couple (Circle, Line, etc.)

-------------------------------
As for my experience and some of my thought process:

I have no experience with ROS and thought this would be a good opportunity to learn how to use it. With the little research I've done, I've come to believe this application warrants using it (correct me if I'm wrong though).

I plan to make a PCB to house the sensors and to use an I2C multiplexer for the ToFs and 2 regular 8 channel analog multiplexers for the IR LED/Receivers.

The robot will be about 15cm in diameter.

One problem I don't know how I will solve is aligning the hexagons so they all share the same angle.

------------------------------------
I believe that is all for my rough outline for my project, I have until about the start of July to complete it.

I would love to know your thoughts on how doable this is, and would gladly accept any resources you think might help me.

[SteveThackery]

Sounds great fun! I can't offer anything helpful, but just a question: won't the IR links interfere with each other? I mean, if two should happen to transmit at the same time.

[AlwaysAbia]

I hadn't thought of that. That could be true of the ToF sensors as well (Or maybe not actually, I'll have to think about that), I'll need to find a way around that.

Thanks!

[SteveThackery]

Also (and I'm NOT being a nay-sayer) I think the timescale is pretty challenging. Each stage in the development will probably take longer than you imagine. Therefore I would recommend that you identify several intermediate points which would demonstrate solid progress and a credible sub-set of the final functionality. Having been a uni lecturer myself, I can say that students are rarely punished for biting off more than they can chew, BUT the person marking can only assign points on what they can actually see, so make it easy for them!

Furthermore, having something "in the bag" from quite early on is a wonderful boost to your confidence and greatly reduces feelings of stress. So, break the project into a series of sub-projects with demonstrable functions, such that even if you run out of time you'll have a virtual bagful of goodies to write up and demonstrate.

I'm very much looking forward to hearing from our expert colleagues - it's an interesting project.