2026; Mini-Strandbeest Robot (Part III)

Introduction

This is Part III of the Mini-Strandbeest series (Part I, Part II here). We will also go back to the chronological ordering here.

The Strandbeest Chronicles

Let's look at what we achieved in the next two weeks:

So yes I built the whole thing including the electronics but this one currently uses a very primitive control scheme.

Week VI: Going Plus Ultra

Most of week VI is actually covered in Part II and here I will just cover the final simulation and overall CAD model here:

The entirety of the simulation model is visible in the figure above. We already discussed the leg module in the previous post. This time another improvement I did was to use a library file where I design everything and then add them into the main simulation as common components so changing one library module updates all the copies.

Most of the parts including main body, crank frame and crank body are mainly just rigid body connections and nothing special. The main body does have a controllable revolute joint which acts as the motor in our case and uses a simple P controller to convert target velocity to torque, which kind of simulates MG995's behavior.

Fusion 360 Designs:

So this time to really really simplify the assembly, I decided to keep a separate motor holder module where both servos and the gears are attached first. Then the left and right blocks of the drive box which contains the main crank gear is attached with M5 screws to hold everything in place

The drive box then looks like this:

As you can see, I added dovetail grooves to attach other parts of the robot that will hold the batteries, microcontroller, electronics and sensors etc. These other parts are shown below:

I have three modules called top, front and rear. The rear module (with the bottom fin) has two cuboidal slots to easily store two 4xAAA battery packs to power the motors. The top module is just for design and is meant to hold the 9V battery and the connector PCB board between the batteries-servos-microcontroller. The front module will hold the PCB and the ESP32 microcontroller, with holes for easy access to the micro-USB port and some sensors (that I left out in the final version).

The full body then looks like this:

Please note that I am not a Fusion 360 power user, my timeline looks like a time-travel movie and that this is my first experience with this software (trial version actually). I will be moving to FreeCAD later down the line (which I also have 0 experience with).

Week VII: Printing and Electronics

It was time to print out all the parts starting with the leg modules

The main drive body is printed in parts starting with the motor holders, and then the left and right parts. After the servos and gears are attached to the motor holders, we assemble the left and right parts from the corresponding sides, making it extremely simple to assemble and also change motors (something I really missed out in my previous design).

I use an ESP32 as the microcontroller connecting to my private WiFi and hosting a local page with just buttons to control the robot around. For power, we have two 4xAAA battery packs providing 6V to each servo and then a 9V battery that is passed through a IN4007+LM7805 circuit to get the 5V inputs for the ESP32 and the Dot matrix display. The dot matrix was what I used originally to test the ESP32 board and I decided it leave it in the final version.

I must say I still really miscounted the amount of wires I would need to pass through the bodies and printed the wire holes too small.

Note that this version does not use an IMU that is necessary for autonomous or at least linear velocity control of this robot and will be implemented late down the line.