Why We're Building a "Basic" Line-Follower Robot

Building an autonomous robot has been on our to-do list for a long time, although our reasons were very different. I wanted to build one because of my fascination with K9, the robot dog from Doctor Who. I imagined building a robot dog who would follow me around and put things in their respective places in my lab or home. Arpit’s reason was simpler - he just loves building robots and that's it.. 

But client projects and existing products always took priority and we never got enough time and resources to put into that project. 

Well, things changed when we shifted our base to the mountains. We realized that sooner or later we are going to need an autonomous robot - in our lab, our greenhouse or in our farms ( yes, we have plans for all those). So we decided that in our free time, we would finally start building one - slowly and properly. Our goal is to build an open-source, industrial-grade robot that anyone can learn from and build themselves.

AGV or AMR 

Our first big question was : “Should we build a simple line-following autonomous robot AGV (Automated Guided Vehicle), or a more advanced AMR (Autonomous Mobile Robot) with LIDAR, cameras and high end processing power?”. 

We spoke to suppliers of autonomous robots to understand what kind of proprietary robots already exist in the market and at what cost. As expected, the AMRs were in the high budget segment. But we were surprised to find that even the simplest line-based industrial AGVs were far beyond the budget of a hobbyist or a small business owner. 

And since our goal is to build an open-source robot that anyone can build from scratch, starting with an AGV makes sense because it has :

• Lower BOM (bill of material) cost 

• Lower complexity in PCB, Mechanical design and firmware

Why a “Basic” Line Follower?

We are often asked - “why build a robot based on such a basic line-follower technology which even a college student can make?” And honestly, explaining our reason to work on something so fundamental when people see dancing robots on every other social media post is not easy. But I would still try to put it into words:

You cannot build complex products without a deep understanding of simple fundamentals.

Knowing that a product exists and how it functions, is not enough to build it properly. Every stage of development requires a deep know-how of certain basic techniques, that is only gained with experience (not from a video or a simulation). 

And surprisingly, if you try to find resources for building even a basic industrial-grade robot, very little information on its circuit boards or mechanical design is freely available online. Most of the information is either related to software development or simulations, and not about actually building one from scratch.

If someone wants to build a high end AMR today, they need to understand how to build a simpler robot first. Only then can they appreciate the real complexity of advanced hardware and software that goes into an AMR. 

So, our aim is to start building from the very basic and then progress to more complexity - documenting everything so that anyone can learn alongside us from scratch.

Our Plan :

So, here is the roadmap for our open source autonomous Robot: 

• Build a small, robust 1ft x 1ft x 1ft robot platform with max payload of 5kg

• Design a custom controller PCB with interfaces for current and future expansion

• Implement navigation functionality

• Add Automatic map discovery 

• Create a web interface to view the map and send point to point movement commands 

• Open source everything: PCB design, mechanical files, firmware.

• Build APIs so others can develop applications on top of it - e.g. for factory floor movement, restaurant table service etc.

• And if possible make video tutorials for our YouTube channel.

In the coming days, we will decide the first application for our robot which I will talk about in my next blog post. 

Thank you for reading. See you next week.