So, I want to build a cool robot, that can do cool stuff, his name will be Rolf. How to achieve that? Both from a technical stand point but also from a project management point of view. What skills do I need to learn? How much labor should I invest in this? How much money is okey to spend on a hobby? How do I stay motivated for something where real results can take years of effort? How do I work on this consistently over long stretches of time?
As I write this I realize that this post might derail completely from what I first intended. Originally I thought that this would be a simple, short post with establishing some easy to reach milestones in the short-term just to get the ball rolling for this project, but I have a sneaking suspicion that the post will morph more into some weird mix of motivational psychology and a personal diary.
So, what’s a roadmap and why do I want one? I think its important to define some goals and to plan a bit ahead. Goals are nice, everyone gets fulfilment by accomplishing their goals, its a good motivator. Setting a goal also takes away the uncertainty by answering the question of what to do when I have some spare time I can work on this hobby, I simply work on tasks that can accomplish my next goal on the roadmap. So why plan ahead, why not just set my next goal and be happy with that? Well, planning ahead allows setting larger goals as you can just make many smaller goals in sequence so that when you reach all your mini-goals you are either close to or have reached your larger overall goal. That’s quite a neat trick. The next trick is that planning ahead separates problems into two bins. The first bin is quite small and the label says “NOW” that is where your current goals lives, and it is where all your focus and attention should be. The other bin is endlessly deep and the label says “FUTURE” that is where all the other goals live and you are not allowed to work on them. In that sense planning ahead actually becomes a way to increase focus on the current task in front of you, it does however require that planning the project and executing tasks in the project has to stay separated in your head-space.
So, creating a plan for what tasks to accomplish seems useful, but how do we make a plan for a project you do not know how to execute successfully? The first solution I could think of was to admit that I don’t know how, but I can try at least! At this point I think of the roadmap more as a motivational poster hanging above my desk. It’s purpose is simply to keep me working on the project and force me to break down my huge ideas into more manageable, bite-sized tasks. The roadmap is in no way final, nor is it detailed, and its certainly not guaranteed that I will stick with it. But here it is in the pretty form of a numbered list:
The roadmap
- Document the Hiwonder SW stack
- Install ROS on the TonyPi
- Create proof of concept application controlling a single motor through ROS
- Create a jobtracker, kanban board or trello on this website to track what I’m working on in more detail
- Set up the basic pieces of a development environment and pipeline for writing, deploying and testing what I create
- Create a gazebo simulation model of the robot
- Investigate the performance of the Hiwonder servos
- Do some reading on strategies for bipedal locomotion in robotics
- Find the simplest possible gait algorithm that can be implemented in ROS for Rolf
- Explore other topics in robotics such as computer vision, path planning and navigation and create roadmaps for those features
- Use what I learnt to see if the mechanical design or servos need to be upgraded to achieve better performance in locomotion
- Play around with AI stuff and machine learning, everyone seems to be doing it now so why not me also
- Try out more advanced control strategies for locomotion
So that’s it then. At the start the list consists of very specific tasks that are small and easy to accomplish, but towards the end the topic is broad and a single item could likely take years, but hey if it gets the ball rolling it has served its purpose. It could be fun to revisit this list sporadically to see how well I stick to the original plan over time. Thanks for coming to my TED talk.