If you are building an Autonomous Mobile Robot, it is quite likely that you will need a navigation system. Off the shelf, navigation comes in two flavors. More on that here. In this article, I am looking at another choice that needs to be made — build vs buy.
Navigation is a hard problem. Often, start-ups in robotics underestimate the effort involved in building a navigation system that is reliable and precise. Many start-ups try to tweak open-source software to meet their navigation needs. But these do not always cut it. Mowito has hours of navigation data from robots on the field. This data helps us rigorously test and fine-tune the performance of our software for various operating environments.
So, why should a bot maker build a navigation stack of his own? We spoke to a few companies and here are some of the reasons we heard:
- I would like to own the IP for navigation
- I thought it was simple to build it
- I had vision engineers who doubled up as navigation engineers
- It is a critical piece of the stack. Would prefer to have total control over it
- I never knew the existence of off-the-shelf solutions
We urge you to ask these questions before embarking on a mission to build your own navigation stack:
1) Do I have the expertise to build “the best” navigation software?
Building navigation software requires learning and rigor. It needs engineers with a specific skill set to predict possible pitfalls and edge cases. While it can be argued that any software engineer can program navigation, there is always a learning curve. And, a ton of data and testing is needed to get it right.
2) Will I be able to launch sooner if I go with off-the-shelf navigation?
It might take many months to build the right navigation stack for your bot. As per our survey, teams have taken as much as 10 months to get the basic deployable navigation system ready. Off the shelf, navigation saves you the time and helps you launch your bot sooner.
3) Will I be able to manage project finances better if I buy off the shelf navigation?
Months of engineering time saved translate to lower build cost and improved runway. When you decide to go with Mowito’s stack, you also get Mowito’s exemplary support. That means happy customers and more dollars saved.
4) What is the focus of my bot/business?
By this, we mean, are you building a bot to target weeds, clean carpets or pluck strawberries?.
If you are making a robot to target weeds in the farm, your core IP will lie in the technology you build to precisely recognize the weeds, locate them and take them out. The same idea can be extended to other specialized tasks. You can pursue those goals while you leave the navigation to us.
At Mowito, engineers work tirelessly to build the best navigation stack for your robot. With Mowito’s best in class trajectory planner, your robot zooms around obstacles without a pause in motion. Email us at email@example.com, for a trial.