Worldwide, the number of elderly people (65 years and older) surpassed the number of children under 5 years of age in 2019. It is likely, in my lifetime, that there will be more people needing assistance in their lives than there will people who can provide it. I don’t expect to be able to afford especially extraordinary care, nor do I expect not to need it as I age. My alternative is to create the needed technology so it will be there when I need it.
Wimble Robotics is a one-person effort to create an assistive robot for my personal use.
I understand that this will be hard. In fact, having spent several years on the effort so far, I know that everything about robots is hard. This is a theme that you will see me remind you again and again.
For years I have been creating various experimental robots, from tiny robots the size of your hand, to robots the size of an electric wheelchair. My expectation was not that I would quickly reach even that dreamed of major goal of the home roboticist, that of fetching a drink from the refrigerator and bringing it to you. Rather, there is so much technology that is required for a robot to be useful even in a tiny way. This is the hardest learning journey of my life, so far.
Creating a robot that is trustworthy (it won’t harm me or my possessions), reliable (it will work when I turn it on) and predictable (it will do the tasks as designed, repeatedly as needed) is a tall order, indeed.
As I’ve often said in my talks to various groups about robotics, just because you tell a motor controller to move the motors, doesn’t mean that the motors actually move. If the motors move, it doesn’t mean that the attached wheels turn. If the wheels turn, it doesn’t mean that the robot will move at all. If the robot moves, it doesn’t mean that it will move as commanded.
Everything about robots is hard. Batteries fail, Electronics fail. Screws loosen or fall out. Wires break from vibration. Networks stop communicating. Wheels get slippery. Robots are a breeding ground for failure.
Every few months, I create a new robot which improves on the predecessor because of what I learned up to that point. Sometimes I want to explore one new aspect, like reliable sensing of the position of the robot over time, relative to some time in the past. Sometimes I want to try out a new bit of technology, like stereo depth-sensing cameras, or time of flight distance sensing. Sometimes I work on just one dimension of the overall problem, such as fallback technology for when the network fails, or reliable startup of the robot. Progress is made with each generation. But the progress is slow. Especially for a single-person effort made by a retired computer scientist.
This blog is a somewhat free form brain dump of the journey to make a personal, assertive robot, to be ready when I need it.