I believe we should develop simple and inexpensive robots that exist to harvest energy, process data, replicate, and ‘die’ like ordinary life forms. We need to research already proven biological systems to provide insight for new directions in the field of robotics. Two great developing forces are BEAM robotics and a practical commercial outlet of them at solarbotics.com
What is important with these types of robots is that there is no ‘brain’ needed to operate. Most species of life don’t have a brain to speak of, notably jellyfish. They manage to survive and reproduce effectively without one. With the simplest robot design in mind we can create a sustainable network of autonomous machines.
Using energy from the sun is obviously the first source. But using this harvested energy is inefficient for locomotive use. Instead, chemical energy should replace it while the electricity is used solely to process information. The fuel should be a digested mass of organic matter. The resulting composition should be full of potential chemical energy to be consumed by cell-like chemical engines.
The use of a microcontroller and radio communication can organize these machines by processing a real-time virtual array. Each robot is aware of each other and a botmaster signal for reprogramming. Within this neural net (array) each individual robot will provide statistical feedback. When a majority of the robots can detect an unexpected drop in light exposure, for example when a large cloud passes, it will trigger every machine instantly to go into a low power hibernation. You could compare this to when a flock of birds take off simultaneously when only one is startled to flight.
The application of these simplistic robots is not commonly approached. Their purpose is to exist self sufficiently and provide a small cloud computing array. Consider the microcontroller as the thought organ, it’s radio is a communicating organ, and the chemical actuators used to move is akin to their muscles. Providing a biological parallel allows these machines to emulate the proven systems in life today.
Why not consider other uses for these machines? Providing difficult tasks requires specific programming which makes it less autonomous. Receiving feedback from the simple data can provide clues to help the machines to evolve. And as the machines become better at their simple task of gathering energy and computing, a massive resource of computing power becomes available. Maybe the PC’s of the future will be sprawled across your lawn, processing this webpage by cutting your grass and soaking up sun rays. Or maybe the nodes are used globally as brain cells of an omnipresent digital organism…