Sensors technologies – part 1
Continuing on in my discussion on Symbiotic Autonomous Systems as part of the Future Direction Committee Initiative I will look now into the technologies that are providing the building blocks for their evolution.
Technologies for sensing the ambient have improved significantly and will continue to do so in the next decade. There is the possibility to “sense” electrical fields, acceleration, displacement, temperature, force –including pressure, touch-, presence of chemical compounds –including radiation-, sounds.
Additionally, image capturing at high resolution and in low light condition is no longer a problem. In the area of “vision” significant progress are still being made in creating sensors that can detect objects and identify their relative position in a field of view. These are exploiting Light Field technologies and decrease significantly the processing required to “understand” what is in an image, something that is crucial for an autonomous system, like a self driving car but more generally in robotic vision.
The technologies adopted in the “making” of sensors are based on mechanics, optics, electronics, chemistry. In each area there has been impressive evolution, both in terms of performance (accuracy, sensitivity), dimension (miniaturisation) and cost decrease. Sometimes this is achieved through software that can make up for hardware limitation, through signal processing and multiple sources sampling.
An interesting evolution, already in sight, is the embedding of sensing capability in materials (smart materials). The trend is towards using construction materials that have sensing capabilities as an integral part of their material structure, like concrete that can “sense” pressure and detect stress, rubber that can become a sort of skin detecting temperature and pressure (touch) variations, material surfaces that detect the presence of light and its “quality”. This evolution is of particular interest in robotics where smart materials can be used both as “skin” –detecting external conditions- and as muscles – detecting strength and position of the robot’s parts.
Another important evolution is the embedding of processing and communications capabilities in sensors. This enables some level of local signal processing and the creation of local sensors networks that integrate the data acquired and generate a meaningful information (like a bridge detecting a “moving stress” on the structure and interpreting it as a heavy vehicle crossing it vs a stress produced by a mud slide on one edge of the bridge).
Sensors networks are becoming a crucial part of any autonomous systems since they can provide data that properly analysed lead to context awareness.
The evolution of sensors is matching the evolution of IoT (that are in a big way made up by sensors or that embed one way or another sensors). This creates volume of scale that will keep pushing performances, production efficiency and cost decrease in a virtual spiral that accelerate innovations. There is no perceivable stumbling block ahead, at least in the coming decade and this will greatly benefits autonomous systems.