“Control is everywhere”. This sentence has been used to promote the importance of Systems Engineering in high-level studies as well as being a propellant of market diversification of specialized equipment manufacturers. There is no doubt that this is true. Control is an essential piece of most aspects of our lives, but the original dream goes beyond.
Visionaries, scientists and writers, foresee a world strongly instrumented with sensors, actuators and computing resources, populated with software entities capable of anticipating the actions leading to a better performance of real-world entities. This is a very complex scenario proper of Artificial Intelligence research involving low-level, entity-level, global and ethical requirements and objectives. This also involves a global tradeoff among systems having a strongly non-linear behavior. Realizing this vision constitutes a major challenge.
As in the first large revolution of Systems Engineering that moved it from the analog to the digital scenario, we face now a new revolution also related to the coupling of Control and Computing. Some of the computing technologies involved in this qualitative change are:
- Pervasive communication and computing
- Cloud computing and services
- Deep learning
- Agents and assistants technology
Some years ago, the experience of moving from the context of “Wireless Sensor Networks” to the “Wireless Sensor and Actuators Networks” has shown us that closing control loops in a complex and distributed computing environment unveils new challenges. While communication delays of sensor information can be measured, and sometimes compensated into the control software, acting information traveling in a communication infrastructure is less robust. This communication must be done in the context of a contract of Quality of Service. To ensure a reliable performance, local actuators (“local” means “wired to the analog component”) must have computing capabilities to deal with this “contract” and algorithms to manage contract violations and all other exceptional situations.
In a similar way, looking at a big scale Pervasive Control reveals the importance of aspects of the design related to the reliability and confidence as well as rethinking the role of the human component. As a previous step to face the Pervasive Control dream, there must be solid developments on at least the following technologies:
- Mixed Criticality management and execution platforms.
- Performance related to resources availability. Graceful degradation.
- Stability guarantee.
The capabilities of computer and control technologies are now stronger like never before. Nevertheless it is not as simple as “putting it together”, there is a lot of work to do in the gap. Surely this is a source of opportunities for research and business.
Article provided by José E. Simo Ten (LinkedIn profile) Universitat Politécnica de Valéncia, Valencia, Spain IFAC Technical Committee 3.1. Computers for Control